New version 84V/200A VESC motherboard, using INA241! testing!

Yes, as the title indicates, I am testing the new VESC motherboard, and the test has been basically completed. In this post, I will only share my process of designing the motherboard, without attaching other links (after the company sells the motherboard, he will provide the schematic diagram Or the PCB files may be released).
I named this motherboard 84200HP (because the design was based on Benjamin’s 75300, I originally planned to name it 75200, so there was a silk screen of 75200 on the PCB, but obviously 84200HP is more suitable for it). The following are the current parameters of the motherboard:

  1. Use INA241A2 high-precision sampling chip;
  2. Use 12 MOSEETs (100V/380A);
  3. Use an aluminum-based PCB board with 5oz copper thickness and 6 4*2 copper strips;
  4. Use 6AWG wire;
  5. Use aluminum alloy shell (under design);
  6. 3S-20S, MAX 200A!
    (In fact, all the components used have a withstand voltage of 100V (DC-DC, MOS, capacitors, current sampling chips, gate drivers), but for safety reasons, we use SMCJ85A TVS for protection.)

Motherboard functions:

  1. Built-in Bluetooth module (test completed);
  2. With AUX interface (test completed);
  3. With DC-DC controllable switch -Or shutdown(test completed);
  4. Equipped with BMI160 attitude sensor (interface reserved, testing in progress);
  5. Phase filter, current filter.

The motherboard has gone through many modifications during the design process. Below I share some photos of my motherboard (happy! This is my favorite part):
First version:

Second version:

Third version:

The fourth version (I haven’t gotten the board yet):
I’m ready to test the BMI160 attitude sensor!

Who knows what I’ve been through:

And the shell:

Tell me, what else do you want to know?

9 Likes

do u work for makerbase

4 Likes

Would you consider screw terminals for the leads, given how much of a pain it is to solder 6 AWG? Crimped wire termination to screw lugs would be much more achievable for attaching to the board, I’m not super knowledgeable about their vibration handling but it seems much easier

At peak power with <1% losses in the controller it’s still >150W to dissipate, have you done any thermal testing so far?

4 Likes

It’s great to see you guys sharing the design progress on this ESC. I also saw that Makerbase recently merged a bunch of hardware configs with Vedder’s main branch, which is very helpful.

So did you also design the 75100 and 75200 aluminum pcb versions too? Are you an independent designer or working directly for Makerbase? Did the 75100&75200 alu designs get sold to Flipsky? Did they make the 75300 on their own? Sorry for all the questions, but I’ve been working on the hardware configs for the 75100 and 75200 controllers for a while and haven’t been able to piece together the whole design origin story.

I’m excited to see a 20s phase shunt ESC finally be commercially available thanks to the INA241. Also with 3 temp sensors and phase filters! The thermals on the 75100/75200 alu are really good, some of the best in fact, so I’d imagine this will be a great performer.

A few questions/suggestions:

  • The DC-DC switch… so is this basically to shut off power to the STM32 and other chips so it can stay plugged into the battery without draining it? What’s the current draw with it on vs off at 20s? BTW for terminology you can use disable instead of disenable… or maybe just call it power on/power off if that’s it’s intended function? Or power on/sleep?

  • The 75200 alu (and it appears this design as well) doesn’t have ceramic dc link caps on the power stage - do you plan to add those here or did you find them not to be necessary?
    image

  • How does the gate ringing look?

  • On the 75200, the decoupling capacitors near the STM32 were placed a bit too far away and it was potentially messing with the motor’s voltage and current readings. An example (the yellow and black lines). Might be worth reviewing on this design too?

  • How is the noise on the 5v and 3.3v rails?

  • From a mechanical attachment standpoint, the 75200 had added epoxy to the standoffs between the power and logic PCBs which is a decent idea… but maybe consider the ability to screw the logic board down? There is a lot of vibration, and it can sometimes be tricky for the fab to consistently solder those standoffs to the power board especially with 5oz copper and an aluminum backing. I have seen these fail by lifting off the power board on other ESCs. Might need thermal reliefs built into the traces as well.

  • +1 for posts instead of wires. 6AWG is very difficult to solder into the XT90 connector, and posts like the the MakerX G300 can make it easier to work with.

  • There has been a lot of progress on the VESC Express which is an ESP32 based bluetooth module, compared to the NRF52 based ones most are using. Not necessarily recommending that, but maybe look into it to see if it makes sense over the NRF52 which isn’t really being updated anymore.
11 Likes

I am very happy to receive such good suggestions, and I am also happy to discuss the development of VESC with everyone.
Regarding my identity, I am indeed an employee of Makerbase, but here I prefer to be an independent designer. As I said at the beginning, I will not attach additional links to the discussion in the future. Because this is not the job I should be doing.
The purpose of my coming to the forum is also very simple, which is to collect the functional requirements of VESC and the content that needs to be improved in the design. For example, the board I designed has functions that need to be improved. I will work on it in the next version or redesign it. I think this is What I should do as a designer.
I’ll answer other questions below:

  1. The shutdown function was actually implemented by Benjamin in VESC MK5. Its function is to turn off the power of the control end. For example, the 84200HP motherboard can be seen more intuitively (because the control board and driver board are separated). When it is turned off After controlling the power supply of the board, there is no power loop with the battery, and the theoretical consumption is 0 (currently I can only display values >10mA, and I have not observed the current);
  2. Ceramic capacitors are not unnecessary, but are limited by the structure, and our tests found that ceramic capacitors are more effective for small currents, while for large currents it mainly depends on electrolytic capacitors, and my design is more concerned with large currents;
  3. The 75200 line has been readjusted, and it is expected that the company will sell a new version in the next two days (of course, I am not sure about the time, because it has nothing to do with me);
  4. Regarding 5V and 3V3 noise, I will upload pictures of the oscilloscope after a while;
  5. Regarding the design of mechanical mechanism, I would like to have a more in-depth study with you, because this method is indeed unstable at present. Can you explain it in detail?
  6. Regarding the screw terminals, I have suggested it to the company before, but it was not accepted. The reason given is: the patch screw posts will produce a large force to pull the copper skin during the screw tightening process, which can easily cause copper skin. came off on the PCB.
  7. It will take me some time to understand the design requirements of the Bluetooth module.
3 Likes

Currently, the 84200HP has not been thermally tested, but the aluminum substrate selection was taken into consideration during the design. The current material selection is 2W/(m·k). Do you think it is necessary to use better materials?

1 Like

Yes, I will not deny this issue. But please don’t ask me how to sell. I don’t think any company’s sales are linked to designers.
If you mind my identity, I will not attach any information about the company in the future.

i did not say it was an issue, but an employee of a company should always be upfront about it. its not hard to start ur thread like this:

hey people, im a hardware engineer for makerbase, and im currently testing new vesc motherboard…

2 Likes