Prius Gen 3 Inverter Logic Board Support Thread

[Bigpie]

Looked back over some of damiens old videos. Hes getting 150 dc amps so either I've more tuning to do or the yaris inverter is less powerful. I forgot, I killed my prius inverter and using a yaris inverter with a prius capacitor. Oops.

[FFMan]

I am attempting to order a replacement board from JLCPCB using the gerber file etc as supplied.

There is an inventory shortage on a few items, most of which can be sorted but there is a component labelled PS1 (SGM3204), which is a negative voltage charge pump which is out of stock. If i can get this elsewhere I can add it later, but for the life of me i can't find it on my dead board to see where it goes.

Can anyone point me in the right direction ?

thanks

[thornogson]

I am attempting to order a replacement board from JLCPCB using the gerber file etc as supplied.

There is an inventory shortage on a few items, most of which can be sorted but there is a component labelled PS1 (SGM3204), which is a negative voltage charge pump which is out of stock. If i can get this elsewhere I can add it later, but for the life of me i can't find it on my dead board to see where it goes.

Can anyone point me in the right direction ?

thanks

this is the device for creating -5v for the current sensors and lives between the current sensors connectors.(IC14)

I have used an isolated 5v device with +5v connected to ground to successfully replace its predecessor

[Jack Bauer]

Do keep in mind that the latest boards no longer use that part but instead the max735. If you intend on building one then thats what I would recommend.

[Pete9008]

Looked back over some of damiens old videos. Hes getting 150 dc amps so either I've more tuning to do or the yaris inverter is less powerful. I forgot, I killed my prius inverter and using a yaris inverter with a prius capacitor. Oops.
Screenshot_2022-09-11_at_10.09.30_pm.png

You've worried me now, if the Prius inverter is that limited on the DC side it would not be good!

I'm hoping to get 300A DC though it, I can't see any reason why that shouldn't be possible but would appreciate it if anyone who has done it, or found that they can't do it, could confirm?

[Bigpie]

The still from the video above it Damien getting 150 DC amps from a Prius inverter. I'll have to get hold of a Prius inverter, mine is currently a Yaris bottom end after user error last year, Id forgotten I'd used the lid and cap of a Prius inverter. Maybe the Yaris is less powerful or there's something wrong with my parameters

[Jack Bauer]

there are no dc limits on an inverter other than those imposed by the user or by the AC side.

[Pete9008]

there are no dc limits on an inverter other than those imposed by the user or by the AC side.

Thanks, that's a relief, do you think 300A DC for brief periods is realistic?

[Bigpie]

It looks to have cutout at lower currents on the graph, unless something spiked and Mr Toyota said now before Johannes gets the info.
2.5 Throtcur, if I'm reading this right, it's happy at about 250 amps then at 380Hz ish Mr Toyota sits down.

Throtcur_2_5.png

4.5 Throtcur, current 100 amps higher, but again at 380Hz ish Mr Toyota sits down.

there are no dc limits on an inverter other than those imposed by the user or by the AC side.

Sorry, I can't get my head around it, probably being dim.

Are you saying for some reason the AC current must be spiking in both cases around 100 DC amps and the hardware protection is shutting it down before the stm32 even gets chance to read it? Graphs are showing (to my understanding) stable ish, but different AC currents.

[Pete9008]

I had a look at your data over here https://openinverter.org/forum/viewt ... 5#p45325 . If you look at the max current rather than the ave, and combine Iq and Id you are getting spikes of up to 500A.

[Bigpie]

Thanks. Do you have a best guess as to why? PID parameters? Note, this is not new since 5.20, I've just been more ginger with the throttle

[Pete9008]

The two reasons I can think of are electrical noise or control loop oscillation (both could cause currents to rapidly vary between say 50A and 150A rather than being a nice steady 100A). The only way to tell is to change something (either electrical screening or kp/ki) and see if the difference between the min and max currents gets bigger (bad) or smaller (good).

The other possibility is a control loop is actually going unstable and the large current created is tripping the inverter.

[Bigpie]

Thanks, think it's starting to make sense. Will do some playing around.

[Pete9008]

Forgot to say, there was also a bit of a regular blip on your motor speed trace that was visible when plotting the max and min values (sorry didn't do a screen grab). Not very big on the speed trace but could correspond to quite a big instantaneous error on the rotor position.

