Doing useful things with the Prius Gen2 buck/boost converter

[johu]

Many hybrid inverters feature a buck/boost converter. It can step down the voltage from the motor to the battery side to charge the battery with a fast spinning motor. And it can boost the voltage from the battery to the motor side to make the motor spin fast from the battery. Battery voltage in hybrids is typically about 200V.

In addition they feature a DC/DC converter from HV to 12V to run the "car" systems. It sits on the battery side of the converter.

The Prius Gen2 inverter features this and due to its not too integrated design we can faff around with it.

First of all, when using the Prius inverter to power an actual EV we will find that the converter is "in the way" as it sits between the battery and motor side and is a bit of a weak link, limiting battery current to about 100A.

Secondly we will find that when using a regular 400V style battery salvaged from a PHEV or EV the voltage exceeds the rated input voltage of the DC/DC converter.

With a simple mod we can fix both these problems

We disconnect the positive battery cable from the boost converter input and reconnect it to the boost converter output i.e. the motor side of the converter. This has 2 consequences
1. The inverter bus is now straight connected to our HV battery and can draw as much power as it can handle
2. The DC/DC converter still sits on the buck converter output and by running the latter at, say, 50% duty cycle will feed it with the 200V it expects

With the Gen2 inverter this has a 3rd consequence: in parallel with the DC/DC converter there is an auxiliary inverter. If we replace its output IGBTs with a bridge rectifier and feed 120 or 230V AC to it we will have a rectified DC voltage of 150 to 300V on the boost converter input. So now when we run the boost converter it will feed current into out HV battery - we have created a simple mains charger.

[Jacobsmess]

This is exciting

[MattsAwesomeStuff]

With a simple mod we can fix both these problems

I have a second inverter with its top off I will try to take pictures of if I can figure out exactly what to do based on your diagram.

Also, does any of what you described require using new pins or telling the buck/boost what to do? Any code or software? Or does it just happen by being connected?

[Keo]

I assume the same can be done with the gs450h inverter? Does it also have the auxiliary inverter to run the same setup as with this?

[johu]

Also, does any of what you described require using new pins or telling the buck/boost what to do? Any code or software? Or does it just happen by being connected?

Sorry, just found this. With the SMT version of the Prius controller this is integrated. In drive mode the "user PWM" pin runs the boost converter at a fixed rate. In charge mode the regular PWM runs it in current control mode.
Here we do indeed need additional hardware: an extra current sensor that measures current delivered by the AC/DC.

I assume the same can be done with the gs450h inverter? Does it also have the auxiliary inverter to run the same setup as with this?

There is no replacement logic board for Lexus inverters. I don't know if the buck/boost converter can be controlled with the Serial protocol? And I'm pretty sure it has no aux inverter.

[Keo]

Ok thanks. I'll moving forward with the gen2 Prius since it's relatively cheap where I live. I'm thinking of running a 100v 100ah to the boost converter to get anything around 25kw to run the motor. I don't need much power since there's no traffic where I stay and 60kmph top speed is enough. Between running via the boost converter vs direct to inverter I don't know which is best since I can't get a high voltage battery pack here in Africa without importing and that will more than double my costs

[johu]

The booster is limited to 100A, so this will get you 10 kW

[MattsAwesomeStuff]

With the SMT version of the Prius controller this is integrated.

Good news, I also bought one of those 2nd hand from someone who went a different way. So I have both hardwares.

In drive mode the "user PWM" pin runs the boost converter at a fixed rate. In charge mode the regular PWM runs it in current control mode.
Here we do indeed need additional hardware: an extra current sensor that measures current delivered by the AC/DC.

Stupid question, how does it know whether it's in charge mode, and is charge mode also what happens during regen?

Do I need to install an extra current sensor, or does the SMT board come with a way to interface with one built into the inverter?

[johu]

Good news, I also bought one of those 2nd hand from someone who went a different way. So I have both hardwares.

Is it version 1.3? Early versions don't have the user PWM connection.

Stupid question, how does it know whether it's in charge mode, and is charge mode also what happens during regen?

Because you explicitly command charge mode https://openinverter.org/wiki/Battery_Charging
Regen=Drive Mode

Do I need to install an extra current sensor, or does the SMT board come with a way to interface with one built into the inverter?

Yes. The board has 4 inputs. You use two inputs for ONE inverter power stage and one for the extra sensor. You need to juggle a bit to have the same current to voltage ratio as the built in ones. You could regard the latter as cosmetic. Say the extra sensor outputs 50 mV/A and the built-in ones 25 mv/A it just means you need to command double the charge current.

[MattsAwesomeStuff]

Is it version 1.3? Early versions don't have the user PWM connection.

*checks*

V1.1

Whelp, nevermind I guess.

[johu]

So charging feature would work but not powering the dcdc converter while driving.

[MattsAwesomeStuff]

So charging feature would work but not powering the dcdc converter while driving.

Oh, I misunderstood.

In the video you shared, you say it uses the A/C inverter to do this? I was somewhat planning on using my A/C inverter to run the A/C. Though that said, I also have the A/C off of a Gen3 which only needs PWM fed to it as the inverter is onboard the compressor.

Does the charging feature both buck and boost too? Or is that feature lost on earlier versions?

[johu]

Via the A/C inverter it will just boost
Via MG1 it will just buck. You need an additional relay to connect the battery to the buck side while charging. At least then you don't need the extra current sensor because it will use the MG1 ones

[MattsAwesomeStuff]

Via the A/C inverter it will just boost. Via MG1 it will just buck. You need an additional relay to connect the battery to the buck side while charging. At least then you don't need the extra current sensor because it will use the MG1 ones

I think I mostly understand this, but not sure I'd apply it correctly.

