Toyota Gen 3 inverter charger

[arber333]

I decided to start and make a charger with Prius inverter.
I will use a complete inverter with casing. It will be water cooled. as intended.
I will connect 3phase input directly to MG2 (MG1) RST contacts. Reverse diodes inside transistors will act as rectifiers to make 580Vdc from 380Vac. Precharge will be made with 3x 20W 220R resistors across contactor pins. Voltage usually rises up to 500Vdc in 20s. I will put some delay in software.
I will use BUCK configuration from AC towards DC side in the same configuration as former EMW DIY charger. I will not use PFC control (no use since it would complicate things much). Main chip will be Arduino Nano which is in every way the same as pro mini of old.
I decided i will not use original current sensors. They are located on MG lines that will be my inputs. I would need to calculate PWM difference etc... Rather i will put one Tamura L31S100S05FS 5V or such sensor after capacitor output line.

There is still a problem where (if) to mount output diode. But if we consent to keep charger permanently connected we can use an NTC resistor across DC contactor to keep positive DC link precharged. I always use diode like VS-RA160FA120. It could be mounted directly to the car chassis to dissipate heat.

Additionaly i will use dedicated relay to turn on DCDC 14V power unit for those cars that have main battery voltage 150Vdc to 310Vdc. Also i will use main cooling system for inverter and charger with single pump driving the system

What we know about signals in Prius gen 3 inverter. Thanks Damien.
- To drive charger i will use EMW setup with some of my twists .
- HVIL = High Voltage Interlock. For DIY built we can ignore this one IF we use our own safety in maintenance - our brain.
- Fault signals are 5v when not faulted and pull low when a fault is detected. I will definitely use this with some NC interface to inhibit PWM
- I will use original NTC and later compare characteristic to semicron in code. If not i see i can use whatever sensor on the coolant line i can imagine .
- Current sensors are 5v centered and react at 10mV/A
- MG and BOOST drivers have their own pullups in the circuit and we can ignore them if we dont use them.
- Since i will be using HCPL-7520 to sense output voltage i will ignore DCBUS contacts for now. Maybe later in adaptation we could just use that I/Os and adapt the code.

[celeron55]

I'm still waiting for the post to deliver my gen3 inverter for attempting the same. Or, well, it's starting to look like I need to order a second one and consider the first one lost forever.

I'll be following this meanwhile. I'm interested in any remotely useful PCB designs!

[arber333]

I ordered one Prius interface board from JB and when i get it i will connect my EMW charger boards and try to charge my Mazda from 3phase.

[sonachan1]

Hey arber,

Looks like an interesting project! I've made a DC motor drive board from Damiens Prius design (not yet published as there are a couple of mistakes I want to correct beforehand). There were a few things I discovered which may be useful to you.

When my inverter arrived, the board was not the same size at the top edge (as shown in your screen grab) as the one Damien had. So I added a few holes to match the board I had and made the layout suitable for the 2 types of board we know about. Let me know and I can send you over the board file to get the hole positions from if you like. There's some early chat about this in Damien's original Prius Inverter thread.

Also, on the top edge, make sure you don't have any tall components as they can clash with plastic framework (like my power supply inductor I tucked away up there!).

Finally, there are a couple of metal inserts in the frame where the original multiway connector fits, so make sure you don't have any through hole parts or tracking in that area.

I'm interested to see how you get on with this as an alternative (and cheap) charging option!

[arber333]

No for now i will only use interface to try the concept. After i confirm it works i will try a cardboard mockup how it will fit. I will use thos contacts as GND connection with case.
I have Auris inverter and i expect to try one Yaris soon also. Can you share which P/N of inverter you have?

tnx

[sonachan1]

Hi arber,

The part number on my inverter is G9200-47190. The control PCB was F1759-47070/05. The 05 is in a white square and looks like the revision number. It came from an Auris according to the breaker.

[arber333]

Hi arber,

The part number on my inverter is G9200-47190. The control PCB was F1759-47070/05. The 05 is in a white square and looks like the revision number. It came from an Auris according to the breaker.

Yes i have the same one.

[arber333]

OK Guys!

I have connected my Ampera battery to Prius inverter and observed DC_BUS1 and DC_BUS2 signals. There are three power supplies needed for operation of power section. I used Lab PSU for 13.7Vdc and 2x simple 5V 1A ACDC power supply. Since we need to supply -5V and +5V i wired outputs in series so i got +5V, GND and -5V. This can be done because both 5V supplies have isolated inputs. I see i will have to make 2 simple split supplies for my board design similarly to what JB did.

Heres what I found out:
1. DC_BUS1 measures voltage on the Battery side, when battery is disconnected it is effectively 0V.
2. DC_BUS2 measures voltage on the motor DC side. When i move outside source its value is 0.554V.
3. When battery is at 213Vdc i get DC_BUS1 value and DC_BUS2 value 1.075V.

I could go up to 380Vdc with Ampera battery and i arrived to BUS voltage 4,758V.
I can see from data that battery voltage is limited to something around 400Vdc. There DC_BUS1 becomes 5V and Arduino AD wont be able to reference above that. I have yet to try over 400Vdc. I have 32V power supply which i can wire in series with battery. This will show how high this inverter can go.
Also i intend to try to connect 3phase 400Vac and inverter should make 600Vdc rectified.

I input rest of values into XLS file to interpolate between them, see the file.
EDIT: I had my first try at interpolation

[Jack Bauer]

Did the EMW charger ever work for you?

[arber333]

Did the EMW charger ever work for you?

I use it every day now for like 4 years. Mainly i use it for quick charging from L2 EVSE at 11kW. Original inductor is a bit chewed now but i have a spare Prius gen 2 inductor in store to replace it. I will have to redo alu box to get it in.... or i would just make a box from plywood, works for my friend.

