Use PWM to control your Jebao DCT pump using brushless motor controller from ebay

[d0ughb0y]

I got this brushless motor controller from ebay for under $10.
http://www.ebay.com/itm/DC-24V-DC-brushless-Driver-DC-motor-speed-controller-3A-/181647891641?hash=item2a4b0edcb9:g:tKAAAOSwYGFUwl4I

It can control a 24v brushless DC motor under 3 amps.
I tested this on my DCT 4000 pump and it works fine.
It takes in a 0-5v dc voltage (5v PWM works as well).
If you want to connect this to an Apex and outputs 0-10v, all you need to do is connect a 10k resistor in series, as the board has a 10k pull down resistor for this pin.

There is this version that works with higher power pump motor, up to 6 amps
http://www.ebay.com/itm/12V-24V-36V-dc-500W-Brushless-Motor-Drive-Board-Balanced-Car-BLDC-Controller-ne-/322140411977?hash=item4b01109849:g:S5gAAOSwuhhXVp1h

They both should work the same (both use the same motor controller chip), but I think for $2 difference, get the one that can handle higher amsp.

The connections are self explanatory.
The A B C motor connections go to corresponding W V U pins.
If you are looking into the 3 pin connector coming from pump with the notch at the bottom, W is left, V is upper/middle, U is right.

The module even has a direction control pin. So this will work in controlling a CP-40 or Gyre. I don't have either one of this so I don't have any further info.

The controller does not have built in soft start, so you will have to ramp up the pwm from your controller. The advantage using this is you can dial down your pump down to about 30%. The stock DCT controller minimum is 70% (one led does not mean 10%, it starts at 70% to 100% in 3% increment).

It has built in over current protection, which is equivalent to preventing the pump from running dry.

Since this is brushless motor with no hall sensors, the "stall" protection is built in. Meaning, the pump will stop sending out voltage if it is physically stopped from spinning.

youtube video (https://youtu.be/MEeC7Jjb1ww)

<iframe width="560" height="315" src="https://www.youtube.com/embed/MEeC7Jjb1ww?rel=0&amp;controls=0&amp;showinfo=0" frameborder="0" allowfullscreen></iframe>

[karimwassef]

Wow. Can you post the image with the connections?

http://i.ebayimg.com/images/g/S5gAAOSwuhhXVp1h/s-l1600.jpg

I'm asking because the powerheads and the old DC pumps were just 24V/CONTROL/GND

http://www.ultimatereef.net/uploader/2013Q4/DSC_0017_6.JPG

but the same connector on the DCT and CP is actually W U V?

What are Z/F and VR?

[d0ughb0y]

For the input side, all that is needed is a 0-5v variable voltage or 5v pwm and Gnd to control speed.

The motor connections that worked for me is WVU corresponding to ABC. Not sure if the green board label will work. It says CBA to WVU.

The z/f is for direction. If you leave it unconnected, it defaults to clockwise.

[karimwassef]

ok. got the higher current controller to drive my DCT15000

I'll use my Apex with a divider to get 0-5VDC.

So from the control board:

VCC goes to 24VDC power supply input positive
GND goes to 24VDC power supply ground negative

and looking at the connector image I posted above:

the pin labeled 24V connects to MA/U
the pin labeled CNTL connects to MB/V
the pin labeled GND connects to MC/W

The output 5V and GND go to a 5VDC supply?
The signal and GND goes to the APEX resistor divider midpoint and ground.

Z/F unconnected

and VR?

[stage3-s4]

great find thanks for the info. Also looking for confirmation on wiring.

[d0ughb0y]

As I described in my 2 previous post,

MA - W
MB - V
MC - U
GND - power supply GND
VCC - power supply 24V

connector from pump -
If you are looking into the 3 pin connector coming from pump with the notch at the bottom, W is left, V is upper/middle, U is right.

I tried switching the W and U and the pump did not spin. I don't think it will do any harm switching connections. If you ever used brushless motors on drone, which runs at much higher amps, they tell you to switch the wires around unti the propeller spins clockwise.

VR - 0-5V analog or 5vPWM. Connect 10k in series for 0-10v signal from apex. The value of 10k is based on the board I tested on. I think for the Green board, it may be 100k. You can measure the resistance from VR pin to Gnd, then just put a resistor of the same value in series.

Gnd - Gnd from controller

No connection for all other pins.The other pins are as described in ebay listing. Do not connect anything else. 5v is an output.
Z/F is to reverse direction. Connecting to 5v will make motor spin backwards. I did not try this on the DCT pump. I think it can be used on CP-40 or gyre. But probably not really needed as reverse flow on these pumps don't really do anything. If someone wants to test it on their pump, they can post their results in this thread.

Connect it as described. If you already have an apex interface that has a voltage divider and you connect it to VR, your lower resistance will become the combined parallel resistance of your lower resistor and the input resistor and you may not get the 10v to divide in half.

[stage3-s4]

As I described in my 2 previous post,

MA - W
MB - V
MC - U
GND - power supply GND
VCC - power supply 24V

connector from pump -
If you are looking into the 3 pin connector coming from pump with the notch at the bottom, W is left, V is upper/middle, U is right.

I tried switching the W and U and the pump did not spin. I don't think it will do any harm switching connections. If you ever used brushless motors on drone, which runs at much higher amps, they tell you to switch the wires around unti the propeller spins clockwise.

VR - 0-5V analog or 5vPWM. Connect 10k in series for 0-10v signal from apex. The value of 10k is based on the board I tested on. I think for the Green board, it may be 100k. You can measure the resistance from VR pin to Gnd, then just put a resistor of the same value in series.

Gnd - Gnd from controller

No connection for all other pins.The other pins are as described in ebay listing. Do not connect anything else. 5v is an output.
Z/F is to reverse direction. Connecting to 5v will make motor spin backwards. I did not try this on the DCT pump. I think it can be used on CP-40 or gyre. But probably not really needed as reverse flow on these pumps don't really do anything. If someone wants to test it on their pump, they can post their results in this thread.

Connect it as described. If you already have an apex interface that has a voltage divider and you connect it to VR, your lower resistance will become the combined parallel resistance of your lower resistor and the input resistor and you may not get the 10v to divide in half.

Great info here. Any reason I cant use a pot to drop 10V down to 5v?

[d0ughb0y]

If you use a potentiometer, you would not know where to stop turning. Turning all the way will feed 10v to the 5v input.

If you are referring to using a potentiometer to get 5v out of the 10v signal, then, as I already mentioned, when you connect to VR pin, your lower resistor value would change. This is just basic parallel resistance calculations. If you use a 100k potentiometer and turn it halfway to 50k + 50k to get 5v, when you connect to VR pin, the lower resistance of 50k would be in pareller to 100k (or whatever the pull down resistor value). Lets say it is 100k, then your effective combined lower resistance becomes 33k. So your output now varies between 0-4V instead of 0-5v, so you will run the pump at 20% below capacity.

The proper way is to simply measure the resistance from VR pin to ground. If it is 100k, then you can put a 100k in series from 0-10v signal to VR pin.

Or just get a controller or controller interface that outputs 5v PWM.

FWIW, I think 0-10v is old technology. Anything recent (within last 3 years) really should be 5v pwm capable. Like Archon controller. I think when looking for a new controller, that should be one of the main feature to look for.

[stage3-s4]

If you use a potentiometer, you would not know where to stop turning. Turning all the way will feed 10v to the 5v input.

If you are referring to using a potentiometer to get 5v out of the 10v signal, then, as I already mentioned, when you connect to VR pin, your lower resistor value would change. This is just basic parallel resistance calculations. If you use a 100k potentiometer and turn it halfway to 50k + 50k to get 5v, when you connect to VR pin, the lower resistance of 50k would be in pareller to 100k (or whatever the pull down resistor value). Lets say it is 100k, then your effective combined lower resistance becomes 33k. So your output now varies between 0-4V instead of 0-5v, so you will run the pump at 20% below capacity.

The proper way is to simply measure the resistance from VR pin to ground. If it is 100k, then you can put a 100k in series from 0-10v signal to VR pin.

Or just get a controller or controller interface that outputs 5v PWM.

FWIW, I think 0-10v is old technology. Anything recent (within last 3 years) really should be 5v pwm capable. Like Archon controller. I think when looking for a new controller, that should be one of the main feature to look for.

I have a bunch of 10k pots and 100k pots laying around. I was just asking if I could use one instead of a resistor and wire it in series( turned up all the way to 10k or 100k). Probably easier to just use a resister I guess. I prefer 0-10v or 0-5v analog outputs because its usually easier for me to program. My Analog output module can do both 0-10v or 0-5v but not at the same time. My lights require 0-10v dimming so any 0-5v i need I just add pots to get it down to 5V.

[karimwassef]

You can use a pot and run 10V from the Apex and adjust the pot until the voltage to ground is 5V.

Then you can measure the pot resistance and replace it with a hard resistor.

[d0ughb0y]

I have a bunch of 10k pots and 100k pots laying around. I was just asking if I could use one instead of a resistor and wire it in series( turned up all the way to 10k or 100k). Probably easier to just use a resister I guess. I prefer 0-10v or 0-5v analog outputs because its usually easier for me to program. My Analog output module can do both 0-10v or 0-5v but not at the same time. My lights require 0-10v dimming so any 0-5v i need I just add pots to get it down to 5V.

ok, in that case you can. I misunderstood your question. You still need to measure the resistance of the VR input to gnd, then set your pot to the same value. Then test by starting with 0v analog output, slowly going to 10v to make sure you end up at around 5v at max 10v analog out.

[stage3-s4]

ok, in that case you can. I misunderstood your question. You still need to measure the resistance of the VR input to gnd, then set your pot to the same value. Then test by starting with 0v analog output, slowly going to 10v to make sure you end up at around 5v at max 10v analog out.

Yea I will certainly measure the resistance between the VR and GND when I get one. I would want to make sure its 5V max between VR and GND with the pot connected right?

How long did it take for it to arrive BTW? Sometimes Stuff I order from China takes like 4 weeks to get here.

[d0ughb0y]

epacket takes 10 days for me (counting Saturday and Sunday). You should get a tracking number. The delivery date shown on ebay is pretty accurate for me, give or take 1 day.

you can get this cable as well
http://www.ebay.com/itm/3-Pin-Male-Ends-3x0-75-SQMM-LED-Light-Strips-Waterproof-Connector-Cable-/201095471852?hash=item2ed23962ec:g:b9AAAOxyTkJSRo1h

and 5.5x2.1mm female DC power jack.

[karimwassef]

I did get those cables but I'm a little concerned that the jebao has a v-notch and this has a rectangular one.

We'll see...

[d0ughb0y]

I also just ordered the higher current green motor control board. I think there are now 3 of us that bought from the seller I linked in post #1. I hope to get the board next week so will test it by next weekend.

If I can confirm the green board works well, I'll scrap the idea of building my own board. I probably could use the time better finishing up my v2 of diy controller.

The white board module is really bare bones component. It does not even filter the back-emf signal and is fed directly to LM339, and I am amazed how the LM339 comparator is still able to output proper digital signal used for motor commutation timing.

The stock DCT/DCS controller also uses the LM339 comparator, but they add capacitors to filter out the input signal noise.

Just by looking at the picture of the green board, I can tell it is much better than the white board. You can see all the resistors and capacitors around the LM339, so it looks like it filters the input. Also this board uses the IR2101, so that means it uses all N-channel MOSFETs, which makes the upper and lower circuit essentially identical, and N-channel MOSFETs have lower Rds, so can run more current through without generating too much heat. I think the listing says can run up to 6amps without a heat sink, That is a lot of current. I think even the largest DCT pump only use 4 amps.

[stage3-s4]

I just order my board this week. I have to order a pump now.

[d0ughb0y]

I just order my board this week. I have to order a pump now.

when I ordered my pump, I mistakenly thought the DCT was newer than DCS, but DCS is actually the newer one. It's pretty much identical except for the impeller. The DCT impeller looks the same as the old DCS impeller (green color).

[stage3-s4]

when I ordered my pump, I mistakenly thought the DCT was newer than DCS, but DCS is actually the newer one. It's pretty much identical except for the impeller. The DCT impeller looks the same as the old DCS impeller (green color).

Should I order it directly from fish street or on ebay?

[d0ughb0y]

shop around.
I've never bought anything from fish-street, or any ebay US sellers based in California, because sales tax is added.

