Toyota Gen 3 inverter charger

[celeron55]

I think the practical charging use case for the gen3 inverter is 1/3-phase charging for about 260V nominal batteries. If the battery voltage gets higher than 300 or so volts you only get 3 phase charging, which is not ideal. Also it's considerably more than what the converter was designed for.

I think for 350V nominal you should probably be using a real 1-phase AC charger and ideally take advantage of the high voltage by adding fast DC charging.

[postmann2002]

How much AC power do you want? Maybe a Tesla PCS would be a better solution given it already has the bidirectional functionality;

I wish 11kW or more... maybe 22kW charging-power.

Yes, Tesla PCS is an option... but now I have the Prius Inverter on the bench and Tesla PCS is to expensive for me...
I will continue tinkering a little bit with the Prius Inverter. Last weekend I managed to drive a small BLDC-Motor ...but uncontrolled, just a 3 Phase Sinusoidal variable frecuency... and it spinns. Next step is control the motor with hall-sensors conneted (FOC).
Then I will tinker with the big inductor and the corresponding switches for chargingfunctions in the future.

[rstevens81]

This thought came to me ...
If we were to use a "240v to 110v site isolation transformer" (they are used in the uk on building sites to drop 240v to 110v for safety)
this would allow us galvanically isolate from the mains produce 110v rms with would be circa 150v once bridge recified.
Therefore the diagram would go like ...

Mains (240v) --> Transformer (110v) --> Build our own bridge rectifier (150v DC)--> Original DC input -->Buck Boost -->Battery (Via J.B revised connection Point). This would still give us a relatively cheap slow charger.

Any thoughts as I'm a mechanical engineer /Mathematician by trade although I have an understanding of electronics (I understand V=L di/dt ) I'm not an Electrical/Electronics Engineer and maybe missing something fundimental

[arber333]

Lets not get carried away with transformers. Such electromechanical components would be heavy and specialised for a single application.

Instead i propose we connect HV battery (360Vdc) directly to the original inverter input. When we connect inverter to charger we get buck/boost path through inductor which can act as a charger.

Now to provide traction in the same configuration we would be limited with what buck/boost stage can provide. Since Damien can attest for 500Arms with the gen3 MG1 power stage buck/boost converter stage becomes redundant and even a hindrance. So to circumvent that i propose that we implement another DC contactor which would provide + HV supply directly BEHIND boost transistor onto the main cap contacts. This along with permannent original connection to the main HV battery. My idea is on the pic.
If we use additional contactor rated at 500A we covered the power section rating with overhead. There are some conditions though:
1. This contactor should be controlled by STM32 traction circuit NOT the charger circuit.
2. There should be a control circuit to inhibit charging and AC contactor trigger in case this contactor is on, for obvious reasons.
3. Likewise there should be inhibition from charging circuit to this contactor turn ON to prevent KABOOM event. Maybe a small DPDT relay which would route signal to both AC and DC contactors, or not...

Advantages... additional ability to charge either from 1 phase or multi phase, of course with main task of moving the vehicle.
Disadvantages... one more contactor... one more possible failure point..... Damien?

Comments anyone?

[clanger9]

That's exactly the architecture I had in mind. That way, you can get a nice power factor when charging via the buck/boost converter, but bypass it for traction mode when you need full power.

If you want to get really clever, you could open the contactor when cruising (assuming running a different DC link voltage would be more efficient), then pull the contractor shut when you need full power...

Main question is galvanic isolation. We don't have an isolation transformer anywhere, so how practical/feasible/safe is it to guarantee that the entire HV system is able to float free of earth with negligible leakage to earth when charging? Will this mean any extra precautions for the installation and wiring? More than a few mA leakage anywhere will risk tripping the AC input...

[arber333]

That's exactly the architecture I had in mind. That way, you can get a nice power factor when charging via the buck/boost converter, but bypass it for traction mode when you need full power.

If you want to get really clever, you could open the contactor when cruising (assuming running a different DC link voltage would be more efficient), then pull the contractor shut when you need full power...

Main question is galvanic isolation. We don't have an isolation transformer anywhere, so how practical/feasible/safe is it to guarantee that the entire HV system is able to float free of earth with negligible leakage to earth when charging? Will this mean any extra precautions for the installation and wiring? More than a few mA leakage anywhere will risk tripping the AC input...

Yeah... well i am using EMW charger 3phase derivative for 5 years now and i only managed to trip FID once. But that was due to exposed contacts and moisture on connector. When i properly isolated connector everything was ok. Really what you need are a set of AC fuses in a car that are able to trip on OC event. Just now i charge with 16kW on air-cooled 3ph charger and every EVSE is ok with it.
On 3 phase of course. At the instant you use it with single phase you start seeing and hearing trouble.
I intend to have separate 3kW charger for single phase charging, because it is more efficient overnight. I assume for quick charging on 22kW EVSE even ZOE doesnt use isolated charger. I mean i can hear its buck stage ringing the same Hz mine is...

