Investigating trips

[FFMan]

I had another post going related to trips of the prius inverter (V1C) that i was experiencing every now and then. It only seemed to happen first thing in the morning, leading to the the theory it was perhaps condensation. I garaged it a couple of nights but results were inconclusive, one morning it worked, the next it didn't.

The problem seemed confined to first drive of the day and once going was ok, until Thurs when after a few starts it was running, then i lost drive 100 yards from the house. Managed to get it home after a restart or three, seems like problem was getting worse.

So i decided to pull the inverter and see if was damp, or if the board had any corrosion, but all looked ok. I had previously cleaned and sealed board and it looks ok, clean and shiny etc

As i could see nothing amiss, i transferred the V1C to my spare inverter, but once reinstalled, had no drive, and the inverter is showing a UDC of 35 on pack of 380v. I pulled it again to check all the internal fittings, there is an easily forgotten pair of bolts inside. Nothing amiss.

So i transferred the v1c back to the original inverter but now i see the same symptoms. 35v showing on inverter, and not surprisingly no drive.

The only thing in the back of my mind is that it is possible the 3 cables from inverter to motor have been mirrored, as i removed some of original casing in order to get a better seal, and in doing so lost my alignment marks. What would the effect be of mirroring the phases, though i think this is secondary to inverter showing 35v ?

Any thoughts on possible issue, where I should look next ?

[jrbe]

The only thing in the back of my mind is that it is possible the 3 cables from inverter to motor have been mirrored, as i removed some of original casing in order to get a better seal, and in doing so lost my alignment marks. What would the effect be of mirroring the phases, though i think this is secondary to inverter showing 35v ?

Any thoughts on possible issue, where I should look next ?

Mixing up the 3 phase wires to the motor won't give you 35v on what should be the high voltage battery.

Before you go nuts, check your grounds.

Can you very carefully check the voltage at the inverter to verify what the voltage is there? If it's ok move forward to the next one. If its low, keep checking each connection / component back towards the battery.
If that seems solid into the inverter, next up would likely be checking the voltage divider circuit.

[FFMan]

yeah i think first step is check the voltage input to the inverter and go forwards or backwards from there. could be a contactor fault etc.

[FFMan]

ok - so some early morning tinkering:-

1) with no HV the inverter registers about 32v, so i think my problem is no HV rather than some,
2) there is full pack voltage on the input terminals to the inverter
3) I have a neon across the HV output in my contactor box. If i close and then open the contactors without the inverter connected the neon goes on and off in a couple of seconds, if i have the inverter connected, the HV takes several minutes to drain away, leading me to believe the capacitor in the inverter is connected and therefore HV is present in the inverter
4) I get the same symptoms on 2 inverters using the same V1C board

I've triple checked all the usual suspects like not bolting up the internal contacts that join bottom part of inverter to top etc

Question:How does the V1C board 'see' the HV. Is there some internal comms to the Toyota boards, or is there a physical connection ?

Not sure where to go next on this one.

[jrbe]

ok - so some early morning tinkering:-

1) with no HV the inverter registers about 32v, so i think my problem is no HV rather than some,

It's a bit hard to follow your description but it seems like the openinverter board is referencing the high voltage reference correctly, it's just not getting above 32v / 35v at the inverter. If this isn't the case, please clarify voltages at the inverter in different states. From this info it sounds like the pre charge and the main contactor are not making connection for some reason.

It might be worth doing a voltage drop test across your main contactor. But you need a way to add a load inline at the inverter (carefully!) Could be a few incandescent household light bulbs (heat bulbs?) in series that fit your pack voltage.
Set your meter to volts dc, red lead battery side contact of high voltage contactor, black lead to the inverter side hv convection of the contactor. What's the voltage it reads when you think the contractor is closed (sending power to the inverter)? If it is indeed closed you likely have a high voltage reading on the meter meaning bad connection or bad contactor. The higher the voltage at the meter the higher the resistance at the connections between the meter probes.

If the contactor is open you need to look into why it's open.

