Building new pypilot motor controller

[dusan]

The boards have arrived!

I've managed to fully populate a few of them, and the next step is programming and testing. I'm attaching a picture so you can see how nice they turned out.
I'm also planning to design a protective cover/case and a mounting support so the board can be securely screwed into an appropriate location on the boat. Once I have a 3D prototype, I'll share the design files and results here.

So far, I'm very happy with how the boards came out!

[sinneD]

Hello all,

I'm building my own motor controller for my boat, as I love to tinker with things, at my dayjob I have a lot of experience designing electronics. Anyone interested in new features, or particular specifications?

Best,
Dusan

I'm interested. I have some KiCAD and fab experience. I have used JLC and Digikey for my projects too. I am happy to collaborate and test and support.

sinneD

[dusan]

I'm interested. I have some KiCAD and fab experience. I have used JLC and Digikey for my projects too. I am happy to collaborate and test and support.

sinneD

Thanks, sinneD!

That's great to hear.

I have a small batch of boards assembled. Would you be interested in getting one or two for yourself? I'm currently moving into the programming and validation phase.

Let me know if you're interested, and we can work out the details.

Best,
Dusan

[dusan]

Update on the boards:

I tested the board voltages and the regulators are working fine. After resolving a few minor issues, I was able to successfully upload the bootloader and firmware.

I've now tested all the main functionality, and everything appears to be working as expected. So far, I've tested the current capability up to around 6A, as I don't currently have a larger load available. No issues whatsoever were observed during testing.

Later today, I'm planning to pick up some 9004 halogen bulbs and connect them in parallel to create a heavier test load. That should allow me to push the board beyond to its intended operating limits. I'll share more results once I've completed those tests.

[sinneD]

That is exciting news! 
Do you have access to a 12V diesel heater? That might provide a tester for inrush loads and steady state loads too. iirc, ~10A. 
sinneD

[dusan]

That is exciting news! 
Do you have access to a 12V diesel heater? That might provide a tester for inrush loads and steady state loads too. iirc, ~10A. 
sinneD

I don’t have access to a diesel heater at the moment, but I think the bulbs should work well. The cold filaments should provide a decent inrush current, and by switching multiple bulbs in parallel I can test a range of load levels.

I'm particularly interested in seeing how the board behaves thermally under sustained load—how much the MOSFETs and PCB heat up, and at what current level active cooling becomes beneficial or necessary. That should give a pretty good indication of the board's real-world operating limits.

[dusan]

High Current Load Test Results

I completed load testing of the board using halogen headlight bulbs as the load.

Code:

Current      Voltage Drop      Temperature
12.2A        78mV              34°C
15.1A        98mV              35°C
19.1A        130mV             38°C
22.2A        153mV             44°C
26.5A        197mV             55°C
29.8A        221mV             60°C

At approximately 30A, the measured voltage drop was 221mV across the reverse-protection MOSFET, the current sensing resistor and the bottom bridge MOSFET path, including the associated PCB traces, connectors, and other resistive elements in that measured section. This corresponds to approximately 7.6mΩ  for the measured part of the path.
The complete active current is estimated to be around 11.09mΩ, that corresponds to an estimated total dissipation of about 9.2W.


The limiting factor turned out to be the software thermal protection, which is currently set to shut down at 60°C. The board reached that temperature at 30A during sustained operation.

I also ran a test at approximately 30A with a fan blowing across the board. In that configuration, the temperature stayed around 45°C (Ambient 28°C) showing that even modest airflow significantly improves cooling. Unfortunately, I couldn't continue that test for longer than 1h because my improvised load bank failed rather dramatically and ended the experiment.

Overall, I'm very happy with the results. The hardware appears to comfortably handle the target current, and the thermal performance is in line with expectations.

Based on these results, I would expect the board to operate comfortably in the 20–25A range without forced cooling when installed in a large enclosure with good ventilation. For installations in smaller or sealed enclosures, or where ambient temperatures may be elevated, I would recommend fitting a fan to provide additional thermal margin.

[seandepagnier]

7.6m ohm seems reasonable, but 110.9mΩ doesn't make sense at 30 amps. This would dissipate more than 100 watts.

In my 30 amp controller the resistance is about 2-3mΩ for the shunt and mosfets on low side, but the connectors and fuse add another 4-5, so almost the same.

In my tests mine did not heat up more than 15C without cooling at 30 amps, so the controller you built is probably more comparable to the 15 amp one (medium power) size.

For a load I used stainless steel wires.

[dusan]

7.6m ohm seems reasonable, but 110.9mΩ doesn't make sense at 30 amps.  This would dissipate more than 100 watts.

In my 30 amp controller the resistance is about 2-3mΩ for the shunt and mosfets on low side, but the connectors and fuse add another 4-5, so almost the same.

In my tests mine did not heat up more than 15C without cooling at 30 amps, so the controller you built is probably more comparable to the 15 amp one (medium power) size.

For a load I used stainless steel wires.

Sorry there was a typo. the total resistance of the board is estimated to about 11.09mΩ. This includes all the connectors, traces, internal MOSFET's Ron and shunts. Not sure how can you do lower with only 2 OZ boards.