EDIT: Also i wouldnt go about switching contactor in mid drive. I am not sure if it is a good idea at all...

[SciroccoEV]

https://www.bender-uk.com/solutions/emobility

[clanger9]

Arber, any particular reason why it doesn't work on single phase? It ought to (as far as I can tell).

It's true that the (now discontinued) Renault Zoe Q90 "chameleon" used a non-isolated charging system. They don't have a separate buck/boost stage, but use the motor as an inductor on the input to the traction converter. Which is why they buzz like mad!

Renault always struggled to get the leakage current down to something acceptable. Minor tingles & shocks from the car body are common. Nothing dangerous, but it shows how tricky isolation can be...

Interestingly, Renault just used a full-wave rectifier on the input. So to mimic this, the Toyota MG2 inverter could just run as a rectifier (no PWM), and control all the charging via the buck/boost converter only. Simpler?

[arber333]

Interestingly, Renault just used a full-wave rectifier on the input. So to mimic this, the Toyota MG2 inverter could just run as a rectifier (no PWM), and control all the charging via the buck/boost converter only. Simpler?

Well this is what i intend. I will just use transistor diodes to rectify AC and i will get 600Vdc on main cap. then i intend to buck down to 400Vdc with 20kW from EVSE. Liquid cooling should let me charge up to 20kW.
Single phase works just fine, but its PF is off. Maybe some 75% efficiency if every planet is lined up...
I dont boost up from 320Vdc though. Rather i use voltage doubler to get 650Vdc from single phase. It works for cca 2.5kW then fuse starts to heatup.
Maybe if we would boost up we could get better PF.

[clanger9]

Another crazy option I've been thinking about is DC charging via the buck/boost converter. This would ultimately allow me to add CCS to my bike (which by necessity will have a battery pack voltage below the minimum CCS level).

NB: fuses & contactors needed on all AC & DC charging inputs for safety - I just haven't shown them on the diagram.

20-30kW AC and DC on-board charging would be a game-changer for a bike; current designs mostly max out at 6kW AC due to space and weight constraints...

[celeron55]

Main question is galvanic isolation. We don't have an isolation transformer anywhere, so how practical/feasible/safe is it to guarantee that the entire HV system is able to float free of earth with negligible leakage to earth when charging? Will this mean any extra precautions for the installation and wiring? More than a few mA leakage anywhere will risk tripping the AC input...

You have to have an AC input contactor/relay when the charger's input is not galvanically isolated from the rest of the HV system so that the charge port is isolated when no cable is plugged in. So you do need yet another contactor on the input side. An AC precharge circuit can be avoided if the inverter main capacitor is precharged from the battery side using the boost converter before activating the AC input relay.

Skipping that seems risky to me. Instead of being always two faults away from getting a HVDC shock, you would be only one fault away from that if you happen to be touching the non-isolated charge port.

[clanger9]

Yes, contactor on the AC charge port absolutely needed! Fuses too. Same goes for the CCS port - I just left them out of the diagram for clarity. Have amended my post, above.
Good point about pre-charging the DC bus before closing the AC contractor. Likewise for DC charging. At least all that can be done in software...

[arber333]

I just use 3x 330R resistors across AC contactor and charger waits for 20s. Then it is all set. I only use N line to run the fans and 12v power supply.

[celeron55]

330 ohms doesn't protect a human from high voltage.

[arber333]

330 ohms doesn't protect a human from high voltage.

No for that I use contactor on DC side on the same trigger than AC contactor together with output diode. Also it solves another problem.
In early EMW iterations you couldn't just unplug the cable from EVSE. PWM would collapse and energy from inductor would wreck havoc on the igbt. Now when I unplug EVSE cable AC and DC paths are cut at the same time and nothing bad happens.

[johnspark]

i have been doing a bit of a desk top study on charger circuits connected to the mains, single or three phase. i have found there have been a few papers on using a current proportional-resonant control strategy to reduce second-order harmonics being presented back to the utility via the power point. The paper i have discusses single phase charging, and claims to get harmonics down to 1.92% using this PR control strategy for a 3kW charger simulation.

Sounds complicated but is not really, just an algorithm to go into the processor. What is required for single phase is an extra inductor to be connected to the power point, using 4 IGBTs for switching, not just 4 diodes, (which we have for free already), voltage monitoring of the 230V ac, and current monitoring of the same 230V ac. This arrangement also reduces current ripple going to the lithium batteries.
For 3 phase, we require 3 inductors, voltage monitoring for 2 or 3 phases and current monitoring for 2 or 3 phases. Harmonics produced with 3 phase charger can then be made lower again as well.

if anyone is interested, will post a paper here.
Paper talks about IEEE 1547 standard, which is about: (Standard for Interconnecting Distributed Resources with Electric Power Systems).