Voltage drop test - The tricky thing about using voltage across a supply to verify that connections are good is that it can be misleading. A 10k resistance in line will show correct voltage, but as soon as there's any load it will quickly drop. Doing a voltage drop test can identify bad connections / bad components. But you need a load for it to work!

[Ev8]

I’m not gen3 Prius inverter expert but I believe others have had issues with the funny multi pin board connector

[FFMan]

On mobile.

Contractors are fine. I’ve measured full hv voltage on them and on the feed into the inverter. Plus the charger is on the same terminal posts and sees full hv. Therefore pretty confident hv is in the inverter just v1c board can’t see it.

Re multiway connector, yeah it’s fiddly but can’t believe hv is present there hence question about how v1c board sees the hv, does Toyota boards tell it, or does it measure it ?

[johu]

Toyota board delivers a low voltage version of the HV on that small connector.

[jrbe]

On mobile.

Contractors are fine. I’ve measured full hv voltage on them and on the feed into the inverter. Plus the charger is on the same terminal posts and sees full hv. Therefore pretty confident hv is in the inverter just v1c board can’t see it.

Re multiway connector, yeah it’s fiddly but can’t believe hv is present there hence question about how v1c board sees the hv, does Toyota boards tell it, or does it measure it ?

From the wiki, it seems the DC high voltage info is a mystery.
https://openinverter.org/wiki/Toyota_Prius_Gen3_Board

Have to trace it down in the schematic. Unfortunately the schematic is an image, no searchable text.

https://github.com/damienmaguire/Prius- ... ematic.pdf

From there,

Verify that connector, wiring, and solder connections are good on your drop in pc controller board. Verify any ground connections on that connector are good as well. Looks like 11, 12, 18, & 22.
If that's all good you'll need to find where that HV_CON connects in the schematic. I looked on mobile for a bit and can't find it. It not being searchable and being huge makes it a challenge. I didn't see a voltage divider on the board for high voltage sensing. I would imagine Toyota kept all the high voltage stuff off the low voltage controller board but I don't know enough about these to say for sure.

Johu's post update, check the connector mentioned above.

[Mhs]

Afaik Hv_con is the for a contactor (controller by the atmega328p) that bypass the buck/boost side of the inverter if you are using the inverter for charging, read further in the link you provided, sorry I'm on the mobile I would've shared screenshots of what I mean.

[FFMan]

Afaik Hv_con is the for the a contactor (controller by the atmega328p) that bypass the buck/boost side of the inverter if you are using the inverter for charging, read further in the link you provided, sorry I'm on the mobile I would've shared screenshots of what I mean.

ok that makes sense - i'm not using buck/boost.

Still wondering how the inverter sees the HV. I know from checking my logging from before it used to see full pack voltage on udc.

I'll check all the connectors, maybe even reflow them with a soldering iron but likely tomorrow eve now.

[FFMan]

ok - so progress of sorts.

I removed the V1c board (again) and reflowed solder on the internal ribbon connector and now the inverter is seeing pack voltage.

BUT still the constants trips. I switched to the spare inverter, identical issue.

I am wondering now if the fact it happens in the morning is related to the fact that this would be the highest voltage point, and it might explain that once going i don't see trips again that day. Is there a parameter I should adjust (mine attached) maybe.

It's currently at 3.93v/cell so pack voltage of 377v.

[johu]

Just wondering, the Prius inverter boards don't have the over current comparator so they can't trip from over current. Only from the fault signal supplied by the power stage can cause a shutdown. So maybe that is dodgy as well.

You can compile a custom firmware with line 406 commented out (enable_break())

Then there is no way for it to ever trip, except the power stage really faulting out and locking the PWM.

[FFMan]

thanks Johannes.

By way of a further update, i pushed udcmax to 435 and whether related or not managed to get the car out of the garage ok (bit stuttery/jolty), and then did 5 mile trip to burn some voltage off without incident, all smooth. I am getting more convinced the morning trips are/were voltage related. I'm going to let it sit a while and then do another run, and then tomorrow i only have to go to station so won't charge it overnight and see how it is in the morning.