[stage3-s4]

Just got a DCS7000 with amazon prime for $92 shipped.

[karimwassef]

Ok. I got the controller and some connectors. Good news and bad news. Good news, I'll wire it up today I hope.

Bad news, major disconnect on connector (pun). The large with a red seal is from one of my old Jebao DC pumps. The large with black seal is from one of my DCT pumps, and the tiny fake connectors in the middle with the rectangle notch are the wrong ones I just got. :(

<a href="http://s1062.photobucket.com/user/karimwassef/media/22EAECA7-BCF8-472C-BD27-C924661FBD2C_zpseytk994p.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/22EAECA7-BCF8-472C-BD27-C924661FBD2C_zpseytk994p.jpg" border="0" alt=" photo 22EAECA7-BCF8-472C-BD27-C924661FBD2C_zpseytk994p.jpg"/></a>

Eh, I'm sure I'll use them for something. Anyone got a link to the right connectors?

[karimwassef]

Nice beefy FETs but I bet they need a cooling plate for full current

<a href="http://s1062.photobucket.com/user/karimwassef/media/EA63F8FE-5655-4122-92B8-55197EF72C33_zpsxc5dorp3.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/EA63F8FE-5655-4122-92B8-55197EF72C33_zpsxc5dorp3.jpg" border="0" alt=" photo EA63F8FE-5655-4122-92B8-55197EF72C33_zpsxc5dorp3.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/D4DDA1E5-199E-481C-B68F-BC4AC7223C46_zps7ftqjvvr.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/D4DDA1E5-199E-481C-B68F-BC4AC7223C46_zps7ftqjvvr.jpg" border="0" alt=" photo D4DDA1E5-199E-481C-B68F-BC4AC7223C46_zps7ftqjvvr.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/620EE52E-4BDB-48C8-827C-89B11088CFE3_zpsvgj3wanx.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/620EE52E-4BDB-48C8-827C-89B11088CFE3_zpsvgj3wanx.jpg" border="0" alt=" photo 620EE52E-4BDB-48C8-827C-89B11088CFE3_zpsvgj3wanx.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/2696C527-E811-4E4D-8264-DBA9CB9F0D92_zpsmbmk2xml.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/2696C527-E811-4E4D-8264-DBA9CB9F0D92_zpsmbmk2xml.jpg" border="0" alt=" photo 2696C527-E811-4E4D-8264-DBA9CB9F0D92_zpsmbmk2xml.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/13E09BBA-EE6C-4999-8E5B-13BE3FFA2A3A_zps4qfgmfbg.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/13E09BBA-EE6C-4999-8E5B-13BE3FFA2A3A_zps4qfgmfbg.jpg" border="0" alt=" photo 13E09BBA-EE6C-4999-8E5B-13BE3FFA2A3A_zps4qfgmfbg.jpg"/></a>

[d0ughb0y]

did you buy from the ebay link I gave above?
I got that cable and it fits fine to the original jebao cable.
The square notch is a lot smaller so it fits into the triangular opening just fine and the size of the connector itself matches.

I am getting the board today as well. I placed ebay order last monday and will get it today. It is not uncommon now for me to get items from china faster than buying from US sellers, and not to mention way cheaper.

those look like TO-220 FETs. The stock controller FETs are smaller TO-253 I think.

[karimwassef]

no. mine was more like this.

http://www.ebay.com/itm/2pin-3pin-4pin-5pin-Black-Waterproof-LED-Strip-Connector-Cable-Male-Female-22awg/371361951388

didn't realize the sizing on the listing.

[karimwassef]

ok. here's what I'm wiring up right now. I like to cover all the details just to be 100% sure it all ties out.

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsdcgx42zb.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsdcgx42zb.png" border="0" alt=" photo 0_zpsdcgx42zb.png"/></a>

any issues?

[d0ughb0y]

I measured the resistance between VR and Gnd to be about 98K, so the board must use a 100k pull down resistor. So you need to short your pins 3 and 4 and connect that to a 100k resistor to VR pin.

[d0ughb0y]

I lifted up the mosfet to see the part number, and its NCE6990.
rated 90amps with Rdson of 7.2mOhms.

So the FETs should generate very little heat at the current rating of the jebao pumps.

And they are all N channel mosfets.

http://i1193.photobucket.com/albums/aa352/jerrysy/controller/8EC906C6-D178-4893-9157-2C8ADD489A7F.jpg~original

I'm going to use 5pin JST connector for input. These have 2.5mm pitch pins.
http://i1193.photobucket.com/albums/aa352/jerrysy/controller/C671FD47-A50B-4B8E-B050-2302B40E9965.jpg~original

And screw terminal connectors for the output. These are 5mm pitch. I don't have a 3 or 5 pin terminal connector, so I have to buy some.
http://i1193.photobucket.com/albums/aa352/jerrysy/controller/9005B376-8CB1-4ECE-8E50-4F836075B0A5.jpg~original

[karimwassef]

ok. I'll use 100K. Other than that, any other changes?

Also - I don't think you meant to say that I would short pins 3 and 4, that would pull the signal to ground. I think this is what you meant:

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsqes5adjx.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsqes5adjx.png" border="0" alt=" photo 0_zpsqes5adjx.png"/></a>

[d0ughb0y]

I wasn't sure how the apex cable is wired. I just assumed by convention that since both are the same color, that they are the same line.


I checked the VR pin connections, and it goes like


VR -> 1k resistor -> 100k resistor -> Gnd
-> capacitor -> Gnd
-> VR pin of chip.

So yes, if you connect your 10v signal from apex to a 100k resistor then to the VR input, you should be fine.

If you want to be exact, you can use a trimmer and adjust till you get 5v at the VR pin while apex is sending out 10v.
You want to adjust from below 5v up to 5v and not apply >5v to the VR pin of the chip (pin 16)

[karimwassef]

Woohoo!! It works!

Ok, I cheated and used the 0-5V from an old Jebao DC controller (already had it hooked up) and recycled the connector that used to be connected to it... Crude but it works!

<a href="http://s1062.photobucket.com/user/karimwassef/media/41BAF185-0478-4C8A-A972-07A668439316_zpsepvhsf4t.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/41BAF185-0478-4C8A-A972-07A668439316_zpsepvhsf4t.jpg" border="0" alt=" photo 41BAF185-0478-4C8A-A972-07A668439316_zpsepvhsf4t.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/1CD1458B-20F9-4B24-AFB3-C4958CC8D835_zpsf17ucqwi.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/1CD1458B-20F9-4B24-AFB3-C4958CC8D835_zpsf17ucqwi.jpg" border="0" alt=" photo 1CD1458B-20F9-4B24-AFB3-C4958CC8D835_zpsf17ucqwi.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/12213C0D-E090-48BE-8059-A3C21FC28F36_zpssarpqqyt.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/12213C0D-E090-48BE-8059-A3C21FC28F36_zpssarpqqyt.jpg" border="0" alt=" photo 12213C0D-E090-48BE-8059-A3C21FC28F36_zpssarpqqyt.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/9879F374-31FF-4CCE-A058-AD287725D78C_zpsacdf5xzy.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/9879F374-31FF-4CCE-A058-AD287725D78C_zpsacdf5xzy.jpg" border="0" alt=" photo 9879F374-31FF-4CCE-A058-AD287725D78C_zpsacdf5xzy.jpg"/></a>

[karimwassef]

I'm uploading the video now. I'm letting it run overnight to make sure it's all good.

I touched the FET pads and got a mild shock, not sure why that would be since the 24VDC shouldn't be enough.

[karimwassef]

Here's the updated version for the final APEX implementation

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpscihurket.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpscihurket.png" border="0" alt=" photo 0_zpscihurket.png"/></a>

[karimwassef]

<iframe width="560" height="315" src="https://www.youtube.com/embed/b7CVC9NgkdE" frameborder="0" allowfullscreen></iframe>

[karimwassef]

Hey doughboy, I listed you as the source for this build on my YouTube vid and linked this thread. Let me know if you want me to add or change anything there.

And thanks for the effort to get this working. I've been struggling with a simple solution for a while and this $11 option ROCKS! I got two more for my other DCT15000s and I'll be using a secondary Apex VDM module for controlling them.

The one from yesterday has been working all night without issue.

I'm thinking we should test the flow compared to a native controller. Basically measure how fast it can fill up a 30gal bucket, for example. Comparing the time from the native and this DIY should be a valuable datapoint.

Not sure how to measure pressure though. Maybe use a valve setup and see if I can repeat the flow rate test with a lot more backpressure?

[d0ughb0y]

very nice!

with the pump being out for 3 years, I thought someone would have come up with a way to control this by now, but all I found was a lot of bs and misinformation and no result.

I'm pretty certain the Z/F direction pin will work in making CP-40 or Gyre run in reverse (even the DCT pumps for that matter). The Z/F pin is pulled up high with a 4.7k resistor. To run in reverse, just connect Z/F pin to gnd.

Once I get my setup working, I'll look at the signal pin and see if I can get rpm info or at least know if it can signal if a pump is spinning or not. I think the ability to detect if a pump stops is a very useful feature.

[karimwassef]

The reverse feature could be very useful with a jammed impeller. If the signal pin indicates that it's not turning, a few seconds in reverse could help unjam it.

[karimwassef]

Any readings on voltages that could explain the mild zap I got from the FET bases?

I'm going to put these in a metal box with a finned back. I'll use insulated gap pad material to heatsinks the FET pads to the box internal walls. I would prefer to use cheaper silver paste, but if there's voltage on the FET pads, I can't really do that :D

[karimwassef]

Also, any thoughts on multi-drive?

If I wanted to drive an array of pumps together, can I use one driver and multiple pumps?

If not, how about using one controller as a driver for an array of FET packs to drive the array of pumps? Think 10 pumps in phase.

[d0ughb0y]

The reverse feature could be very useful with a jammed impeller. If the signal pin indicates that it's not turning, a few seconds in reverse could help unjam it.

Once I get my setup working, then I can plan what else I can do with this. Reefers have used return pumps for years without knowing if they are running dry or stopped spinning, so using the board by itself adding speed control is already a big step as it is.

Any readings on voltages that could explain the mild zap I got from the FET bases?

I'm going to put these in a metal box with a finned back. I'll use insulated gap pad material to heatsinks the FET pads to the box internal walls. I would prefer to use cheaper silver paste, but if there's voltage on the FET pads, I can't really do that :D

not sure, maybe static electricity?
Did you try touching it again after running overnight? is it even warm?
how about run it at max speed for a few hours and touch the FETs. If you have a Killawatt, you can measure how many watts the pump is using at full speed.

[d0ughb0y]

Also, any thoughts on multi-drive?

If I wanted to drive an array of pumps together, can I use one driver and multiple pumps?

If not, how about using one controller as a driver for an array of FET packs to drive the array of pumps? Think 10 pumps in phase.

You cannot use one board to drive multiple pumps.
Not even sure if you can tie all pwm inputs of multiple boards together, as the input impedance will be affected (if you are using a voltage divider).

[karimwassef]

Did you try touching it again after running overnight? is it even warm?
how about run it at max speed for a few hours and touch the FETs. If you have a Killawatt, you can measure how many watts the pump is using at full speed.

It's been running at 100% for over a day now. I used a plastic sheet and touched the FETs... warm, but not hot.

I'll measure the current tonight. I'll also make the adaptation to the Apex drive.

[karimwassef]

Not even sure if you can tie all pwm inputs of multiple boards together, as the input impedance will be affected (if you are using a voltage divider).

No, I meant use the A B C outputs to drive banks of A FETs, B FETs, C FETs... basically create a multiplier effect with multiple switches for multiple pumps


A to drive the gate of 4 FETs -> A1 A2 A3 A4
B to drive the gate of 4 FETs -> B1 B2 B3 B4
C to drive the gate of 4 FETs -> C1 C2 C3 C4

A1, B1, C1 go to pump 1
A2, B2, C2 go to pump 2
A3, B3, C3 go to pump 3
A4, B4, C4 go to pump 4

So one output can be used with a bank of 3x FETs to drive x pumps.

So 10 pumps would require 30 FETs that are being turned on and off by this controller board.

So $12 x 10 = $120 (conventional, but need 10 signal sources)
or $12 x 1 + $2 x 30 = $72 (assuming $2/FET assembled with board and components, can use one signal source for all).

[d0ughb0y]

The WVU outputs are fed back to the control chip as back emf signal to control commutation timing, so it would not be possible to use one controller to drive multiple pumps.

