openinverter forum

I've been involved in electronics for nearly 40 years now and I still make mistakes. Don't be too hard on yourself. In any event its a beta test and you have just tested something I would not have thought to do. I'll hope to have some new boards in a few weeks so no sweat if you need a replacement.

With the help of a friend in the US I bought a "960" inverter so hopefully we'll have some info on the HV side power supplies soon and I can get the design updated and a batch of boards ordered.

I hum and hawd about re using the oem gate drives off the orginal board, the confromal coating remover did not work as well as i hoped. Ended up hist buying new drivers for around 10$ each.(cad) Went ahead and assembled my dev board, and relised I didn’t get temp sensors as I thought I could harvest them from the orginal board! Oops. Everything is looking good so far, once temps sensors are in should be up and running.

So been pouring over some high res pictures of the 960 board this morning and my best guess on the supply to the HV side 5v reg is......they replaced the whole hv derived psu with a single 0805 resistor and called it done!

There are a few other changes I've spotted in the gate drive section such as different resistors vals for the front vs rear etc but more on that when I get my hands on the boards.

On the FDU, can you grab a SPI trace from the gate driver chain as well? We've no idea if they are doing anything different with the configuration.

I will once the boards arrive. I just have pics at the minute.

Looks like the next M3DU boards from jlc will have the gate drivers and transformer fitted meaning only temp sensors , current sensors and main 30 pin connector will need to be fitted by the end user. All through hole.

Jack Bauer wrote: ↑Fri Jan 23, 2026 4:23 pm

Do you have timeline when boards will be available for purchase?

Hopefully in the next few weeks.

Just a quick note folks as I have been getting lots of contacts asking about availability of the design files for this project. I will make them available once the design is finalised. That should be in the next few weeks. I am currently waiting on post from the US of two inverter boards so I can do a little more reverse engineering and verification before ordering the next batch from JLC.

temp sensors arrived! im on the final stages of assembly.

just doing a test point voltage check.

everything but the temp sensors are installed

il1 and il2 are reading 1.57v

tp6 is reading 12.86v (along with the rest of the 15v test points) with the oem transformer ----this seems sufficient ----
tp8 is reading -5.15v

no gate driver faults detected. (led has faint glow)

ye haw!

Back in the saddle. Boards arrived from the US. Confirmed how they do the 5V HV side supply now on the "960" is indeed derived from the low side driver supply. Bad news is I can't get a capture of the spi as this board came minus the gate drivers , transformer and current sensor block so am guessing the Elon brain is upset and not bothering to send config. Will see what I can sort out.

So basically it looks like I can send the V3.2 to JLC for production...

Attached 3d view and pdf schematic. Going to push the button on ordering these tomorrow all going well.

Looks like there's something up with the schematic. Lots of components aren't connected around the DC bus discharge. Has anything changed with the system behaviour that would affect the firmware?

Sorry don't know what happened there. Should be fixed now. No code changes.

Good Day Gents,
Im a newbie in this and I bow to your hardware and controls prowess.
Im working on a boat conversion right now. Im using an ev controls unit and a model 3 battery and drive unit.
Still a work in progress but I got it in the water at the end of the year. It drove but I have a couple of challenges
1. Propshaft speed. Easy mechanical fix.
2. Reverse max speed. Limited by the tesla controls. Might be good enough once I gear it but I was losing a battle to a 20knot wind...
3. Decel to a speed at which I can shift from D to R. It takes forever to spin down to a speed in which I can fo from fwd to reverse. It could be that I was in a battery induced limp mode but I suspect there is some sort of Max decal speed where the controls modulate the regen.

It leads me to a couple questions.
1. For problems 2 and 3 does your pcb open up calibration possibilities.
2. Could I really just be seeing the effects of regen being shut down due to a bms limp command?

Id love your thoughts.

Dave

Usttdth wrote: ↑Sat Jan 31, 2026 12:04 pm 1. For problems 2 and 3 does your pcb open up calibration possibilities.

Yes. It should be possible to fix these behaviours with the OpenInverter board. The operation of the inverter is configurable and is (I believe) being used in boats currently with other motors.

First spin. Very impressed gents, in the end i dont think the technical overhead will be very high for most folks.

First order of ten V3.2 boards now placed. I should have them in around 2 weeks and all going well be in a position to ship by end of Feb. To that end I have updated the webshop and added an option to purchase the full design source files for those who prefer to order their own and/ or contribute to the project.

I know I'm taking a risk making the files available in this way. The last time I tried something like this it did not take long for a genius to realise all he had to do was order and pay , wait a few days and file a bogus claim to get a refund. This is by far and away the biggest opensource project I have ever ran. It would not even be remotely possible without countless people giving of their time and resources so please think about that. That said given the main aim is to make these drive units more accessible to folks and the way the world seems to be going I feel its a risk worth taking. Please don't make me regret it.
Needless (and prob useless) to add these files are provided for personal use only.

As always the schematic and board layouts are freely available on Github.

https://evbmw.com/index.php/evbmw-websh ... el-3-du-32

https://evbmw.com/index.php/evbmw-websh ... v3-2-files

Hello fellow model 3 board swappers,
I have been stuck on a mistake I made a few weeks back, my beta board has blown the fuse from a bad power cycle (overvoltage)and I tried bypassing the 3amp fuse to replace the one I blew from a power up overvoltage but I am still now burning out the new fuse everytime I try to power up. Any suggestions to follow back the next shorted part? Thanks in advance.! And congrats to the others on their successful motors turning!!

As I stated I will ship you a replacement board from the new batch once they arrive and are verified.

In production now at JLC. As you will see the gate drive transformer is missing. Sadly they seem to be out of stock everywhere until March. If I can find a source I'll order and fit them myself for this batch to maintain the "no smd soldering". If not them the end user will need to harvest from their OEM board.

I've screwed up a bit and have only just checked the latest schematic in detail. Unfortunately it's missing fixes for two problems identified early in the beta.

The biggie is the lack of a pull down on the base of Q16 which causes the board to power on but never power off (if the MCU fails to initialise properly, boot loops, etc). I fitted a 10K from MAIN_PSU_EN to ground here.

The other problem is lack of a pull up on GATE_DIAG causing the LED to glow and risks tripping the inverter. I fitted a 10K pull up to 3.3V here (across R265 and D58). I've tested this in normal operation and when shorting out phase MOSFETs to cause fault conditions. The LED lights correctly and the extra current is minimal and well within the spec of the STGAP1AS.

Don't sweat it Dave. Whole host of people could have spotted that and chipped in. I can do those mods on the current ordered board before shipping.