> Only from the fault signal supplied by the power stage can cause a shutdown. So maybe that is dodgy as well.

when you say dodgy, do you mean a fault specifically on mine (seems unlikely as i have tried 2), or by using the inverter at this voltage (I don't recall what it runs at in the Prius) it's gone beyond what works reliably ?

I don't have capability to compile STM code, beyond my knowledge and I'm all PC based which i think makes it harder ?

[johu]

when you say dodgy, do you mean a fault specifically on mine (seems unlikely as i have tried 2), or by using the inverter at this voltage (I don't recall what it runs at in the Prius) it's gone beyond what works reliably ?

Dodgy as in cold solder joint on your Damien board

[Bigpie]

Think i've got a spare gen 3 board if you want to try it.

[FFMan]

Think i've got a spare gen 3 board if you want to try it.

that would be really useful - process of elimination. I'll email you an address. Let me know if you want me to send a courier for it, save you a trip to the post office.

I forget what voltage do you run ?

[Bigpie]

I charge to 398V. I'll dig it out this afternoon.

[FFMan]

I charge to 398V. I'll dig it out this afternoon.

thanks

Interestingly from the param set i last had from you, your udc max is 520, where mine is 435.

Not entirely sure what that affects but given by issue may only occur on a fuller pack can't help wondering.

[FFMan]

Ok so I think I debunked my charge related theory as this morning short trip to station. Didn’t charge it overnight so set off on 3.8v

Got a trip 2 miles in accelerating gently. Restarted, got immediate trip then after another restart it’s ok

I now think something is just plain faulty as I’ve done approx 1000 miles without incident before.

[Bigpie]

I've not got the little white connector on the spare board, I'll have to soldering it on then I'll get it sent to you to try.
Have you got any logs from during a cutout?

[FFMan]

thanks

funnily enough i came across that white connector new in packet on my desk at home yesterday, do you want it sent to you to save you buying one.

I will have one of my logs from the trip, car is at the station overnight, i'll retrieve it tomorrow and attach here.

[Bigpie]

If you don't mind, I do have one somewhere but seem to have misplaced it.

[FFMan]

Attached is the log file from the trip on the way to the station. This was unusual as it wasn't at home after a charge, but rather, i stopped at the kerb to let a car out, and then when i went to accelerate it tripped. You can see the error flag is set at row 2832

Thinking about it, it always trips on a light throttle, once on the move it doesn't do it, if its going to, it is always pulling away. I did wonder if there might be a throttle pot issue. I graphed it in the web interface and it looked smooth but there wasn't much granularity. Should i raised pot min a bit perhaps ?

[FFMan]

ok - so some progress on this one - can't say i understand quite why but...

I have spent days on this, inverter in and out multiple times, and i tried a loan V1c boards (thanks bigpie) but to no avail. In fact the problem is worse than ever, complete failure to pull away at all, rather than the occasional morning trip it started with. So i put it all back together to what I knew and could replicate the issue, and then prompted by bigpie started dialling back some params.

I tried cutting thotramp from 5 to 2 but this made no difference. So i went back to the tuning time when helped by pete9008 i was upping currki/idkp/iqkp until it tripped and then dialling it back. The values i have been running for a while are 5200/80/80

Going back a few steps to 4000/25/32 it started driving and didn't trip until after a mile or so when i gassed it hard. So i dialled curki back to 3500 (no idea really what these do but retracing earlier steps), and drove another 3 or 4 miles without incident. Performance feels about the same.

Is the problem fixed or is it just the intermittent nature of it. I tend to think i'm barking up the right tree now as it was tripping every time before i made the changes, and then it was ok.

So the question remains, having had like 1000 miles of faultless driving on my params, what has changed. I ruled out addition of second pack (by trying on 1), and ruled out my new contactor box by bypassing the contactors. Could ambient temp make a difference, certainly looking back it went from occasional trips to near constant over a period of a month. Had i previously tuned it close to the edge, and a change in ambient or humidity was enough to tip it over the edge.

anyone else with my setup care to share some params ?