$120 would just buy you 1 icecap interface.

Considering the JCT15000 is probably running at 4amps, warm is a good sign. Once you get the killawatt reading, then we know for sure how many amps the pump is using.

[karimwassef]

Ok. So I can't go cheap. One controller per pump...

How many controllers do you think we can drive with a single Apex signal?

[d0ughb0y]

Ok. So I can't go cheap. One controller per pump...

How many controllers do you think we can drive with a single Apex signal?



The only way to find out is to test. Worst case you can add a buffer chip (maybe 74HCT245) with all inputs connected to apex signal and each output connects to a board.

[karimwassef]

good point. I think there'a a limit of 20mA on the V1/V2 outlets though. Exceeding that could damage the Apex. That's what I remember from Russ a couple of years ago.

[karimwassef]

ok. Here's the output current and power as a function of the signal VDC (0-5VDC)

<a href="http://s1062.photobucket.com/user/karimwassef/media/1_zps9lpwm4nj.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/1_zps9lpwm4nj.png" border="0" alt=" photo 1_zps9lpwm4nj.png"/></a>

[karimwassef]

The supply voltage was 24.6VDC consistently.

After running for a couple of days, the FETs are "painfully hot", but not immediately burning. :)

So I'm guessing 80C? I think a metal plate and little fan would be a good idea for multiple controllers in a box.

[karimwassef]

Another question... On another thread, a CP pump user said that the makers of the wireless controller are pushing to 120% output (more than the native Jebao controllers). That suggests that they're boosting the voltage input to 29VDC or more and the motors can handle it...

So now, I'm curious about "overdrive". I ask because I plan on running these pumps in cyclical mode between 0% and 100% over about a minute period. In fact, I'm going to phase them by 180 degrees too (two of them) or 120 degrees (if using three). That implies that it may be possible to overdrive at the peak since the pumps won't be running at that level constantly.

So how do we test if we can push 20% or more out??

[d0ughb0y]

I would not trust any claim made by fish-street. They have been known to stretch the truth a lot. (they once claimed they are selling a newly designed power supply they designed themselves, but it is actually your generic cheap laptop power supply, like the ones that comes with JCT pumps now) That probably means the stock controller is not capable of running 100% and the new device they are selling can, in order to promote the sales. The other possibility is, I know the stock controller uses P-Channel N-channel mosfet pairs. This is not as efficient as using N-channel mosfet pairs. So perhaps the T1 is using N-channel mosfets only. The green board is using all n-channel mosfets.

All you need to do to maximize this is make sure pin 16 of the motor controller chip gets the max 5v. You can try putting on voltage higher than 24v, but I don't think that is what the fish-street controller does, as you still use your stock 24v power supply. I doubt there is any voltage boost Applying a higher voltage may risk burning out the motor coils.

If it is getting hot, then using a heat sink makes sense. I know I found a heat sink on ebay specifically designed for these boards, but can't seem to find it now. I should have bookmarked it.

[karimwassef]

Since I'm doing multiple drives for multiple pumps, I'll use a larger plate, U tubes and a fan.

[d0ughb0y]

I was trying to find the listing for the heat sink, and it was just a sample picture of the board attached to a heat sink and not selling the actual heat sink.

and the other listing for the same board says to use a heatsink if load is > 60watts. So I incorrectly said 6A. So it makes sense now since your graph shows you are going over 60 watts, that you need to use a heat sink.

[d0ughb0y]

Can you test if the pump still operates if you power it with 12v?
I think it still should, with PWM at least 50%, or maybe closer to 100%.
The board is rated 12-36v.
The stock controller only works on 24v, so it cannot be used with 12v backup power.

I just found UPS battery on sale at a local store, like $12 for 7AH 12v battery,so I might get some and use it as backup power.

Bad news for me, I think the green board I got is defective. I will contact the ebay seller.

[karimwassef]

:( sorry to hear that. I just ordered three more... Hope my odds are good.

We do need to exercise the controllers to make sure they're robust.

I'll try the 12V.

I'm running a 24V backup battery, by the way, so 12V isn't necessary.

[d0ughb0y]

did you diy your backup battery?

I am looking for a simple charging switch over circuit.

[karimwassef]

No. I found a really cheap charge controller on amazon. $11

https://www.amazon.com/gp/product/B00L37KZI6

https://images-na.ssl-images-amazon.com/images/I/71sKqn6Q%2BpL._SL1000_.jpg

It's originally intended for solar panel charging, but it doesn't matter.

[d0ughb0y]

Good idea. I was just looking under lead acid battery chargers.
Does this power the load if there is power from solar input and charge your battery at the same time, then automatically use the battery to power the load if there is no input from solar?

Does it also know to shutdown completely if battery voltage goes down to say 10v? (or 20v if you are using 24v). As you do not want to completely drain your battery.

I tried searching online for a manual and can't find one, but based on the description, it looks like it will shut off the load at 10.8v. This is perfect.

[karimwassef]

I haven't tested it under all different conditions.

I wouldn't use a normal lead acid battery though. Also, the energy storage/$ is important to me since I would like to survive for several hours on a charge.

A deep discharge marine battery is really best here. This is the one with the best value IMO

https://www.amazon.com/gp/product/B00S1RT58C

You can stack them two deep to get to 24V.

This charger is limited to 20A though, so that's not much power throughput IMO. I was looking for a higher current bypass so the 20A only limits the charging and not the normal running current.

This is a big topic for me - you might want to start its own thread :D so this one can stay focused on the motor drive.

[karimwassef]

As an aside - I don't use the cheap 12V or 24V power adapters if at all possible. They're horrendously expensive and very poor quality.

I prefer refurbished power supplies, but you need to do some work to get them up and running.

I wouldn't recommend the same for everyone, but this is the best solution I've found for real DC power. I use it for 12V and 24V grids

http://www.rchelination.com/setting-hp-dps-1200fb-power-supply/

[d0ughb0y]

I don't need anything that heavy duty.
That battery is a bargain compared to the icecap and vortech battery.

[karimwassef]

Yup. Two in series is 24V at 100A-hr so it would run 2400W-hrs (assuming slow discharge so the discharge penalty is minimal). So a 100W pump (or two running at 50W) can run for a day. That's enough time to get a generator primed and gased in case of a longer gap... Or it would ride through most normal outages.

[d0ughb0y]

ok, good news, my pump is now running using the green board. It was a user error. Somehow I programmed my arduino to output 25% pwm, since I don't want it to splash water out of the bucket all over the room if I start out at 100%. After I changed it to gradually increase to 100%, the pump is now running!

I tried to Gnd the Z/F direction pin, and the pump spins. I can't tell though if it is actually spinning the reverse direction or not, but the water still comes out of the top, but definitely not as strong as the forward direction. So maybe the impeller is actually spinning the opposite direction. In any case, I still believe this will work with crossflow and gyre.

I checked the signal pin, and the signal is the same as what I saw on the white board. If the pump is running, there is a signal that seems random, but probably is synchronized with each commutation. One thing for sure though, once the pump stops, the signal either stays low or stays high. So I think that can be used to detect if the pump is running or not.

[d0ughb0y]

And I just tested on 12v, the pump did not work.

[karimwassef]

Mine runs but it's very weak. Let me measure the current

[d0ughb0y]

I will try again later. I can't find another 12v power supply so I used a splitter.
The green board specs says it works with 12v. I think the pre-drivers requires 10v bias, so I think this should work with 12v. So the only reason this won't work is if the pump itself does not want to work with 12v.

[karimwassef]

Hmmm.. Ok, the results require some interpretation... I'll share the data and leave the interpretation for later.

So, first, I set up the measurement for the 24VDC first. Remember that I was seeing about 70W a couple of days ago (~3A)? Now it's only drawing ~1.8A at the same setpoint (signal = 4.5VDC). Here's a few details:

1. The pump has been running continuously for 2+ days.
2. When I first tested the pump, it hadn't been running for a couple of months after a deep cleaning. It may have had a lot of resistance to turning?

Ok. So that establishes the new baseline of 45W (at least for my old recently cleaned pump) at a signal of 4.5VDC. I have a new pump too (out of the box) - so I'll rerun the tests on that one too.
<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_6255_zpsr3flmpn2.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_6255_zpsr3flmpn2.jpg" border="0" alt=" photo IMG_6255_zpsr3flmpn2.jpg"/></a>

Then I ran the same setup on my 12VDC. A little detail on my 12VDC source: it's capable of driving up to 800W, so its current capability is exceptionally high. First, when I plugged it in, the current went negative ~-0.3A. The pump was actually running but with a negative current? Very weak, but it was running at about 3W. I realize that the wire is covering up the negative sign before 0.28 in the picture, but it's there if you look closely.
<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_6252_zpsv66lf8fu.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_6252_zpsv66lf8fu.jpg" border="0" alt=" photo IMG_6252_zpsv66lf8fu.jpg"/></a>

I decided to disconnect the other loads on my 12VDC source, and after a few seconds, the current reversed and went positive until it got to 0.75A. That's about 9W (so 20% of the 24VDC state). The flow is significantly lower than the 24VDC too.
<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_6257_zpsbeqflzma.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_6257_zpsbeqflzma.jpg" border="0" alt=" photo IMG_6257_zpsbeqflzma.jpg"/></a>

That's the data - interpretation?

[karimwassef]

I made a video if you guys think it's worth posting to compare the flow.

[d0ughb0y]

Is your meter set to measure DC current? I take it that it is set correctly.
Your current connection may be reversed. I think the sign should not matter.

I think you need to rewire your Amp meter connection.
The power supply negative goes direct to green board negative.
The power supply positive, connects to meter + then meter - to green board +.

[karimwassef]

I'm making my measurement on the ground (-ve), not the +ve line.

The normal 24VDC measurement works fine.

The current flips without changing the connections, so it's a real flip. I repeated it several times.

It is on the 20ADC.

[karimwassef]

<iframe width="560" height="315" src="https://www.youtube.com/embed/NQrE0Moqhw8" frameborder="0" allowfullscreen></iframe>

[d0ughb0y]

I think we can say it is not feasible to run the pump at 12v.

[karimwassef]

I agree. While it turns, it doesn't do much for flow.

Still makes me wonder about running higher though in a cyclical mode. Just not sure how high would be safe.

[karimwassef]

One more tidbit on the current measurement. I'm using modified meter wiring to reduce the resistance in the current loop. Last time, I used the standard meter wire probes. I think that could contribute to the higher current reading.

[d0ughb0y]

Can you measure power and current using stock controller?

BTW to test running at 100%, simply connect the VR pin to 5V pin with a jumper wire.

[karimwassef]

I thought of that, but I wanted to characterize over different setpoint voltages 0-5V

I'll collect a lot more content this weekend

[d0ughb0y]

I was thinking more of how the max setting between the 2 compares, since you mentioned the fish-street T1 can drive the pump 120% (120% more flow than stock controller that is).
As these are BLDC motors, the amount of flow is proportional to the rpm and is proportional to power usage. And if you can get a hold of a T1, you can measure and verify if the claim is true or not.

[karimwassef]

good point. I need to find someone local with a T1 since I'm not spending that much for a test :D

[karimwassef]

ok so here are the actions/tests for this weekend:

1. set up the Apex connection on a VDM so I can run two pumps at the same time.
2. measure the current on a brand new jebao controller and measure the associated flow for each of the setpoints. (new/new)
3. measure the current and flow on the new jebao controller with the old pump at the same setpoints. (new/old)
4. measure the current and flow on the Apex/driver with the new pump for different 0-5V setpoints. (apex/new)
5. measure the current and flow on the Apex/driver with the old pump for different 0-5V setpoints. (apex/old)

Then I'm going to create a phased flow with the two pumps out of phase and run them off one power supply... that's my actual application :D

I think that'll cover all the questions for now?

for the Apex-driven voltages, I'm thinking:

0.00
0.50
1.00
1.50
2.00
2.50
2.75
3.00
3.25
3.50
3.75
4.00
4.25
4.50
4.75
5.00

I suspect that the 2.75 to 5.00 (10 settings would line up with the native controller's setpoints). I could open it up and just measure it, but I'd rather not.

That would correspond to the Apex running

0.0
1.0
2.0
3.0
4.0
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
9.0
9.5
10.0

[d0ughb0y]

I used a killawatt and compared the power usage of the green board to the stock controller. Green board uses 25 watts, stock controller uses 35 watts. There may be difference in the flow, but it is hard to quantify. The stock controller with all 10 leds on will probably pull a few more watts, but I was not expecting 10watt difference.

