Building new pypilot motor controller

[dusan]

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

[dusan]

Changes so far:

Updated transistors to NTMFS5C628NL MOSFETs, targeting 30A operation
PCB special plating of 4.5oz copper for high current
Switched to standard-size fuse for easier replacement
Added red LED fuse-blown indicator
Supports both 12V and 24V systems
Merged rudder sense and E-stop connectors into a single waterproof connector/cable assembly
Using JST JWPF waterproof connector series
Added cooling fan control via PB5 pin
Fan activates whenever the pump is running or autopilot is engaged

Main goals are reliability, serviceability, and keeping everything marine-friendly.

Still open to ideas for other features or protection/monitoring functions people would find useful.

[dusan]

I just made a trial for ordering the boards. Unfortunately the manufacturer doesn't have all the parts in stock, so I will need to order missing parts from DigiKey and populate them myself. Anyway the board price comes to about $75.

Anyone interested?

Here is a rendering:

[Lastknownposition]

I just made a trial for ordering the boards. Unfortunately the manufacturer doesn't have all the parts in stock, so I will need to order missing parts from DigiKey and populate them myself. Anyway the board price comes to about $75.

Anyone interested?

Here is a rendering:

I would be interested in buying one (or possibly 2) of your boards.
Bruce

[dusan]

I would be interested in buying one (or possibly 2) of your boards.
Bruce

I'm just finishing the design and ordering the boards. Most likely JLCPCB needs about a week for the order and another one for shipping. As soon as I have the boards, I'll populate the missing parts and keep 2 for you.

Best,
Dusan

[seandepagnier]

1) smaller size fuse will have more contact resistance creating more heat. This is why the large maxi fuse was selected.
2) using a single mosfet (of higher on resistance) will make much more dissipation. My 30 amp board has 500 micro ohm on resistance (because 4 parallel) this has 2 milli ohm. So it will create basically 4 times more heat? Why not put 4 in parallel? For me, the 30 amp rating means the board will not heat more than 10C and remain very efficient.
3) Adding a fan makes it less reliable because moving parts have lower reliability. This is why I used 63v capacitors because they will have a very long life in a 12/24v motor controller. If this depends on a fan for operation because of higher heat dissipation because of fewer mosfets the reliability is reduced not improved.
4) is there any longer a buck converter with overcurrent (short circuit protection so if rudder feedback or other sensors short out) for the 5v side and sensors? This reduced quiescent (standby consumption) to 2mA, but I dont see any inductor for this.

Maybe with 4.5 oz copper compared to 1oz, it can dissipate heat much better, but now the BOM cost is $75?? nearly double the original design that was also 4x more power efficient. Is this really "improved"? Maybe the pcb is a bit smaller?

I think you will find that hand soldering components will create a bottleneck in production. I learned this the hard way after consuming more than 8 pounds of solder by hand soldering. At least my current design can be fully assembled.

if something is better or can be improved and I can get feedback for why I would like to adopt it but I am skeptical, maybe the connectors (JST JWPF) are an improvement over the chinese waterproof connectors I've been using, can you say why?

Please continue the discussion! Also consider brushless motor controllers.

[Lastknownposition]

I would be interested in buying one (or possibly 2) of your boards.
Bruce

I'm just finishing the design and ordering the boards. Most likely JLCPCB needs about a week for the order and another one for shipping. As soon as I have the boards, I'll populate the missing parts and keep 2 for you.

Best,
Dusan

Thanks Dusan,

I see that Sean has commented on your design, unfortunately I cant make an intelligent comment as I am not knowledgeable on this subject what so ever, although I am keen to start my education. Let me know what you need from me.

Thanks,

Bruce

[dusan]

1)  smaller size fuse will have more contact resistance creating more heat.  This is why the large maxi fuse was selected.
2)  using a single mosfet (of higher on resistance) will make much more dissipation.  My 30 amp board has 500 micro ohm on resistance (because 4 parallel) this has 2 milli ohm.  So it will create basically 4 times more heat?  Why not put 4 in parallel?  For me, the 30 amp rating means the board will not heat more than 10C and remain very efficient.
3)  Adding a fan makes it less reliable because moving parts have lower reliability.  This is why I used 63v capacitors because they will have a very long life in a 12/24v motor  controller.  If this depends on a fan for operation because of higher heat dissipation because of fewer mosfets the reliability is reduced not improved.
4)  is there any longer a buck converter with overcurrent (short circuit protection so if rudder feedback or other sensors short out) for the 5v side and sensors?  This reduced quiescent (standby consumption) to 2mA, but I dont see any inductor for this.