I don't have the experience to know whether it's normal or not but it might be worth giving your resolver connections a once over. If there is a problem with them then that could be causing the current spikes too.

[Bigpie]

Woah you're right.

Think the resolver wiring is pretty solid though, but will investigate.

[Pete9008]

There's always going to be a bit of noise there but it was these bits that made me wonder (only marked a couple but there are several in the plot):

It just seems off that on each of the blips the max increases and the min decreases. It kind of suggests that the controller saw a sudden step in the motor position (the reading before the step gives the maximum and the reading after the minimum - or vice versa depending on whether the step is forwards or backwards). If the controller does see a step change where there isn't one (due to noise) it could lead to odd voltages being driven onto the motor, odd currents being produced and possibly a trip?

The blips look fairly periodic and the period doesn't seem to align with the trips but then we might not be seeing all of them due to the limited sample rate on the graph.

Just a thought but might be worth checking if everything else looks ok.

[Bigpie]

Those seem quite insignificant next to the yellow one. Those 3 close sharp downward spikes must be the cutouts I'd say though.

[Pete9008]

I'd assumed that the orange one was a cutout too, if it wasn't then you're right it's much more worrying than the blips!

Edit - yellow not orange!

Edit2 - did you dip the clutch on the cutouts, did the motor speed really drop to zero or is that the inverter restarting?

Edit3 - could the yellow one be some sort of plotting discontinuity? The values each side don't seem to match up?

[Bigpie]

No clutch. Straight coupler. Inverter cuts out and I've set auto restart, so just lifting off restarts. Next time I go out I'll get a log with just one cut out just to be sure.

[thornogson]

Thank you - that would be useful before i go to the expense of replacing it. I'm in Bicester but can post it to you. Another set of eyes on it would be helpful in case i am being a muppet somewhere.

Can you PM me an address and I'll find some sensible packaging. I'll probably dpd it given the post is intermittent, so if you cold include a mobile phone no that is helpful as they'll text when they will deliver etc.

Alex

@ bigpie @FFman I'm having exactly the same problem with a recent blockIV auris board , was flashing when I got it and after I attached the connectors etc and tested voltages, for some reason I had to reload the bootloader firmware , set up the STLink with a fresh install of stm32cube on my Windows 10 machine and the latest stm32loader.bin from Johanes github. Stm32cube tells me the write was successful and it looks just like the read from a known good board but no blink... LED is good , If you find something , please publish it here !

[Jack Bauer]

Please use .HEX files with stlink NOT .BIN

[Bigpie]

If the LED isn't blinking the firmware isn't running. First thing to do is try flashing it again.
If you didn't tell stm32cube to skip erase when you flashed the bootloader, you'll likely be missing the firmware.

[Bigpie]

I'd assumed that the orange one was a cutout too, if it wasn't then you're right it's much more worrying than the blips!

Edit - yellow not orange!

Edit2 - did you dip the clutch on the cutouts, did the motor speed really drop to zero or is that the inverter restarting?

Edit3 - could the yellow one be some sort of plotting discontinuity? The values each side don't seem to match up?

log (4) (1).csv

(57.47 KiB) Downloaded 86 times

This is a log from a few days ago, with only a single cutout even.
I think the cutout happed at the massive spike in amp. Theres 2 long flat spots, I stopped briefly twice, the cutout occurred after moving off for the second time.

[Pete9008]

Have had a look at this. On the current traces:

You can see there is almost a factor of two between the min and the max in places, don't have much experience looking at these plots but that seems too much to me?

In the region where you had the trip there is nothing obvious:

But looking at the voltage amplitude values:

It's pretty clear where the trip was, the amplitude was driven to a very high value which would have cased a current spike which tripped the inverter. Not obvious what caused it though.

I still think it is either an electrical noise problem or a loop stability problem, I'd probably err towards the later as I've seen oscillations like this on some simulator runs:

If you image what these would look like if you just had the min, max and ave for each 0.6sec period it might look pretty similar to your plots. Worth having a go a tweaking the gains to see if it gives an improvement, if not then look for noise problems.

[Bigpie]

Thanks. Hoping to get a look at following your PI tuning strategy tomorrow. Will post updates.