Maybe when you do the comparison, you can test say time to fill up a 5 gallon bucket in order to quantify any difference in flow.

[karimwassef]

that's exactly how I was going to test flow. GMTA man!

[d0ughb0y]

I connected the white motor controller, and it uses 35 watts as well. I will try to resolder the connecting wires to the green board to see if it makes any difference.

[karimwassef]

ok. Here's the first batch of results. Either the Jebao native controller is a power hog, or it does create more drive. Tomorrow, I'll try and get the green controller wired to my Apex to test.

So - first the setup:

<a href="http://s1062.photobucket.com/user/karimwassef/media/C3F9716E-7441-4B25-83F9-FAB6076A051F_zpspphphjvk.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/C3F9716E-7441-4B25-83F9-FAB6076A051F_zpspphphjvk.jpg" border="0" alt=" photo C3F9716E-7441-4B25-83F9-FAB6076A051F_zpspphphjvk.jpg"/></a>

basically, I let the run pump in a closed loop to get to full speed and stabilize the current consumption. Then I use a sliding gate valve to suddenly shut that path off and simultaneously push the timer on the stopwatch. When the water level in the 5 gal bucket crosses the 5 gal mark, I stop the watch and take the reading. The container around it is because I didn't want to be focused on reopening the gate valve. It's secondary to catching the stopwatch reading so I needed some protection in case the bucket overflowed.

Then reset and repeat...

Here's an example reading

<a href="http://s1062.photobucket.com/user/karimwassef/media/69B848E9-CC6F-41F5-84B6-ECC307730B78_zpsllt99urh.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/69B848E9-CC6F-41F5-84B6-ECC307730B78_zpsllt99urh.jpg" border="0" alt=" photo 69B848E9-CC6F-41F5-84B6-ECC307730B78_zpsllt99urh.jpg"/></a>

and the results

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsunlxunql.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsunlxunql.png" border="0" alt=" photo 0_zpsunlxunql.png"/></a>

So there's a little drift in the 1-10 relating to power drive, but overall, the power and flow line up. The old pump just seems to have more power for the same set point.

I'm sure there's some pump to pump variability.

[d0ughb0y]

that is all using the green controller board right?
what does the 15000 pump specs say for flow at 3' or about 1meter?
If you test with stock controller, you'll be able to confirm if the pump is over rated or not.
your power reading seems to match the specs 105watts.

[karimwassef]

No. This is with the native Jebao controller

[karimwassef]

http://www.fastlight.co.uk/acatalog/DCT%20Graph%20510.jpg

So 1m ~ 11,500 liters per hour ~ 3000 gph

With the Jebao controller on 10, I peaked at ~2600 gph. That's off by 13% from the advertised rating but it may be limited by the controller

[d0ughb0y]

I think that flow is fine compared to the chart.

I'm curious to see if the green board generates a comparable amount of flow. I think from your chart, it uses 70-80 watts at max.

[karimwassef]

We'll find out tonight. I set up the new VDM module. Still need to make the cables.

I think the difference in flow is ok given that my flow measurement error increases at high flow rate because 5gal is a very small volume at 2600gph, so fractions of a second errors are more impactful.

[karimwassef]

Btw, measuring flow is a PITA! But I'll do it for the common good.

:D

[Dactylopterus]

I think the pump flow is meassured without any back pressure other than the height. Your setup has two elbows creating a lot of back pressure. Your measurements show that the flow of these pumps may be rated correctly.

[karimwassef]

Got the Apex control working and able to provide 0-5VDC (0-100 APEX signal).

Unfortunately, at APEX=100 (drive signal = 5VDC), the current draw is only 2.8A @ 24.3V and I can tell that it's not as strong as the native Jebao controller.

I'm going to try using the 5VDC output on the controller itself to see if I get anything different.

[karimwassef]

running off the internal 5VDC only increases the current draw slightly up to ~3A (maybe 3.07A max) . I checked the signal and it's 5.00V.

It's a far cry from the native controller's 4.57A for the same pump.

Going to baseline the native controller again just to see if anything's changed.

[karimwassef]

Jebao controller = 4.57A
Green controller = 3.07A

same conditions...

[seldon]

Have you tried measuring the current draw at the pump rather that at the wall? That way you can take out any controller inefficiencies.

[karimwassef]

The pump power is 3 phase square pulses. I would need to capture all 6 voltage and current waveforms and determine the RMS power.

[karimwassef]

I'm thinking of what to try next.

I can try the other pump, but their behavior was similar.

I can dig up my data acquisition USB board and wire it up to compare the pump waveforms.

I have two more green boards on the way... I can try them out.

I can add a plate and fan in case any component is thermally limiting output?

I can add a ferrite bead at the DC input to experimentally limit any noise going back to the power supply and limiting it?

[karimwassef]

Doughboy- ideas? There aren't many other control points that I see?

[d0ughb0y]

The only true way to compare is by flow.

On my setup, the stock and white controller runs the same 35 watts, and green 25 watts.

The stock and white controller both use transistor + N and P channel mosfet pairs, and the green uses IR2101 pre driver + N channel mosfet pairs. So it made sense that the white and stock controller showed the same power usage. Both white and green board are based on the same JY01 motor controller chip.

Both circuits are pretty standard and every article I read about BLDC driver uses one or the other circuit and I frequently see mention that the N channel pair circuit is better.

On my small pump, I can't tell the difference in flow. I take it on the bigger pump you are able to tell without doing a flow test right?

If that is the case, I can analyze the green board circuit further to see if any mod can be done.

FWIW, I think we both see about 25% less power on the green board from stock.

[karimwassef]

I didn't measure the flow because I could immediately tell that it was significantly lower.

I can run the test but I don't think it's worth it unless there's a path to getting more power into the pump

[d0ughb0y]

no need to run the test if you can tell right away.

I'll check if there is anything that can be done to improve it.
otherwise, I can still go back to my original plan to build a new board from scratch.

[karimwassef]

Or try another board vendor? If we're looking at a new board that would basically be a variant of this one, can we debug it and improve it's performance instead?

[karimwassef]

If it's generating less power drive, is it a frequency mismatch maybe? Then a fix could be changing a timing resistor/cap?

What are the other control variables?

The FETs can handle it, so... there's got to be a path forward.

[d0ughb0y]

the board I got is marked JYQD_V6.3E1
and I can't find any other version for this board.
I know there is a version of this board for motors with hall sensors and that won't work with the jebao pumps. The LM339 chip is missing from the board.

The board is advertised as used in electric scooters and bikes.

The speed is proportional to voltage/pwm. I checked the commutation signal with pwm connected to 5v, and I still see pwm like signal. I think that is reducing the output voltage. I'll check the BEMF circuit, maybe later this week, to see if I can find anything.

[d0ughb0y]

The standard driver circuit, as per datasheet looks like this

http://circuits.datasheetdir.com/65/IR2101-circuits.jpg

I found an article with a slightly modified circuit (but using IR2110).

http://4.bp.blogspot.com/-0oqvn95WCf8/UPvndlg4kfI/AAAAAAAAAZw/pz0uShYp3qw/s640/IR2110+-+2.png

It should not be hard to solder a 1k resistor between the Gate and Source of all 6 MOSFETS. Adding the diode might be a little bit trickier as you need to solder to SOP-8 smd pin. I will try adding the 1k resistor. I won't be able to do it till later this week.

[karimwassef]

the board I got is marked JYQD_V6.3E1
and I can't find any other version for this board.
I know there is a version of this board for motors with hall sensors and that won't work with the jebao pumps. The LM339 chip is missing from the board.

The board is advertised as used in electric scooters and bikes.

The speed is proportional to voltage/pwm. I checked the commutation signal with pwm connected to 5v, and I still see pwm like signal. I think that is reducing the output voltage. I'll check the BEMF circuit, maybe later this week, to see if I can find anything.

Mine has the same marking

[karimwassef]

I wanted to document the Apex connections and clean up the overall design outside the board

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsye5ajie2.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsye5ajie2.png" border="0" alt=" photo 0_zpsye5ajie2.png"/></a>

with color coding to make it easier for others to do the same.

My VDM was outputting 0-10.8VDC so I needed to modify the resistor divider network to get exactly 0-5VDC (as close as I could). I checked it with two different multimeters and it's real.

Here is the resulting APEX programmed input (0-100) and measured VDC

<a href="http://s1062.photobucket.com/user/karimwassef/media/1_zpss5qyrcua.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/1_zpss5qyrcua.png" border="0" alt=" photo 1_zpss5qyrcua.png"/></a>

Here are some pics of the connections

<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_6317_zpswgdrdwda.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_6317_zpswgdrdwda.jpg" border="0" alt=" photo IMG_6317_zpswgdrdwda.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_6320_zpszlvm9xtn.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_6320_zpszlvm9xtn.jpg" border="0" alt=" photo IMG_6320_zpszlvm9xtn.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_6321_zps1vsnon1i.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_6321_zps1vsnon1i.jpg" border="0" alt=" photo IMG_6321_zps1vsnon1i.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_6322_zpsatw0kh4f.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_6322_zpsatw0kh4f.jpg" border="0" alt=" photo IMG_6322_zpsatw0kh4f.jpg"/></a>

[karimwassef]

The standard driver circuit, as per datasheet looks like this

http://circuits.datasheetdir.com/65/IR2101-circuits.jpg

I found an article with a slightly modified circuit (but using IR2110).

http://4.bp.blogspot.com/-0oqvn95WCf8/UPvndlg4kfI/AAAAAAAAAZw/pz0uShYp3qw/s640/IR2110+-+2.png

It should not be hard to solder a 1k resistor between the Gate and Source of all 6 MOSFETS. Adding the diode might be a little bit trickier as you need to solder to SOP-8 smd pin. I will try adding the 1k resistor. I won't be able to do it till later this week.

Is the idea that the 1K will improve the commutation timing to get more throughput power?

[karimwassef]

I think I'll add a small wire connection to the gate first and then wire the 1K to that.

Eh - it's a $12 board and there's two more on the way. If it doesn't work, at least we'll know.

[karimwassef]

Here's a closer look at the Ethernet resistor network

<a href="http://s1062.photobucket.com/user/karimwassef/media/karimwassef001/0_zps4gjnlhly.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/karimwassef001/0_zps4gjnlhly.png" border="0" alt=" photo 0_zps4gjnlhly.png"/></a>

If I could go back, I'd get a 1K-10K pot instead of measuring and stringing resistors.

[d0ughb0y]

I think I'll add a small wire connection to the gate first and then wire the 1K to that.

Eh - it's a $12 board and there's two more on the way. If it doesn't work, at least we'll know.

All you need to do is solder 1k to pins 1 and 3 of each MOSFET. You can do it either on top side or bottom side. The resistor is to prevent surge current. The diode is for faster discharge of the FET capacitance, or something like that. In the above circuit, the mosfet gate resistor is 10 ohms, the one on the green board is 510ohms. Perhaps the value may be too high, so I think the diode could make a difference.

I think that's about all you can do with the driver circuit.

I have to check the back emf feedback circuit.

[karimwassef]

ok. based on the datasheet:

http://www.maxim4u.com/download.php?id=2054467&pdfid=EE6580B32C714D45F34EE77B83D2F7B9&file=0528\nce6990_8192231.pdf

I'm going to connect gate to source with 1K:

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsnebrj6xt.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsnebrj6xt.png" border="0" alt=" photo 0_zpsnebrj6xt.png"/></a>

right?

[karimwassef]

Here's a closer look at the Ethernet resistor network

<a href="http://s1062.photobucket.com/user/karimwassef/media/karimwassef001/0_zps4gjnlhly.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/karimwassef001/0_zps4gjnlhly.png" border="0" alt=" photo 0_zps4gjnlhly.png"/></a>

If I could go back, I'd get a 1K-10K pot instead of measuring and stringing resistors.

The image got corrupted. Here it is again:

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zps4gjnlhly.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zps4gjnlhly.png" border="0" alt=" photo 0_zps4gjnlhly.png"/></a>

[karimwassef]

would this controller not work without the hall sensors?

http://www.ebay.com/itm/NEWEST-DC-12-36V-500W-Brushless-Motor-Controller-Hall-Balanced-Car-Driver-Board/231954428640

or just go beefier?

http://www.ebay.com/itm/12V-24v-30A-brushless-DC-motor-controller-power-high-current-motor-driver-board-/111954645718

[stage3-s4]

I'm in the process of setting up a flow meter to measure flow. Just waiting on my signal converter to come in. Once I have that, I can test flow rates fairly easily.