Could you diagram the change like you did in the first post, so this is idiot(me)-proof?

I'm stepping closer and closer to hooking up a battery and driving here eventually.

[jrbe]

So these are the 2004-2009 Prius inverters right?

Has anyone figured out what year / serial the split to v1.3 is?

Could this potentially boost a 48v solar panel (500-2500 watts) up to 400v to directly charge a traction battery?

Edit,
Is it possible with the gen 2 v1.3 to pull 48v out of the 400v battery if needed?

[johu]

Yes and yes. Of course that will loose the DCDC converter because that doesn't run at 48V.
MPP tracking won't be possible as you can measure input or output current.

[imwoody36]

I think I mostly understand this, but not sure I'd apply it correctly.

Could you diagram the change like you did in the first post, so this is idiot(me)-proof?

I'm stepping closer and closer to hooking up a battery and driving here eventually.

Im also looking for a diagram for grids charging with the gen2 inverter, and im unclear if it can be stand alone. because my application is a harsh marine environment where there is zero support and in remote locations. target voltage 160 to 220 vdc. but gen 2 inverters are agreeably cheap and rugged. i had difficulty following johu's german language youtube on that even with text enabled.

[JeremyMakesThings]

I'm inching closer to doing the charging side of this on my build (once I get my BMS actually working...) and I naturally have a ton of questions. I'm planning on documenting it all in video form and can take a stab at writing it up for wiki if i manage to figure it all out.

I'll stick to one question for now- looking and poking at my "spare" (water damaged) inverter, I'm wondering if the AC inverter board is doing anything important for the rectification side of things, or if I can just stick a bridge rectifier next to it and just hook up the +/- from the rectifier to the terminals on the board to connect to the wires that go to the top side of the inverter?

[johu]

I'm wondering if the AC inverter board is doing anything important for the rectification side of things, or if I can just stick a bridge rectifier next to it and just hook up the +/- from the rectifier to the terminals on the board to connect to the wires that go to the top side of the inverter?

Yes, rather do that.

[moebelschmitt]

Hi,
since I'm also in the process of my conversion (Yaris Gen3), I'm very interested in your attempt and hope to follow your progress.
I've already subscribed to your YouTube channel
Peter

[JeremyMakesThings]

I finally got my BMS back from Ewert, got everything hooked up and drove around the woods a little bit yesterday, so I’m back to thinking about adding a charging system. I had a couple more questions, but in sitting here writing them, I realized I was misunderstanding where the current sensor goes. I was thinking it was on the line going to the battery, but now I realize it goes between the rectifier and the buck-boost converter.
So now my questions are:
1) If I run a separate cable from the buck-boost to my rectifier (rather than going though the ac inverter), would I still need a relay to switch the current sensor, or can I just let it hang out when it’s not charging? I guess I should do some testing with the current sensor I have to figure it out.

2) Before I understood where the current sensor lived, I was wondering if it would be possible to feed the controller a value for idc via CAN to control the charge current. So now I’m wondering if that would theoretically work, and how much coding it would take?

[johu]

1) in that case it shouldn't disturb motor operation
2) no. Current is sampled at 8.8 kHz to get a somewhat tolerable power factor. So it decreases PWM at rectified peak and increases around 0 crossing

[JeremyMakesThings]

Ok, I think I’ve got a plan, opened up the inverter and then realized I’m not 100% sure how the current sensor gets hooked up.

My understanding is GIWA and MIWA share an input, as so GIVA and MIVA. Does it matter which one I connect to?
And also a question on the burden resistor- I assume this is based on the particular current sensor I’m using, and I should be targeting a particular voltage range? And what is the range? Or I guess what I really need to know is what is the control board looking for from the current sensor?

[johu]

Yeah well the current sensor part of the board is up in the air really. The original plan was to use MG1 as a 3-phase rectifier and then buck into a 400V battery. But it then turned out that funneling the motor current through the inductor would burn up the latter. So now the battery is always directly connected to the DC-bus.
With the original plan MG1/MG2 or GIWA/MIWA current sensor would have been sensed by the same analog input because one is only active when charging and one when driving.
Now ideally you want a current sensor that acts like the MG2 sensor. +-/15V and the same current to voltage ratio. I happened to have such a sensor lying around that I tuned to be Prius-like with the burden resistor. But this only works with a current mode sensor (I call it that, the actual name might differ) i.e. say, for 1A primary current it creates 1mA secondary current.

I wonder if you could extract an MG1 sensor and route your charging current through it?

[JeremyMakesThings]

I’ve got a junk (water damaged) inverter that I had previously tried to pull oen of the current sensors from, and it was a bit of a dead end for me. It would be tough to get it out without destroying the inverter, it has 5 pins on it and the only one I can identify is the ground. The only markings on it are “KOHSHIN” and what looks like a Toyota part number (g9247-47050 and g9247-4740…different numbers on mg1 and mg2, interesting). Searching that go me nowhere.

But….I’ve got the current sensor, it’s already wired up, so my current plan is to try to use it in situ. The way the Busbar connects to the igbts doesn’t really give enough clearance to disconnect it from the igbts and connect a cable, but I think I can pull the Busbar out entirely and run multiple stands of 16awg (~1.3mm) wire though the current sensor. (I just need one sensor, correct?)

I’ll report back.