GS450h Buck/Boost Converter control

[xp677]

Also note that this is not the first time I've seen a converter with the appearance of this failure.
...
or I was just unlucky and have had a faulty converter all along!

So has there been another case of this or not?

I saw one for sale with the same damage, from a GS. Which is why I said "appearance of this failure". This gives credit to the thought that it may be a fault with the unit rather than the method - if the same thing has happened in OEM use.

[drprox]

I would be interested to know if anyone has had long term success using this method.

The 330CI project which started off the GS450h trend did use the buck/boost converter for both charging (boost) and for powering the DC/DC (buck). The guy didn't report any long term issues but there was no detail on the method used.

Is there any context to the mods and what they are for? E.g Are we breaking the tracks from the original MCU control to insert our own? What happens if you do no mods, does it just buck down to 288v all the time? (This is what I actually want).

[xp677]

The modification to the PCB is to provide a ground path to the converter control PCB, this was previously switched via the MCU on the inverter board.

The 330Ci project had it's replacement MCU wired to the back of the PCB connector. It would be impractical to find and cut traces in that location. If traces needed cutting, a better option would be to cut and splice into the wiring harness.

Damien is doing a similar control setup on Prius converters right now. The IPM for those looks to be similar to the GS, likely the same/similar control setup.

I am unsure what caused the failure of my unit, having spoken to several electrical engineers about it, it was decided that the most likely cause was one IGBT failing short, allowing a path to ground when the other IGBT was triggered. The cause of this is unknown, the Toyota logic boards prevent this and should have detected the issue. The failure could be as simple as a routine failure of the unit - this was a junkyard inverter, after all.

I know that Damien & others in the Prius "cant kill an inverter" thread was looking into ringing within the inductor, and investigated adding deadtime to the two PWM signals to the IPM, however I don't know if that is relevant for this unit.

What I do know is that the wiring and code posted in this thread does work. For simple bucking to 288V the code could be simplified.

The main issue found with the code is that any cycle which includes serial writes (for diagnostic output) increases the cycle time for the code, which can allow the output voltage to rise or fall further than expected. I'd recommend using minimal diagnostic output during testing.

The important thing to take from this, is that the internal inverter <> converter bus bars should not be used, and instead a separate, fused power feed to the converter should be supplied. My inverter failed short, pulling 350A at 630V, if I had added a 20A fuse, I could have saved the converter.

[drprox]

Thank you for this information, it helps my understanding.

Do you know if the converter requires the main DC link capacitor?

I think it will be easier to supply the converter via a separate fused feed than to try and splice a fuse in-line with the original busbars. How did / would you do that?
I only need around 1kW max, so at 600v we are talking a few amps, though for that voltage the fuse is still not particularly small.


Also, how does the buck converter work under no load conditions - is there a static load built-in to keep it stable or is there a risk of it just rising to full DC voltage and blowing up?

[mfox]

Is it possible phisicaly remove buck-bost converter from inverter (wich is used to run drivetrain). Or both are using the same PCB, so is not possible ... just on diferent sides..
TNX

[arturk]

It is possible to physically remove electronic components of the inverter such as PCB, Capacitor, Inductor, IGBTs.
However both: Inverter and Converter share "cold plate".

Good info and pictures here:
https://www.osti.gov/biblio/928684-eval ... ive-system

I wonder why you ask.
I was contemplating separating both at some point to make Inverter smaller since I have no need for Converter but it seems to much hassle but not impossible.
It is definitely recommended to disconnect HV bus bars between two.

[xp677]

You can easily remove all the converter components other than the inductor (which is potted).

The space left behind is a nice spot for a DC junction box or even just a HV fuse or whatever.

I was thinking of maybe looking at designing an alternate VCU which could mount where the converter hardware used to live, and wire through to the inverter PCB. This would give an "all in one" solution. But then I figured it just changes which connections you need to the outside world, and doesn't really help all that much in the end.

[mackoffgrid]

To complement the work on the GS450h inverter, here is the control method for the 30kW Buck/Boost converter.

I control this converter using a bare Arduino Nano mounted to the inverter control PCB.

The following pins need to be connected (for the attached sketch)

• +5v
• GND
• HV Voltage Read (A0)
• LV Voltage Read (A1)
• PWM High (D6)
• PWM Low (D5)

XP677,

Because of this project / Topic I have been able to fill many gaps in the pinout information of a GS450/Camry breakout board I'm putting together. So I have included your Nano project into the breakout board-
https://github.com/mackelec/LexusGS450_ ... r_Breakout
I hope this in both Correct and interesting.

[xp677]

Nice work!

[TonyV]

I'm about to embark on my EV conversion soon and been collecting some parts but, I've been reading alot about charging EV batteries using the 450h inverter,(and I must say I'm more confused then ever ), question is, to charger the batteries do we need another inverter to charge the batteries or do we use the one that drives the MG1 and MG2 and somehow convert it to charge the battery?
If its the later is there a wiring diagram?

Thanks in advance for your time.

[ggeter]

Because of this project / Topic I have been able to fill many gaps in the pinout information of a GS450/Camry breakout board I'm putting together. So I have included your Nano project into the breakout board-
https://github.com/mackelec/LexusGS450_ ... r_Breakout
I hope this in both Correct and interesting.

mackoffgrid, do you have experience or recommendations to run the GS450h transmission and do charging and dc/dc with the Camry or GS450h inverter?

[Dilbert]

Outlander charger and dc to dc is cheap and readily available, not worth doing a diy if you can source one of them

[mackoffgrid]

I made that breakout for other purposes but added the boost control because I could. I've not had any chance to power this board up because of other projects

So I'm sorry I can't advise, but Dilbert makes good sense - I'd like to get a outlander charger but they are not so cheap in Oz.

[Jacobsmess]

Out of curiosity has there been any more developments with this?

[thomaa]

Really cool project! I'd also be interested in using the converter to charge from a DC source like a LV solar battery. It'd be necessary to implement a current limit.

I see two options

Option 1: add a current sensor of some kind and implement a current control loop to your arduino code.
Option 2: Maybe the converter in its OEM use case already does current limiting, in which case hacking the comms to get it to buck/boost as we please. (altough I suspect the current regulation is entirely handled by the inverter and how much power it is sourcing/sinking, then the converter just "follows" by keeping the bus at its 650v setpoint. If thats the case then the converter hardware doesnt implement current sensing and the idea is out of the window)

Wondering if anyone has tried to crack the OEM communication scheme for the converter and knows more about exactly what can be played with. If that can be done then the comms can be integrated into a VCU like the zombie and offload the switching and control loop stuff to the OEM logic board, which seems like a more bulletproof solution compared to hijacking the IGBT gate drive signals.

[thomaa]

xp677 I've done some more web scraping and found your old thread on the diyelectriccar forums, I didn't realise you were a major contributor to the reverse engineering of the gs450h, amazing!

What came of your rx7 conversion?

Curious how you went from trying to control the inverter via the sync serial interface to hijacking the DC/DC power stage control directly?

edit: I found your other thread that is probably want im looking for viewtopic.php?t=205&hilit=rx7
I'll have a read when i get a chance!