[d0ughb0y]

would this controller not work without the hall sensors?

http://www.ebay.com/itm/NEWEST-DC-12-36V-500W-Brushless-Motor-Controller-Hall-Balanced-Car-Driver-Board/231954428640

or just go beefier?

http://www.ebay.com/itm/12V-24v-30A-brushless-DC-motor-controller-power-high-current-motor-driver-board-/111954645718

The first one requires hall sensor, so it won't work with jebao pump.

the second one is rated 18v max only.

I'm thinking another thing to try is to replace the mosfet. The specs of the mosfet used on the board looks fine, but maybe the quality may not be good.

I am also going to capture the 6 motor drive signals for the green board and compare it to the data I captured from the stock controller just to make sure the phase is similar.

I think it is just a matter of component value choice to get this working.

regarding the addition of the diode to help discharge FET capacitance. I just check the specs again, and the input capacitance of the NE6990 is 2 to 3 times higher than a couple other MOSFETs I checked. So perhaps either adding the diode or changing the mosfet may improve the power.

[karimwassef]

ok. based on the datasheet:

http://www.maxim4u.com/download.php?id=2054467&pdfid=EE6580B32C714D45F34EE77B83D2F7B9&file=0528\nce6990_8192231.pdf

I'm going to connect gate to source with 1K:

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsnebrj6xt.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsnebrj6xt.png" border="0" alt=" photo 0_zpsnebrj6xt.png"/></a>

right?

Here's the missing image

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsrjd6cvue.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsrjd6cvue.png" border="0" alt=" photo 0_zpsrjd6cvue.png"/></a>

[karimwassef]

I'm in the process of setting up a flow meter to measure flow. Just waiting on my signal converter to come in. Once I have that, I can test flow rates fairly easily.

Which flow meter? Where did you get it from?

[stage3-s4]

https://www.amazon.com/DIGITEN-Effect-Sensor-Flowmeter-Counter/dp/B00VKAT7EE/ref=sr_1_5?s=industrial&ie=UTF8&qid=1467041469&sr=1-5&keywords=flow+meter

People claim that this G thread or BSP thread wouldn't work with NPT fittings. That was my only concern, but I got it over the weekend and it fits perfectly with a 1" female NPT fitting. Put some teflon tape on it and it will seal just fine.

Anyway, its just a simple hall effect senor that outputs a PWM singal. I need the signal conditioner to converter it to a 4-20mA signal to input into my PLC.

[stage3-s4]

The first one requires hall sensor, so it won't work with jebao pump.

the second one is rated 18v max only.

I'm thinking another thing to try is to replace the mosfet. The specs of the mosfet used on the board looks fine, but maybe the quality may not be good.

I am also going to capture the 6 motor drive signals for the green board and compare it to the data I captured from the stock controller just to make sure the phase is similar.

I think it is just a matter of component value choice to get this working.

regarding the addition of the diode to help discharge FET capacitance. I just check the specs again, and the input capacitance of the NE6990 is 2 to 3 times higher than a couple other MOSFETs I checked. So perhaps either adding the diode or changing the mosfet may improve the power.

I'm with you on this. I think it may just be that the components used are not the best quality. I have found this to be the case on lots of electronic parts coming in from china.

[d0ughb0y]

https://www.amazon.com/DIGITEN-Effect-Sensor-Flowmeter-Counter/dp/B00VKAT7EE/ref=sr_1_5?s=industrial&ie=UTF8&qid=1467041469&sr=1-5&keywords=flow+meter

People claim that this G thread or BSP thread wouldn't work with NPT fittings. That was my only concern, but I got it over the weekend and it fits perfectly with a 1" female NPT fitting. Put some teflon tape on it and it will seal just fine.

Anyway, its just a simple hall effect senor that outputs a PWM singal. I need the signal conditioner to converter it to a 4-20mA signal to input into my PLC.



I take it this does not have any metal part in contact with water right? This would be a nice addition to a controller.

[stage3-s4]

I take it this does not have any metal part in contact with water right? This would be a nice addition to a controller.

Its pretty much all plastic. I haven't taken apart the entire thing yet to verify. The shaft for the wheel inside could be stainless steal but its so small I'm not worried about it. You could easily add something like this to your controller. I think its great to have with a variable pump. You could start to program PID control loops and stuff if you really wanted to get crazy.

I went to Reefapaloza this weekend and of course neptune was there displaying the new flow meter that they are coming out with. Its almost identical to this one. I asked when I could buy one and no one could confirm a date or price yet. Why wait for theirs?

[karimwassef]

Ok. I got the 2" version. I'll need to get my arduino to count pulses. The flow of the 1" is too low. The max for the 2" is 2700gph and that's close to what the DCT15000 runs at

[karimwassef]

I take it this does not have any metal part in contact with water right? This would be a nice addition to a controller.

Yup!! Going to experiment. I'm thinking this is key for critical loops like main circulation and then maybe to capture flow in test setups (like testing my surge!).

[stage3-s4]

Ok. I got the 2" version. I'll need to get my arduino to count pulses. The flow of the 1" is too low. The max for the 2" is 2700gph and that's close to what the DCT15000 runs at

Yea I figured the 1" may not flow enough. I'm never going to need more than 900GPH on my 40 gallon tank so the 1" was fine. This thing will deff have a pretty significant pressure drop though. It seems very restrictive.

They also make a digital display if you just want to use it to read flow rates directly without a controller.

[d0ughb0y]

I can see why the flow can be restrictive.
I found this picture of what the inside looks like. Looks like it has a metal shaft.

http://301o583r8shhildde3s0vcnh.wpengine.netdna-cdn.com/wp-content/uploads/2014/05/P2231345_011.jpg

At higher flow rate, it might be negligible, but I imagine this will be another thing you have to take out and clean periodically. I don't know if it is feasible to use this permanently installed. Maybe it is, only time will tell.

I just cleaned my return pump yesterday (so I can figure how to replace the ehiem pump with the jebao pump) and the valve above the union is completely stuck, and when I put my finger to see why, it feels like a lot of calcium build up. So the flow meter might get that as well over time, and may need to be removed for a vinegar bath. If you are able to read the flow, then you will probably know when its time to clean it. And you are right the stainless steel shaft should be fine. The MJ pumps have been used in saltwater for a long time and these have steel shafts. I think my MJ pump has been in saltwater for over 5 years now.

[karimwassef]

Looks like a typical powerhead cleaning job. They all have SS shafts. Just need two unions and a standby fitting that's the same length during maintenance :D

I may not put it on my main flow, but on the aux I keep in place in case the two primary pumps fail.

I have three DCT15000s. I prefer to buy cheap and add in redundancy. So, even if one pump fails while I'm traveling, the likelihood of all flow stopping is minimal. Same with pH probes, surge loop actuators, etc...

I still have a few single points of failure but I'm slowly adding massive redundancy where I can.

Still haven't figured out how to do that with my Apex. I have two of them, but I have to physically disconnect and reconnect the other in case of failure... Redundancy is the only path to cost effective quality and uptime IMO

but that's probably an aside for your other thread.

[karimwassef]

So .. Last check: is this correct?

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsrjd6cvue.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsrjd6cvue.png" border="0" alt=" photo 0_zpsrjd6cvue.png"/></a>

[d0ughb0y]

yes that is correct.
you are quite handy with making all those illustrations. :)

right. I agree. I just edited my post.
also agree that this would be similar to what apex is offering at a fraction of the cost.

btw, I calculated the RC time constant based on specs of the mosfet, 3400pf and 510ohm, is about 1.8us and will work with freq up to 90 khz. The motor controller chip pwm is 20khz. I have to look at the signal in an oscilloscope to see if the capacitance is really affecting the output voltage or not. Most schematics for this version of driver I found use 10ohms resistor and mosfet with 1300pf input capacitance, so the RC time constant difference is quite significant.

[stage3-s4]

I'm going to try running this flow meter on my main return of pump of my tank. If it ends up getting clogged up and not function in a 1 year or 2 I can just replace it for $10.

I'm curious to see any upgrades you will have for the green motor board. Hopefully something easy will turn up.

[d0ughb0y]

If I can figure out how to use the signal pin to get the RPM info out of it, it can be another way of measuring flow. the flow should be directly proportional to the rpm. this will eliminate the need for a mechanical flow meter.

Karim, can you post front and back picture of the stock controller board? I want to see if they use TO-220 or the smaller TO-252 mosfets like the stock 50watt controller.

[stage3-s4]

If I can figure out how to use the signal pin to get the RPM info out of it, it can be another way of measuring flow. the flow should be directly proportional to the rpm. this will eliminate the need for a mechanical flow meter.

Karim, can you post front and back picture of the stock controller board? I want to see if they use TO-220 or the smaller TO-252 mosfets like the stock 50watt controller.

I would be for that as well. You would obviously have to calibrate the signal to amount of flow which may be easier said than done.

[d0ughb0y]

this is the commutation signal from stock jebao controller at the lowest setting (70% duty cycle)

6 commutation = 1 revolution and takes 80.56ms. So that's 60000/80.56=745 rpm
I got to say, the signal from stock jebao controller board is clean as a whistle. The one from ebay board looks like completely random signal. This is one reason I am still considering building a controller from scratch.

http://i1193.photobucket.com/albums/aa352/jerrysy/controller/jebaotiming.jpg~original

[stage3-s4]

this is the commutation signal from stock jebao controller at the lowest setting (70% duty cycle)

6 commutation = 1 revolution and takes 80.56ms. So that's 60000/80.56=745 rpm
I got to say, the signal from stock jebao controller board is clean as a whistle. The one from ebay board looks like completely random signal. This is one reason I am still considering building a controller from scratch.

http://i1193.photobucket.com/albums/aa352/jerrysy/controller/jebaotiming.jpg

Do you have a screenshot of the ebay controller?

[d0ughb0y]

I did not save the data for the ebay controller. for the stock controller, I soldered the test points, so I got a clean reading.

I got lazy on the ebay board and just used crappy ic test clips and the signals were not as clean. I have to solder wires to the test point and capture the data again.

[karimwassef]

If I can figure out how to use the signal pin to get the RPM info out of it, it can be another way of measuring flow. the flow should be directly proportional to the rpm. this will eliminate the need for a mechanical flow meter.

Now that is an EXCELLENT idea.

Karim, can you post front and back picture of the stock controller board? I want to see if they use TO-220 or the smaller TO-252 mosfets like the stock 50watt controller.

ok. I really didn't want to open it, but for science and the common good.. I will.

[karimwassef]

I would be for that as well. You would obviously have to calibrate the signal to amount of flow which may be easier said than done.

Actually, the flow meter you found would make it easy to calibrate the pumps.. :D

[karimwassef]

I did not save the data for the ebay controller. for the stock controller, I soldered the test points, so I got a clean reading.

I got lazy on the ebay board and just used crappy ic test clips and the signals were not as clean. I have to solder wires to the test point and capture the data again.

yes. the connections can make a big difference.

[stage3-s4]

Actually, the flow meter you found would make it easy to calibrate the pumps.. :D

Oh yea true. I guess you could use it just for that and then remove it once its calibrated.

[karimwassef]

<a href="http://s1062.photobucket.com/user/karimwassef/media/DC6E8806-9DDF-46FC-ACC5-8D9A0A0BFE74_zpsdijdqmcy.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/DC6E8806-9DDF-46FC-ACC5-8D9A0A0BFE74_zpsdijdqmcy.jpg" border="0" alt=" photo DC6E8806-9DDF-46FC-ACC5-8D9A0A0BFE74_zpsdijdqmcy.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/85E6EBF1-7657-4A5B-BA1E-FFB6691248B0_zps5ghxnlrp.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/85E6EBF1-7657-4A5B-BA1E-FFB6691248B0_zps5ghxnlrp.jpg" border="0" alt=" photo 85E6EBF1-7657-4A5B-BA1E-FFB6691248B0_zps5ghxnlrp.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/D7849EB9-B112-44B3-B5D5-CD4B07E988BD_zpszhvf9tvl.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/D7849EB9-B112-44B3-B5D5-CD4B07E988BD_zpszhvf9tvl.jpg" border="0" alt=" photo D7849EB9-B112-44B3-B5D5-CD4B07E988BD_zpszhvf9tvl.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/DE7593D5-6167-4987-ACF2-8F7EE0CC5E6B_zps1lrp37us.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/DE7593D5-6167-4987-ACF2-8F7EE0CC5E6B_zps1lrp37us.jpg" border="0" alt=" photo DE7593D5-6167-4987-ACF2-8F7EE0CC5E6B_zps1lrp37us.jpg"/></a>

[karimwassef]

Looks like DPAK TO-252

[karimwassef]

Opening up the stock controller, I can't help but think it would be easy to hack and use an arduino to signal up and down and then take the gate signal off one of the FETs to create a counter of cycles (~flow counter).