[clanger9]

That sounds really interesting. As you say, we have IGBTs available, so if we can fire them in a way that reduces harmonics then all the better...

[arber333]

I have seen EMW using 4 IGBTs with 2 Inductors. And they have had problems with losses up to rev. 4 firmware when they introduced PID control which dampened the oscilations before inductor. I guess you could say it was a sort of algorithm. Now i use this same firmware without boost stage transistor on 3phase, just buck.
This is why i think inverter could be used primarily as 3phase charger. Otherwise you must think of an algorithm to interlace buck and boost transistors to get correct current output. Motor stage transistors are only used for their diodes as rectifiers.

[johnspark]

i think the way it can be done is use extra inductor/s depending upon 1 or 3 phase with an existing motor IBGT set. Use this to set base voltage on motor IGBT DC bus, using the PR control strategy. Then use buck boost IGBTs with inbuilt inductor to change this fixed DC voltage to the voltage the lithium batteries require, using techniques developed on this forum. That way the goal is segregated into two distinct independent goals, which makes the overall solution simpler and quicker to achieve, i believe. PS the solution i saw used PR and PID together.

Inductors have to be carefully designed. Need to consider peak, and average DC that inductor must pass, and same for AC. I have quite a number of inductors to experiment with down the track.

for the file attached, i got this from researchgate. i believe i am allowed to post because i am not making money on it.
Second file is sample c code i found for the PR part.
Also, there are references in the back of the paper.


here is a youtube explaining how to design one type of PR system.

Kind Regards,

PS: I am very interested in this theory so will be working in this direction for charging.

[rstevens81]

Am I missing something? The creater of the forum(johu) has already given us the solution to single phase charging, it's even on the website .... Use the motor as an inductor, then use his charging function of the openinverter firmware (pf 0.93) to boost it (say 400v) then use buck converter to get it to the voltage we want. ( I know not as good harmonics but the parts count is kept low)

I realise this comes with the obvious disclaimer this is mains voltages so you need to give it the necessary respect. ( Can bus AC charger is infinitely safer for only a few £££, this would always be the advised option)

Note: Edited to remove reference to chademo firmware as it is built into standard firmware as if release 4.85r

[Leo M]

Sorry, is this how the charger works on the Gen3 Board or not? Someone make a video on this issue please...)

[arber333]

Am I missing something? The creater of the forum(johu) has already given us the solution to single phase charging, it's even on the website .... Use the motor as an inductor, then use his charging function of the openinverter firmware (pf 0.93) to boost it (say 400v) then use buck converter to get it to the voltage we want. ( I know not as good harmonics but the parts count is kept low)

And you need to keep the motor from wandering off. Remember motor is PM and you are moving current through windings. That should create some torque with the magnets. Your car may not be on the same parking lot when you finish charging .

[Bigpie]

Sorry, is this how the charger works on the Gen3 Board or not? Someone make a video on this issue please...)

The charger code for the Gen 3 uses MG1 to rectify mains (code is hardcoded for 3 phase, but a few simple changes can hard code if for single) then this is buck'd down to a set battery voltage.

I've played a little with it on single phase, and blew my igbt for the buck/boost and haven't looked since as I'm still working on the car.

Single phase (in the UK) rectifies to ~320VDC so your full battery voltage needs to be below this or maybe boost the input or rectify your self and change the code to boost.

As far as I'm aware, only Celeron55 has used it to charge, as far documentation, at the moment you're on your own on that front, there's a fair amount of comments in the code but there's no step by step guide. If I end up using it and getting it working I'll write something up, but it'll be a while.

[rstevens81]

I have a plan what I'm doing going forward, which consists of the following options (depending on how successful I am):

1. Use a (spare) inductor on the secondary circuit (probably bodge an stm32 blue pill into the existing board) then buck using celerons code.

2. Leave the car in neutral and use the motor, depends how noisey it is

3. Use voltage doubler to get over the 320v thing, although power factor is terrible I love the simplicity. And it begs the question do we actually care that much about power factor on a DIY conversion as used OEM charger can be had for circa £300 if you are that worried.

First test will be using a transformer from a broken car battery charger to get a suitably low AC voltage for testing. Then use the motor to boost using the open inverter boost charger function, then the buck converter to lower the voltage to charge a 12v lead battery. This way I can test the principal and get it working (or not) before even thinking about high voltage.


This is all a bit of a way off at present as I'm having adventures in can bus

[Jacobsmess]

Would the information here also relate to the lexus IS300H inverter which is also a gen3 inverter?

I foolishly bought one because it was cheap and my finger was faster than my head and not I'm wondering what to do with it....