Maybe with 4.5 oz copper compared to 1oz, it can dissipate heat much better, but now the BOM cost is $75?? nearly double the original design that was also 4x more power efficient.  Is this really "improved"?  Maybe the pcb is a bit smaller?

I think you will find that hand soldering components will create a bottleneck in production.  I learned this the hard way after consuming more than 8 pounds of solder by hand soldering.  At least my current design can be fully assembled.

  if something is better or can be improved and I can get feedback for why I would like to adopt it but I am skeptical, maybe the connectors (JST JWPF) are an improvement over the chinese waterproof connectors I've been using,  can you say why?

Please continue the discussion!  Also consider brushless motor controllers.

Thanks Sean, all valid points and I appreciate the feedback. A lot of your design decisions clearly come from extensive real-world experience and long-term field reliability considerations. My goals are slightly different in some areas, so I wanted to explain the reasoning behind some of the choices.

Fuse choice
I agree the larger fuse solution has lower contact resistance and better thermal characteristics. My reasoning for switching to standard automotive-style fuses is mostly availability. In many marine situations, being able to source and replace a fuse quickly matters more than achieving the absolute minimum resistance.

MOSFET selection
You are correct that the dissipation numbers are higher compared to paralleling lower Rds(on) devices. The tradeoff I was aiming for was simplifying the layout and reducing balancing/switching complexity while still staying within acceptable thermal limits.
The heavier copper PCB I think will offset the thermal margin due to somewhat higher Rds(on), and should still be acceptable for the intended use case. The goal here was more robustness than absolute efficiency optimization. Having more parts increase the probability of something failing.

Fan cooling
I completely agree that passive cooling is ideal whenever possible. The reason I added optional fan support is because some installations end up in poorly ventilated compartments or very warm environments, especially with larger hydraulic systems.
The intention is not to rely on continuous fan cooling, but rather to have controlled airflow available only when needed. Lowering MOSFET and capacitor temperatures significantly can improve overall lifetime, even if it introduces one moving component.

Cost increase
That’s definitely a fair concern. I think this mostly comes down to different priorities. The original design is very optimized for efficiency and low BOM cost, while I’m prioritizing marine-oriented robustness, connector quality, serviceability, and easier field maintenance.
Even with the increased component cost, the overall system is still significantly below commercial autopilot hardware pricing.

Hand assembly
Agreed — this is not optimized for large-scale assembly. Right now I’m still in the prototype / low-volume stage where flexibility and iteration speed matter more than production efficiency. If the design stabilizes and volume increases, I’ll definitely revisit assembly optimization.

Overall I absolutely appreciate the feedback. 

Best,
Dusan

---

I would be interested in buying one (or possibly 2) of your boards.
Bruce

I'm just finishing the design and ordering the boards. Most likely JLCPCB needs about a week for the order and another one for shipping. As soon as I have the boards, I'll populate the missing parts and keep 2 for you.

Best,
Dusan

Thanks Dusan,

I see that Sean has commented on your design, unfortunately I cant make an intelligent comment as I am not knowledgeable on this subject what so ever, although I am keen to start my education. Let me know what you need from me.

Thanks,

Bruce

Thanks Bruce,

I really appreciate the interest and support.
At this stage I’m still waiting for the jlcpcb boards, they had some feedback on the design. Once the first boards arrive and I complete assembly/testing, I’ll let you know.
For now there’s nothing you need to do — I’ll keep you updated as the boards arrive and the project progresses.

Best,
Dusan

[seandepagnier]

I can certainly agree but, not sure having mosfets in parallel reduces reliability much if any... maybe stress test it with continuous 30 amps, and see how much it heats up and what the voltage difference from the input to output is. Then maybe put at least 2 parallel not only one.

Anyway, yes the fuse is a good question and the connectors as well, I enjoy people exploring alternatives to my choices.

The fan keeping it cool can increase lifetime but really the way to decrease heat and increase lifetime is as I mentioned before, not adding moving parts, but its still ok to explore!

[dusan]

Update on the boards: The manufacturing by JLCPCB is finished, they are shipping today. It took longer than usual, they were saying because they don't have many orders for the thicker 4.5oz copper boards. As soon as I have them, I'll populate the missing parts, test them and then they will be ready to be shipped. Assuming Bruce want 2 of them, I'll have 6 for other people. I'm selling them at cost, $75 + shipping.