[d0ughb0y]

the board looks exactly the same, except yours have 2 1000uf capacitors on the bottom side while the 50watt controller only has 1. I really think the controller is the same, only a higher amp power supply is provided. Considering the biggest jebao pump use less than 5 amps, those TO-252 mosfets should be able to handle it.

You can hack this now by tapping on the signal output (output 1 2 or 3, output 4 is not connected) of the LM339. Every 6 pulse is 1 revolution. Just filter out the blip noise in the signal (see my screenshot of the signal above).

or if you don't want to deal with the BEMF signal noise, you can tap on to the high side signal, that's pin 7,9,or 11. You can use any one.

[d0ughb0y]

I just noticed on your picture of the fets, you can see diodes next to each fet.

[karimwassef]

ok. what's next?

[d0ughb0y]

I think I got my rpm calculations above wrong.
6 commutation is only equal to 2 high side pulse, not 6. So multiply 745 by 3 = 2235rpm at 70% (lowest setting). if there are internally 6 coils instead of 3, then divide that by 2. I don't really know the typical rpm of a water pump. It may not matter, you can calibrate to determine the number of pulse to get 1 gallon. the head pressure would not matter, as 1 gallon is 1 gallon and will take the same amount of revolution to deliver, the difference being only time.

as to what's next. I'm going to buy some parts to test (diodes, fets, etc).
I can't really do much now since I have a house guest using the room where I have all my stuff for this project setup.

[karimwassef]

Ok. Other than the 1K Ohm, is there anything else I should try?

[d0ughb0y]

wire the gate diodes like this. use a 1A fast recovery diode. (UF4007)

http://i1193.photobucket.com/albums/aa352/jerrysy/controller/greenboard.jpg~original


if you feel like taking a risk, you can just wire a jumper wire between the two points. The white controller is bare bones and is wired directly.

I don't have a diode symbol. The arrow points to direction of current flow, just like diode symbol.

[karimwassef]

I don't know if I have any fast recovery diodes. Let me check my inventory.

[d0ughb0y]

I snuck into my quest room to do some soldering. :)

This really put my soldering skill to the test

http://i1193.photobucket.com/albums/aa352/jerrysy/controller/9D13F3DE-4303-471C-9854-031DC9B187DE.jpg~original

http://i1193.photobucket.com/albums/aa352/jerrysy/controller/101258D6-E45E-4A26-9440-829FCE366C0E.jpg~original

This is the commutation signal. Running at 100%, the RPM is 60000/25.2367 = 2365 rpm, which is about the same speed as the lowest setting on the stock controller. So definitely the green board runs slower.
Not sure if those blips/noise on the commutation signals are degrading the power or not. probably does, it messes up the magnetic field.

http://i1193.photobucket.com/albums/aa352/jerrysy/controller/greencommute.jpg~original

[karimwassef]

so can we make it go faster?

[d0ughb0y]

I will compare the circuits of the stock, white and green boards and try to make changes and see if anything works.

[stage3-s4]

whats the max speed you can got out the pump with the ebay controller?

[d0ughb0y]

the max I got on the green ebay controller was the min on the stock. The whilte ebay controller I am assuming the same as stock since the power usage is exactly the same.

I think in order to fix this, the hall input signal coming from LM339 must be fixed to reduce or remove those isolated pulses. The stock controller has it as well, but they are quite narrower and I think the controller chip knows to ignore those, but the one on the green board is twice as wide. It may be the capacitors in the resistor divider is causing the signal to switch slower, so I may try to remove the capacitors and see what happens.
Also, adding the diode or just shorting the gate resistor will also speed up the switching of the FET.

this is the stock controller. On the last two signals, the noise is a small high pulse, and a small low pulse within the larger high pulse.
http://i1193.photobucket.com/albums/aa352/jerrysy/controller/jebaotiming.jpg~original

on the green controller signal, it looks like there are two high pulses within the larger low pulse, and two smaller low pulses within the larger high pulse.
http://i1193.photobucket.com/albums/aa352/jerrysy/controller/greencommute.jpg~original

I only have an 8 channel logic analyzer, so HU signal is not captured. I could use a 16 channel logic analyzer for this application.

[d0ughb0y]

I compared the signals of the white and green board using an oscilloscope, and the wave form itself looks similar, but the narrow pulses are a bit wider on the green board. That difference might be causing the control chip to slow down the motor.

The bemf circuit is almost identical, even resistor values. The white board has fewer capacitors, and I'm thinking the added capacitor might be causing the pulse to get a little wider. I'm going to try to remove the 3 capacitors on the green board and see if it makes any difference.

The white board also connects the driver signals directly to the transistor/fets without resistor. I can also short the resistors out to see if it makes any difference.
If that is still not enough, the next step is to replace the N channel FETs with ones that have lower gate capacitance.

I think these are all the possible mods that can be done on the green board to see if it improves flow.

[d0ughb0y]

BTW, on the stock controller, if you wipe off the heatsink glue on the fets, you can see the part numbers. on the 50watt controller, it's AOD413A and AOD454 and these parts are available from digikey. The N channel is rated 40v 20A and P channel is rated 40v 12A so I think this is the same fet used on the 100w stock controller as the max current is less than 5A. Karim, you can confirm on 100w controller in case you still have it disassembled.

If I have to build a controller from scratch, I will use the same FETs.

[Dactylopterus]

I Hope this helps. It is the 100W contoller from a Jebao DCS7000
http://uploads.tapatalk-cdn.com/20160629/95948bad6ef4f2a10a9018a82497ba25.jpg
http://uploads.tapatalk-cdn.com/20160629/d778c798fa58b2fdfb4f107a5a696c05.jpg
http://uploads.tapatalk-cdn.com/20160629/a3bd3c6237078fdafc92df67d67e823d.jpg http://uploads.tapatalk-cdn.com/20160629/c32da6f04662377c497e5d84944feafa.jpg

Sent from my SM-G900W8 using Tapatalk

[d0ughb0y]

Thanks. That helps.

I see it is pretty much identical, except for high power mosfets

AOD4186 N channel mosfet 40v 35a
AOD4185 P channel mosfet 40v 40a

and the second 1000uf capacitor.

Not sure why they don't just use the 100w controller for all pumps and just include different wattage power supply.

[karimwassef]

The higher power mosfets may cost more or be less available.

[karimwassef]

2" flow sensor is in..

<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_6404_zpssbb5yiq5.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_6404_zpssbb5yiq5.jpg" border="0" alt=" photo IMG_6404_zpssbb5yiq5.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_6403_zps828xsgvc.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_6403_zps828xsgvc.jpg" border="0" alt=" photo IMG_6403_zps828xsgvc.jpg"/></a>

it has a SS clip, but I think I'd use it and find out how it works out.

The fan has "no resistance". I barely breathe into it and it turns like a turbine!

<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_6400_zpsxa6ibgqz.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_6400_zpsxa6ibgqz.jpg" border="0" alt=" photo IMG_6400_zpsxa6ibgqz.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_6402_zpsbl6n3l6l.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_6402_zpsbl6n3l6l.jpg" border="0" alt=" photo IMG_6402_zpsbl6n3l6l.jpg"/></a>

[karimwassef]

doughboy - since you're the master Arduino dude.. would you please write the code to turn the sensor measurements into flow (GPH)?

:D

Sorry - working too hard and need some help.

[d0ughb0y]

if you google search, there are plenty of arduino code that works on this type of flow meter. Even adafruit sells this type of flow meter.

[stage3-s4]

wow. The 1" looks way different than that. It basically cuts off half the cross sectional area at the exit of the meter. That one should work fine.

[d0ughb0y]

I compared the signals between the boards, and the white and green board signals are actually similar, which makes sense since they both have the same chip and bemf circuit is identical.

The problem seems to be the ebay board is skipping one commutation (you can see that in the logic analyzer signal I posted above), so instead of on, off, on, off, on, off it gets on, off, off, on, off, off. Thus is getting less power.

I have to figure out why it is commutation signal is skipping and if this can be fixed by component change.

[karimwassef]

Do you still want me to try the resistors and diodes? It's been a busy week - quarter close today.

[d0ughb0y]

Do you still want me to try the resistors and diodes? It's been a busy week - quarter close today.



No need. I want to make sure first that it will make a difference.

[karimwassef]

Ok. Let me know.

[d0ughb0y]

This is the jebao BEMF signal.
They are longer and overlaps. That's how it should be.
http://i1193.photobucket.com/albums/aa352/jerrysy/controller/jebao-bemf.png~original

This is from the ebay controller. They are shorter and don't overlap. So the coils are not energized as long.
http://i1193.photobucket.com/albums/aa352/jerrysy/controller/ebay-bemf.png~original

So I need to find a way to make the ebay controller BEMF signal be like Jebeo's.

[d0ughb0y]

Karim,

can you try to connect

MA - U
MB - V
MC - W

I think that fixed the problem. I only tested on the white board. The killawatt reading went from 35watts to 45 watts.

I have to desolder the cable from the green board to test this on the green board.

Just checked my connections again. It should be as shown in the green board connection picture on page 1. I had switched U and W.

[d0ughb0y]

I just rewired the connection on my green board, and now I am getting 37 watts reading on the killawatt, and the BEMF signal now overlaps, but the flow seems even less, and the pump makes a whining noise. Tomorrow, I'll try all 6 combinations of connections and see which one works best..

I just shined a flashlight on the pump outlet, and the connection UVW-ABC spins in reverse. Somehow the water still comes out the outlet. That's whats causing the whining noise, and less flow. With the original connection, I can see the pump spin the same direction as the stock controller.

[karimwassef]

Any updates?

[d0ughb0y]

Any updates?



none since my last update. I will try all 6 WVU-ABC connections to see which one is really the right connection. If that does not help, my next step is to try to modify the BEMF ciircuit and change it so it is like the one on the stock controller. Are the resistors 0603 or 0402 size? I need to get 50k smd resistors to modify the circuit. If modifying the circuit works, it would not be a practical mod.

[stage3-s4]

none since my last update. I will try all 6 WVU-ABC connections to see which one is really the right connection. If that does not help, my next step is to try to modify the BEMF ciircuit and change it so it is like the one on the stock controller. Are the resistors 0603 or 0402 size? I need to get 50k smd resistors to modify the circuit. If modifying the circuit works, it would not be a practical mod.


I know I don't have the capabilities to work with SMD devices.I tried once for a project I did in college, It didn't turn out too well. Hopefully its just the connections.

[karimwassef]

Just because the existing implementation is SMD, we can still use wire resistors. Machine placement only matters if you're making hundreds of boards but if it's one two, a little elbow grease is ok.

[stage3-s4]

What are the 6 possible connections? I can test them when I get home.

[d0ughb0y]

you just do the permutation of connections from WVU to ABC on the board.

You can shine a flashlight on the pump outlet to see the direction of spin when it starts up.



WVU

WUV

VWU

VUW

UWV

UVW

[stage3-s4]

here is what I found:

UVW: spins backwards, draws 1.5 amps, bad flow, very noisy.

WVU: draws 1.13 amps. decent flow, somewhat noisy.

UWV: spins backwards, draws 1.5 amps, bad flow, very noisy.

VUW draws 1.23 amps, flow seems to be almost as good as stock controller, more noise than stock controller.

UVW: same as above.

WUV: spins backwards, draws 1.5 amps, bad flow, very noisy.

All configurations make the pump wine much louder than with the stock controller. Also, when I try to adjust the speed with the VR input, the pump doesn't really adjust speed smoothly. Seems to struggle with keeping up with RPMs.

[karimwassef]

I've been running WVU

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsye5ajie2.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsye5ajie2.png" border="0" alt=" photo 0_zpsye5ajie2.png"/></a>

how much louder and more powerful is VUW or UVW compared to WVU?