Yes one mod from original i did to save myself trouble was to make the charger straight buck. No pesky PFC and strange feedback. I just use transistor in push pull config before inductor. Perfect config for Prius inverter.
Well see how it goes on weekend...

[arber333]

I got Toyota Yaris inverter at home. Thank you Damian.Io.

I immediately went to disassembly. I used a flat screwdriver to pry open the seal and took the lid off. I was surprised how small main capacitor is.
Next i took off the current sensors which look the same as the ones on Auris inverter btw.
Then i came to main board which is EXACTLY the same as my Auris main board. I even switched them on their mounts and each connector fits.
So now that i know those things i can safely make a universal charger board. Some pics...

[Berdi]

We are looking forward to success in the project.!!! )

[arber333]

As Damien told in one of his posts, Gen 3 uses +5V fault signal when everything is good. When something bad happens this signal is cut and PWM shuts down. I will use this to tap into EMW setup at UVLO signal. This was meant to sense 12V voltage and in case charger would drain car battery too much it would shut down PWM before brounout would happen. I just have to fix resistor divider.

[celeron55]

Got my inverter today as I ordered from a different seller. Finally after 5 months I can start working on this...

I used the original board to make me all the voltages for testing. Very nice of Toyota for giving me all of them without any modifications or communication with the board.

I tapped the boost converter low side control a few times by hand and voltage on the main rail increased from 11V to 24V. Good start. Now I just need an Arduino to tap it for me.

Where's the challenge, Toyota? This is too easy...

[arber333]

Main boards for hybrid EMW charger were sent to manufacture. I used additional relays independent control of DCDC converter (if inverter will be OK with voltage), coolant pump and finally external AC and DC contactors.
I will have to add 100A 5V Tamura current sensor for native current control. Prius current sensors are too complicated.
After boards return i will populate them and play some with code to calibrate charger brain to Toyota voltage report.
Then comes 600Vdc rectified into charger!

[celeron55]

I tested converting about 500W from 25V to 10V or so and it worked perfectly, so now i tried to move to testing the converter on mains voltage.

The boost converter works fine when I feed some small test DC (12V, 24V, 35V) on bus 2 through the mg2 transistor diodes and send some phase correct PWM signals to the boost high and low side inputs (pulling them low to activate). There is a voltage proportional to PWM on bus 1, with lots of capacity in amps.

But now when I put 230VAC mains on bus 2 (resulting in about 300V rectified), the boost converter will make a weird chack-chack-chack sound and does almost no converting at all when measured on bus 1, where I am hoping to see a voltage proportional to PWM, as I do see when using the lower DC sources on bus 2.

Has anyone tried anything similar yet, with success or failure?

[Jack Bauer]

Ah, you have annoyed the Toyota gods. The clacking is them showing their displeasure:) Do you have the big capacitor fitted? It's located in the lid but can be removed. If not the 100hz ripple will annoy the Toyota gods. They also get angry if you don't leave enough deadtime between the high and low side. As you are bucking down from bus two you can just drive the high side and leave the low side off.

[celeron55]

I do have the big capacitor on.

I'll see if controlling only the high side helps. I'm a bit doubtful as I did try different deadtimes already.

The frequency this is supposed to run at is 7.45kHz, right?

I think I'll need to check my 12 volt levels also. I wonder what the minimum is.

[arber333]

I do have the big capacitor on.

I'll see if controlling only the high side helps. I'm a bit doubtful as I did try different deadtimes already.

The frequency this is supposed to run at is 7.45kHz, right?

I think I'll need to check my 12 volt levels also. I wonder what the minimum is.

I still wait for my boards to come from foundry. I intend to drive only high side for buck. Low side must stay closed so diode would eat the surges and send them forward through inductor again. However i left the option to connect the low side too if i want to experiment with boost stage later on.
If you only have one transistor operating then you dont need deadtime do you .
In my charger i use 80us period for 12kHz operation, but i also used 120us for 8kHz.

I dont think there is an option of making 600V from 1phase here, so i think we will have to drive the boost from 1phase.

[Jack Bauer]

The booster normally runs at 5khz but I've had it over 10khz. also worth noting if the igbt temp rises above 80c they will shutdown.

[celeron55]

I got it working by not using the low side switch at all. Did a test converting about 2kW into a 2kW 230VAC heater just now which it did at 39% PWM from 230VAC mains rectified to 300VDC. 39% of 300VDC equals 230VDC? Ehm... well it converts power, it's fine enough to me.

Now I need to find the encouragement to connect this to 3 phase. Already have the cable and connector, and I do have a 32A 3-phase RCBO (from china, of course) and a set of input side contactors with precharge resistors sitting on the inverter, with automatic control. So it should be a fairly "hands-off, look elsewhere" type of situation.

The sound is awful.

[johu]

Bucking is always weird. I tried calculating the output voltage of inverter buck mode as input*dutycycle and failed miserable. It works perfectly for current in boost mode, output=(1-dc)*input.

good luck with 3-phase, can relate to what you're saying

[Jack Bauer]

Cool. Glad it worked. I think this could form the basis of an excellent charger.

[celeron55]

While my control software isn't the most stable, it's likely I bucked about 7kW from a 16A 3 phase supply into two parallel 2kW heaters, outputting about 25A @ 270VDC.

That's quite useful. I had some brass rod, welding cable and stuff ordered. I'll make a DIY CHAdeMO connector so that I can use this as an external charger on the EVPrevia, as I'm too lazy to install it on the car right now and making a grossly out-of-spec portable CHAdeMO charger seems like fun.

[Jack Bauer]

Excellent. Getting to Tesla charger power levels:) did you have any coolant flow? Are you using the phase current sensors for the ac side?