For thermal conductivity, more is better. It’s watts/ (meter * deg-K). For insulation, lower is better.
1 Like
No, higher is better.
The machine pin headers are OK. That wasn’t the problem.
Did you find out the problem?
@ziploc any progress on the enclosure you could share?
I gave up. I don’t have the info I need to make the design. If someone wants to provide that, I’d be happy to do it.
You said you needed pictures with measurements right?
Which measurements are missing that you cannot get from den ECAD?
I’m happy to measure it if I can. I cannot provide good pics (my prototype is not how it should be assembled). I can print the model and provide feedback.
I’m starting to get a little annoyed by these machine pin headers that I introduced myself. The fact that a connector isn’t locked in position (like the jst) is a real world problem.
Anyone has knowledge of some kind of machine headers that can be locked in position? (where you are still able to detach the mcu board (pill))
Maybe I should just make a complete new pill design where all external connectors are on the pill and of type JST, with the exception of those going to the power board, which can stay as machine headers. - and then move CAN to the pill.
Maybe the IMU shouldn’t even be part of the board, but something that can be attached with an I2C or SPI connector. It’s just causing so much trouble with regards to the price of PCBs.
Going to make a pinout for Vbat, to allow the ability to store vital parameters between boot cycles. It’s very important in robotics - incremental encoders is a good example.
5 Likes
Hmm, Machine headers for the power board has caused me issues.
- The pins interfer with the body of the MosFet lying flat.
- It’s so easy to bump the board, nocking it off and bending the pins. I now have done it onve to often, and pins are broken completely, so going to have to remove them …
I’ll tell you what… if you take some pics of your unit from several angles, I’ll mark them up with the dimensions I need, and ask a few questions. From that we can come up with a design that works for you.
What do you mean by “interfer”? Like electromagnetic interference or ?
No. Mechanical inteference. One MOSFET body cover’s one set of 3-pins on the power module. Very difficult to solder those pins in to be flush with the bottom of the board, so the MOSFET body can lay flat.
Does not suit me. I am using pressure rather than screws to mate the FET’s with the heatsink, so I need everything perfectly flat.
(I have a spring arrangement holding the boards down from the top, far simpler than adding individual screws and works well.)
Not the best solution. The mosfets needs pretty big pressure to make good contact with the heatsink. It will bend the pcb.
2 Likes
Not necessarily, the pressure is put in line with the fets… my 2c
Not optimum and should be fixed - but it will have to wait till next version. Anyway, I managed to fix the issue like warloo93 in the image above.
Sorry for the late reply. Ok great 
As i said previously its still only a prototype with material i had at hand. It looks a bit dodgy but hey it works xD
- Pwr board will not be mounted as high. It will be lowered by marked length
- These will be replaced by screws…i just did not have any
- These will be removed. The idea is to place the MOSFETs onto a heatsink/aluminium plate.
- The FreePill will be mounted on machine pin headers, either on this or the opposite side. The whole thing will be significantly lowered.
- Pins and JST housings will be removed. But for JST housings we should probably account for since its how its modelled.
- As we’ve heard theres different approaches. The case should probably support different mounting heights for the MOSFETs
I only bothered to solder a single capacitor. There is one on each side though. The SWD connector you see on the FreePill will not be soldered.
HTH
What about replacing controller phase wire with some stud (maybe copper) and connect the motor phase wires directly?