[d0ughb0y]

That looks right that 3 of the connections make the pump spin in reverse.



One more thing to try though is with one of the connections that runs the pump in reverse, connect the Z/F direction pin to 5v to make it reverse, hence making it run in forward direction.



I am thinking since the wiring that runs reverse use a higher current, if it maintains the higher current but runs in forward direction, that would give the best flow.



The commutation signals of the lower current (forward spin) do not overlap, the signal looks more like ON-OFF-OFF-ON-OFF-OFF.

On the reverse spin, the signals look like ON-ON-OFF, ON-ON-OFF, hence the higher current.



From what I understand, to reverse spin, you simply reverse the sequence of commutation signal. So if that is the case, the signal for each line will still be ON-ON-OFF, but in reverse sequence, so pump runs in forward direction at higher current. At least that's what I am hoping.

[karimwassef]

oooo smart.. worth a try.

[stage3-s4]

That looks right that 3 of the connections make the pump spin in reverse.



One more thing to try though is with one of the connections that runs the pump in reverse, connect the Z/F direction pin to 5v to make it reverse, hence making it run in forward direction.



I am thinking since the wiring that runs reverse use a higher current, if it maintains the higher current but runs in forward direction, that would give the best flow.



The commutation signals of the lower current (forward spin) do not overlap, the signal looks more like ON-OFF-OFF-ON-OFF-OFF.

On the reverse spin, the signals look like ON-ON-OFF, ON-ON-OFF, hence the higher current.



From what I understand, to reverse spin, you simply reverse the sequence of commutation signal. So if that is the case, the signal for each line will still be ON-ON-OFF, but in reverse sequence, so pump runs in forward direction at higher current. At least that's what I am hoping.

Ill try that. What about the louder wining noise from the pump? Any explanation?

[stage3-s4]

I've been running WVU

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsye5ajie2.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsye5ajie2.png" border="0" alt=" photo 0_zpsye5ajie2.png"/></a>

how much louder and more powerful is VUW or UVW compared to WVU?

Its pretty much the same but you deff can notice the difference in flow. That setting doesn't seem correct to me. I have my pump setup to flow water through a piping setup I made.

[d0ughb0y]

I will be able to resume work on this Sunday night and try the different connections.

On the forward spin, I do not hear any whine, only water flow sound, but maybe that is masking the whine.

If nothing works, I will still use the ebay board for now as even the max flow on it is more than I need. If the ebay board turns out it cannot work at least the same as the stock controller, I will try to build a controller from scratch, which was how this started before I stumbled into the ebay controller.

[karimwassef]

ok! I'm hoping this works. I need to get it up and running this weekend due to other obligations for the next couple of weeks.

[vxmvm]

If you use a potentiometer, you would not know where to stop turning. Turning all the way will feed 10v to the 5v input.http://yeahonlinemarketing.com/apple/images/33.gifhttp://yeahonlinemarketing.com/apple/images/6.gif
http://yeahonlinemarketing.com/apple/images/4.gif

[d0ughb0y]

I just tested all connection combinations for the green and white ebay boards.

It boils down only to 2 unique modes. For the green board, forward spin runs 26 watts, reverse runs 37 watts. For the white board, forward runs 34 watts, reverse runs 45 watts and gets better flow than the green board. If I ground the direction pin to reverse it, it just goes from 37 to 26 or 26 to 37 watts.

The stock controller runs 34 watts, and flow is higher than the white ebay controller.

I don't think it is worth modifying the ebay boards circuits, unless I stumble into something simple to do.

I'll probably just use the white ebay board as is.

I'm going back to work on my controller project, and will look into this again, perhaps build a controller from scratch.

[theatrus]

Looks like a lot of thread happened while I was out :bounce2: !

I suspect the EMF comparator is simply triggering from the noise, and its throwing the controller into a loop. It would potentially help to trace out the Jebao driver its self to see its sense circuit. Swapping some SMD caps and resistors isn't that hard.... The filtering may need to become more aggressive (larger cap, etc) or maybe more than one pole.

As for a gate-resistor, that will slow the FET down but also reduce ringing on the line (big square wave dumping into a gate capacitor is an EMI disaster). The slower the FET switches, the more energy you'll burn in it while its in an intermediate state. I don't think its truly essential here. I haven't studied the circuit enough to understand the diode use case...

Through the miracles of Amazon Prime Same-Day I will be the proud owner of a smaller DCT pump tomorrow to look at whats up with its control scheme.

For a ground-up replacement, the Allegro A4960-A looks like a great candidate - its designed for back-EMF measurements and doesn't need much external components to do it (lot is software settable). Its not going to end up costing $10, probably more like $30 to build it yourself. (because.. China)

http://www.allegromicro.com/en/Products/Motor-Driver-And-Interface-ICs/Brushless-DC-Motor-Drivers/A4960.aspx

I run a Vectra M1 on my tank and have a new-in-box spare, so I cracked open the controller. Its a step up from what I've been seeing in the Jebao. First off, the pump has either an inductive pickup trace or a single hall sensor (haven't checked, two wires). DRV8301 gate driver,

http://theatr.us/images/vectra/vectra1-1k.jpg
Full size:
http://theatr.us/images/vectra/vectra1.jpg

http://theatr.us/images/vectra/vectra2-1k.jpg
Full size:
http://theatr.us/images/vectra/vectra2.jpg

The main MCU is a Cortex-M4F (120MHz) (MKV31F256) http://cache.nxp.com/files/microcontrollers/doc/ref_manual/KV31P100M120SF8RM.pdf?fpsp=1&WT_TYPE=Reference%20Manuals&WT_VENDOR=FREESCALE&WT_FILE_FORMAT=pdf&WT_ASSET=Documentation&fileExt=.pdf

http://www.nxp.com/products/microcontrollers-and-processors/arm-processors/kinetis-cortex-m-mcus/v-series-real-time-control-m0-plus-m4-m7/kinetis-kv3x-100120-mhz-advanced-3ph-foc-sensorless-motor-control-microcontrollers-mcus-based-on-arm-cortex-m4-core:KV3x

It has a number of motor control features as well.

The BOM cost here is probably a good 2-3x the shipped-from-china cost of the eBay drivers (because.. China). The same MCU is probably used in the Vortech QD drivers (just stamp out the Freescale BLDC reference design and libraries, success).

[theatrus]

Also, EcoTech apparently loves Apple's PCB motif - black with no reference designators ;)

[theatrus]

Curiously, the big honking NFET is part of the MIC2587 hot swap controller:
http://ww1.microchip.com/downloads/en/DeviceDoc/mic2587-87r.pdf

I need to trace this out some more to see if its part of the battery backup input or if the main power input is being hot swapped (likely to act as a current limit at input, nice!). You usually don't waste money on this unless it really prevents some disastrous problem.

But back to the Jebao pumps!

[d0ughb0y]

The bemf signal on the jebao board is very clean, almost textbook.
I'll get some 0603 resistors from a local store to change the circuit on the green board to make it similar to the jebao circuit and see if it makes any difference.

[karimwassef]

fancy shmancy pumps and controllers! :D

So is there any way to make the green controller boards work?

[karimwassef]

EXCELLENT!

Hey - how does this look?

http://www.ebay.com/itm/BLD120A-Brushless-DC-Motor-Driver-30V-120W-BLDC-Controller-f-Brushless-Motor-CNC-/262364293044?hash=item3d162167b4

or this one

http://www.ebay.com/itm/CO-BL23B46-02-02RO-BRUSHLESS-DC-MOTOR-Driver-Board-28M127-/181764825174?hash=item2a52072056:g:4n4AAOSwKrhVbwzj

much cleaner layout. It's 25V min, but that may be close enough?

or

http://www.ebay.com/itm/1pcs-DC12V-24V-brushless-motor-controller-High-Power-Driver-360W-700W-30A-/152134063101?hash=item236be593fd:g:B2oAAOSw3YNXZLl8

would need an electronic potentiometer mod...

[d0ughb0y]

contact the seller and ask for specs and what motor controller chip is used. They can try getting it from their supplier.

[theatrus]

#2 is a step controller (note 2 full H Bridges):

http://www.alltronics.com/mas_assets/acrobat/28M127.pdf

#3 is gone.

[karimwassef]

#3 is still available.

How about #1?

[theatrus]

Looks like #3 came back - was a removed item when I clicked through.

Really isn't enough information to figure it out from the listing, haven't tried googling to see if someone is using these in a project somewhere else.

[karimwassef]

#1 uses hall sensors.. :(

So only #3 is possible.

[karimwassef]

and now it's gone ...

[theatrus]

Amazon delivered my DCT-4000 (50W controller). Looks like someone sanded out the main controller chip.

I wonder, if we were to do a new controller, why not just put it back into the same case as the existing controller. Plenty of room in here, could even re-use the little heatsink. I'm not sure how much bigger than controller thats above 50W is.

[stage3-s4]

I'm deff in if you guys want to build a controller. I just dont have the means to do that kind of stuff myself

[karimwassef]

I am too, but maybe have multiple channels for multiple pumps?

[theatrus]

Just doing some basic testing using the controller. The dry pump is, as I suspected, pulling significantly less power as a whole vs one in any liquid (at the lowest speed on the controller, 2W dry, 16W wet). The current shunt on the backside of the board is likely used for this measurement.

[d0ughb0y]

did the controller stop the pump when you ran dry?

[theatrus]

After several minutes, yes.

[theatrus]

Just for grins, I made a drawing of the major features of the 50W Jebao controller board:

http://theatr.us/images/blueacro/dcpump/jebao_50w_pcb_footprint.PNG

I can put up a DXF or similar file. How different is the larger pump PCB?

All dimensions are mm (since 95% odds it was designed in metric). The mounting holes are all 3mm (M3). The square features are the 1206 LEDs and the 6mm square buttons, respectively. The heatsink would be on the back side of the board from this view, and measure 28mmx41mm

[d0ughb0y]

I replaced some parts and it sort of worked.

It ran in forward direction at 36 watts, so that's good news that it at least matches the stock controller. But it does not start most of the time. I think the reference signal is too weak to generate a good LM339 output.

There is one more difference, the stock controller powers the LM339 at 18v and the bemf voltage divider drops the output voltage down to 8v. The ebay controller LM339 is powered by 5v and the bemf is voltage divided down to about 2v.

I'm guessing if I change the voltage divider value or even power the LM339 at 18v like the stock controller, this will be able to startup the pump every time.

red arrow points to the replacement resistors, it replaced capacitors that were there originally.

http://i1193.photobucket.com/albums/aa352/jerrysy/controller/IMG_2036.jpg~original

I am thinking now the best way to proceed is to build a controller from scratch.

[karimwassef]

I replaced some parts and it sort of worked.

It ran in forward direction at 36 watts, so that's good news that it at least matches the stock controller. But it does not start most of the time. I think the reference signal is too weak to generate a good LM339 output.

There is one more difference, the stock controller powers the LM339 at 18v and the bemf voltage divider drops the output voltage down to 8v. The ebay controller LM339 is powered by 5v and the bemf is voltage divided down to about 2v.

I'm guessing if I change the voltage divider value or even power the LM339 at 18v like the stock controller, this will be able to startup the pump every time.

red arrow points to the replacement resistors, it replaced capacitors that were there originally.


http://i1193.photobucket.com/albums/aa352/jerrysy/controller/IMG_2036.jpg



EXCELLENT!! We can use these controllers with a few component changes!


I am thinking now the best way to proceed is to build a controller from scratch.

Wait - what? Confused.

[d0ughb0y]

so I tried running it out of the water, and it now starts almost all the time, and it runs at 3 watts. So it is still possible with some resistor value tweaks to get this to work.

[karimwassef]

whew.. you know I have 4 of these green boards :)

[d0ughb0y]

this is how it is now. it still does not start all the time.
It has to start all the time for this to be feasible/acceptable mod.

<iframe width="560" height="315" src="https://www.youtube.com/embed/Hf0qnLADyWg?rel=0" frameborder="0" allowfullscreen></iframe>

[d0ughb0y]

now that I'm not taking a video, it starts every time again.
Anyway, I'll hook up the lm339 signals to an oscilloscope to see what is going on and maybe figure a way to get this to work all the time in water.

I just put the pump back in the water, and now it starts pretty much all the time. I think only once or twice it did not start out of maybe 20 tries.

and it is definitely putting out more flow than before the mod. maybe even same as stock, although it is hard to quantify visually.

[d0ughb0y]

Testing in water. Seems to start pretty much all the time now.

<iframe width="560" height="315" src="https://www.youtube.com/embed/LKvdTXJ5HiQ?rel=0&amp;controls=0&amp;showinfo=0" frameborder="0" allowfullscreen></iframe>

[theatrus]

Possible the new motor controller doesn't have a long enough open loop period to get things going, so the lack of a strong BEMF signals breaks it. Or the ramp is too fast for the pump.

What's the new capacitor/resistor value vs the old?

[karimwassef]

I would say that if it didn't always start, a smart controller can detect and restart.

Unfortunately - it depends on why it's not starting...

Sometimes, if the starting pulse isn't strong enough AND the motor phases are at the most distant state (trough of the flux distribution), the motor won't start. It's a random event and not easily reproducible.

In that case, we need to increase the starting drive.

[d0ughb0y]

This is the original BEMF circuit. I only show 1 of 3 op amp circuit. The other 2 are identical to the first one.

http://i1193.photobucket.com/albums/aa352/jerrysy/controller/bemf.png~original

I removed R3,R4,R5,R6,C2,C3,C4
The resistors are near the pcb edge next to the LM339.
The capacitor locationss are as shown in my earlier post.

solder 10k 0603 resistor in place where the capacitors were.

The new circuit use the voltage divided MA signal, all 3 tied together and fed to REF input of all 3 op-amps. This is essentially the bemf circuit of the stock controller (with some minor difference).

You need a good magnifier and a hot air rework station to do this mod.

links to youtube video so it shows on tapatalk
https://youtu.be/Hf0qnLADyWg
https://youtu.be/LKvdTXJ5HiQ

[Johohnn]

great find thanks for the info. Also looking for confirmation on wiring.

[karimwassef]

Ok. Different direction for a minute...

I'm experimenting with reverse flow through DC pumps, so I have one pump working normally and the other "off". The second pump line acts as a "drain" while the main pump is on. I didn't think much of it, until I noticed that the "off" pump's controller was flashing - as in a fault mode.

How can that be... There's no 24VDC on it. Sure enough, disconnecting the input power didn't stop it since the off pump's impeller was rotating due to the reverse flow and generating enough voltage in the power board to trigger the fault mode flashing lights.

Question - if I let this happen regularly, will I fry the controller?

This is the native Jebao controller.

[theatrus]

Looks like the green controller was trying to read AC coupled BEMF signals then. From some tracing on the Jebao there is a decent sized LPF on the signal (though the PWM harmonics still bleed into it a bit - need to be careful probing as well).

[theatrus]

Ok. Different direction for a minute...

I'm experimenting with reverse flow through DC pumps, so I have one pump working normally and the other "off". The second pump line acts as a "drain" while the main pump is on. I didn't think much of it, until I noticed that the "off" pump's controller was flashing - as in a fault mode.

How can that be... There's no 24VDC on it. Sure enough, disconnecting the input power didn't stop it since the off pump's impeller was rotating due to the reverse flow and generating enough voltage in the power board to trigger the fault mode flashing lights.

Question - if I let this happen regularly, will I fry the controller?

This is the native Jebao controller.



The pump is now an alternator, and the body diodes in the FETs are acting as your rectifier. It's probably not going to be too terrible unless the speed got really high.

The best solution would probably be braking the impeller - having a controller which can hold the impeller is an electrically braked position (power on).

[karimwassef]

I actually like having the reverse water flow through the pump. It flushes any debri back out. It's not the only drain, of course. But how can I tell if I'm stressing the controller?

A middle path option here would be to let it run in reverse for a minute & then brake or run in very low output mode.

[theatrus]

Stalling the impeller shouldn't dramatically change the reverse flow.

I need to trace out some more of the circuit to determine what else is powering up or if it's only looping back through the main power rail (the back EMF circuit is all high enough impedance from the looks of it). There is a ceramic cap and a big electrolytic on the rails so transients shouldn't be an issue.

If you check the voltage on the DC jack when it's in this mode, how high does it get? Probably at least 6+V to power up the main controller MCU (which I still suspect is a PIC, but haven't checked the header next to it).

If there is no real load, I don't think you're causing much damage. The FET body diodes aren't very robust usually.

[d0ughb0y]

I think the stock controller has more flow than the green controller with the mod.
It looks like the ebay board controller does not do any kind of filtering of the bemf signal, unlike the stock controller where it knows to ignore the narrower pulses (probably a cpu program). It is still usable and viable option, with or without the mod, just less flow and if you only need the variable speed controllability without the bells and whistles.

I'm going to use the controller as it is for now. Once I finish my diy controller, I'll try to build a pump controller.

[d0ughb0y]

electronic brake simply applies high/low power to any 2 of the 3 outputs to hold the magnet in place right?

[theatrus]

electronic brake simply applies high/low power to any 2 of the 3 outputs to hold the magnet in place right?



That's one idea. Most of the time the controller supports dynamic brakes (shunt coil energy to ground) so the motor slows faster. Simply leaving the FETs in brake mode would be plenty and not consume any extra power.

[karimwassef]

I'll just run in very slow flow. That's good enough.

I'll measure the voltage on the controller power input when I get home in a couple of weeks.

On the road, so can't help much for a bit.

[d0ughb0y]

I installed my dct4000 pump in my sump over the weekend and it works nicely. I am using the stock controller for now.

I just ordered a cp-40 today. The price dropped to $87. That's like 1/3 the price of a gyre. I'll test the ebay controller on the cp-40 to see if it works.

[karimwassef]

What are all the mods that improved performance?

[Dactylopterus]

What if we use the stock controller and control the power supply voltage from 12-24V with the APEX? Is it possible to do that? I have a WP40 running for a long time at 12V without any problems. I could try driving the DCS7000 pump with stock controller at 12 volts to see if it runs and how much flow it has. Will this damage the controller?

[d0ughb0y]

I tried using 12v power supply on the stock controller and it did not work.

[d0ughb0y]

What are all the mods that improved performance?

just what I posted in post 213.
you will need tools for surface mount soldering to do it.

[theatrus]

I tried using 12v power supply on the stock controller and it did not work.



The stock controller shuts down if the input voltage is less than about 18V from experimenting on the bench. Otherwise would have been a good idea :)

[karimwassef]

But you know, the concept of modifying the stock controller is intriguing.

Can't we access the 0-5V speed pin on the controller chip and disconnect the button controlled signal???

I think there are two controllers- a simple one for managing the buttons and feed input and converting it to a DC level signal. The power controller takes that signal and drives the phases through the expensive FETs...

Why waste all that? Just connect the APEX 0-5V onto the pin.

[d0ughb0y]

There is no 0-5v input on the stock controller chip.

you can probably build something to press the buttons for you.

[karimwassef]

What's the stock controller IC? It must have some way to control speed

[d0ughb0y]

the label is wiped out. it's probably a motor controller chip that can accept additional firmware program.

it's even possible it is not a motor controller chip at all but just an mcu.

[karimwassef]

How does it connect to the button outputs?

[d0ughb0y]

its a direct connection to the buttons, just as you would do connecting a button to an mcu.

[theatrus]

http://uploads.tapatalk-cdn.com/20160720/b7f5e70d71bed03e24ed99b83cbbc5ef.jpg

Excuse my probe lead wires. SW3 and SW2 is all you'd need. You can either use a FET across the button, or just feed 0/5V to the controller pin.

Note that the LEDs are multiplexed, so if you wanted to confirm the pump status you'd need to filter that out (Q14 and Q15) in firmware.

[karimwassef]

That's viable, but the idea was to achieve greater granularity by using a 0-5V signal directly from an Apex V output divided.

To recreate button pushes that emulate the desired Apex voltage would need a tiny arduino.

[karimwassef]

So why are there two controller chips?

[d0ughb0y]

There is only one, U3 is the controller. U2 is LM339, you can see the chip marking if you shine a light on the chip at an angle.

From what I remember when I traced the signals, each of the 11 LEDs connect to the mcu. 3 switch, 6 motor driver, 3 bemf, 1 current sensor, 1 voltage detector, +5, gnd. that's 27 pins. I can't remember what the last pin connects to.

So in order to get the same functionality (or to use the same size board), the replacement mcu must have at least 26 IO lines. Plus a pwm and a direction pin, that's 28 IO lines, more if you want more functionalities. So an atmega328 will not work as replacement. But if you want no frills simple controller that is controllable by pwm, then atmega328 will work just fine. I plan to build a no frills one for starters.

[theatrus]

U2 is a comparator (need to check the part number) used for the sensing.

I'm mostly thinking U3 is an MCU due to:

The five 0.1" pads next to the chip smell like a programming header.

Q6/Q7/Q8 are part of the main FET driver circuitry. A dedicated motor controller would likely have these on chip.

[theatrus]

Also, the 150W version of the board has an identical PCB. The third sense resistor is populated on the backside, and the second electrolytic cap is present. The heatsink is slightly larger. I haven't checked for any part number differences.

[theatrus]

There is only one, U3 is the controller. U2 is LM339, you can see the chip marking if you shine a light on the chip at an angle.



From what I remember when I traced the signals, each of the 11 LEDs connect to the mcu. 3 switch, 6 motor driver, 3 bemf, 1 current sensor, 1 voltage detector, +5, gnd. that's 27 pins. I can't remember what the last pin connects to.



So in order to get the same functionality (or to use the same size board), the replacement mcu must have at least 26 IO lines. Plus a pwm and a direction pin, that's 28 IO lines, more if you want more functionalities. So an atmega328 will not work as replacement. But if you want no frills simple controller that is controllable by pwm, then atmega328 will work just fine. I plan to build a no frills one for starters.



Only 8 lines are used for the LEDs in the original design (6 cathodes, 2 anode muxes).

The comparator is marked as an LM2901 but that's basically a variant of the LM339 4 channel. http://www.ti.com/lit/ds/slcs006t/slcs006t.pdf. Cleaning the board with acetone or xylene makes fast work of the conformal coating.

I'm going to do the "plug and play" driver replacement and use an A4960 and toss in an nRF52 (cause Bluetooth LE is how I'm running everything now).

[karimwassef]

So there's no way to big in a flow speed voltage reference ?

[theatrus]

So there's no way to big in a flow speed voltage reference ?

Not that is apparent - speed is all digital in this design, driven by the mystery MCU.

[d0ughb0y]

A cheap Arduino pro mini can probably be used on a diy controller.

if I am not mistaken, pro mini has 21 or 22 usable IO lines.

6 for motor driver
3 for bemf
1 current sense
1 voltage sense
1 pwm speed
1 direction
2 for I2C can be used for optional lcd display
2 for serial can be used for low speed esp8266 wifi
1 maybe for a buzzer for alarm
the rest for push buttons

I can get some FETs and wire this up on a breadboard and see if it will work.

If it does and is feasible, the plan is to make this controllable from any pwm and digital pin source to control speed and direction, and have a few basic built in program modes for use with crossflow and DC pump to work stand alone.

[karimwassef]

I thought the problem with digital motor control is the processor speed. You need to line up the timing of the drive transitions against the signals being received through the same wires and you need a very fast processor to make that work?

[karimwassef]

I now have 5 DCT15000. Three for flow and two that I'm using in series to create pressure for my skimmer penductor injectors.

So... Control would be excellent. :D

[theatrus]

A cheap Arduino pro mini can probably be used on a diy controller.

if I am not mistaken, pro mini has 21 or 22 usable IO lines.

6 for motor driver
3 for bemf
1 current sense
1 voltage sense
1 pwm speed
1 direction
2 for I2C can be used for optional lcd display
2 for serial can be used for low speed esp8266 wifi
1 maybe for a buzzer for alarm
the rest for push buttons

I can get some FETs and wire this up on a breadboard and see if it will work.

If it does and is feasible, the plan is to make this controllable from any pwm and digital pin source to control speed and direction, and have a few basic built in program modes for use with crossflow and DC pump to work stand alone.



Don't forget gate drivers, especially if you're using all N-FETs. The gate charge of even the low side FETs is too large for PWM duty on an AVR.

[d0ughb0y]

Don't forget gate drivers, especially if you're using all N-FETs. The gate charge of even the low side FETs is too large for PWM duty on an AVR.

That is the 6 motor drivers.

[theatrus]

That is the 6 motor drivers.



Got it - have a part in mind?