Prius Gen3 Inverter Hacking

[johu]

Can you make sure that I can reliably detect the hardware variant in software e.g. by pulling a free pin to a known voltage? That way we can keep the unified software

[Jack Bauer]

Will do. I have a spare pin we can use.

[MattsAwesomeStuff]

Damian - Thanks for the updates!

People on DIYEV forums have been asking again and again what they can do for an inverter, and I've been saying "Damian's working on something to hijack Prius Gen 3 converters"... and then have been saying "... but he's sold out right now I think".

I really do think this is the best cheap DIY option for AC motors, maybe the only cheap option. Lots of people are waiting on this to become a thing.

Just wanted to let you know that, when you do these things, people notice and it does matter.

[johnspark]

Those of us waiting for boards, are we getting V1 or V2 now? I'm easy either way.

Had it been available in kit form, I would have gone for that. With regards to using the clone boards, there's no sense in soldering up the circuits when they can be had for so cheap ready made.

I am the same as Bigpie, i am happy either way. Kind regards.

[Jack Bauer]

As long as it suits everyone they will be v2 boards. I have no bare v1 left as the pcb maker ruined them.

[johnspark]

I would like to use the gen 3 inverter with as higher battery voltage as possible. It seems from this forum nominally 308 volts is the maximum that can be safely used, or with lithium batteries with max allowable voltage of 3,6V per cell, 3.6*85 = 306 volts. Probably stick to 83 cells for more safety.

Power Perspective
Next I would still like to use the boost converter, which is rated for 27kW. I only need it to go to 500 Volts, so it only has to cover 300 to 500V not 201.8 to 500 Volts. If I want maximum 50kW power at 500V, means I need 100 Amps at 500V. 300V battery provides 30kW, DC-DC inductor only needs to provide remaining 20kW -which is within the 27kW rating.

From current perspective, according to the "Evaluation of the 2010 Toyota Prius Hybrid Synergy Drive System" by ORNL, the capacitors of the DC DC link were tested to 200Amps at 2khz page 35. This suggests the inductor can take at least 150amps. Locked rotor tests were performed up to 250amps (very short time) page 61. All these investigations tell me the inductor IGBT capacitor system can take 150 amps with a little extra for short time.

i intend using plenty water cooling..

Am I on the right track?

[arber333]

I am sure that if you use 500Vdc battery you can control IGBTs directly then phase amp limit is IGBT capability. You would have to leave out DCDC section though, but you can use one Volt DCDC unit in that role.
You can then use the other IGBT bridge as a fast (buck) charger to charge from 3phase 580V rectified to 500Vdc battery. Inverter will think it is in regen .

[johnspark]

Thanks arber333 for that . I still want to use the other IGBT bridge because my set up is for the Gen 3 inverter to control a Gen 2 Prius transaxle. I have had a shaft made to join the generator and motor together. So plan to use the motor most of the time, then for higher speeds use the generator as a motor to help provide a bit more power at the higher speeds. I will be using this shaft to couple to a 5 speed gear box. So i could do test runs to see what is the highest voltage i want and only build a battery pack for that voltage.
The workshop i am getting to do most of the mechanical build are transitioning to an EV maintenance workshop and giving me EV club rates . Also going to let me use their new dynamotor for free...
Then as you suggested use DC DC converter for charger. Happy with slow charge for first EV. I will have an electric oil pump circuit which i can have a cooler built into as well as a big radiator for the water cooling circuit. So with all this cooling, hope to make this hybrid (part-time) motor work as a full time EV motor.

[arber333]

Ok, fair enough. I expect you will get more RPM from DCDC rig. But with 500Vdc directly you would get much more torque. BUT you must not use more than 540Vdc since 3phase voltage sag at 30A is like 550Vdc rectified (together with conduits and cap ESR).
I am not sure why use another transmission. It means additional drag and weight.
Say you put 4 gears in the PSD to couple MG1 and MG2 directly then PSD device will become a rigid clutch. No need to weld it. But then you would only have rpm up to some 100km/h or maybe 120km/h. Still for light EV would be enough.
However if use gen 3 transmission P410 you have an option to turn MG2 against MG1 and use the ratio to get more speed out of the system. I think it is the better option.

[johnspark]

Higher torque at higher speeds will be much more important, so that sounds better to just have batteries directly onto the HV bus (<= 500Vdc). I am only using the Gen 2 transaxle because it was the first one i got and would like to learn what I can from it first. I was also keen to reuse the Prius transaxles because they were readily available and reasonably priced, and i thought it was useful to try and reuse something that would otherwise end up at the tip and waste those permanent magnets. I am not using any gears to couple MG1 and MG2 together. I have had a shaft made that mates the MG1 and MG2 splines together, with a smooth 25mm dia. round surface which I will use to connect to the manual gear box. I am only interested in 120km/h because we can only go up to 110km/h on the roads where i come from.
Interesting what you state about the gen 3 transmission P410, but will try to get some return from what i have first.

[arber333]

I see then. It will be like two motor drive in line. You will have to be careful to idle MG1 or drive it with torque in relation to MG2. That way it will not drag the shaft in regen.

So if you make a coupling for both motors than you can take the chain and rest of the stuff out. It is only weight and resistance...

A

[johnspark]

That's exactly right, so that saves losses from the chain and gears. But as you state, will need to idle MG1 or drive with small amount of torque. So from my experiences, would set MG2 as master and MG1 as follower. So my idea is to use MG2 most of the time and MG1 only at the higher speeds where the torque is low.

[johnspark]

At 6006 rpm,
gear losses are: 1645.3W (25C)
motor rotor losses are: 472.8W
planetary gears,
gen rotor,
sun gear losses are: 264.9W (don't have gen on its own)

According to: pg 24 Evaluation of 2004 Toyota Prius Hybrid Drive System ORNL

[Jack Bauer]

Schematic for the V2 board now done.

[arber333]

That's exactly right, so that saves losses from the chain and gears. But as you state, will need to idle MG1 or drive with small amount of torque. So from my experiences, would set MG2 as master and MG1 as follower. So my idea is to use MG2 most of the time and MG1 only at the higher speeds where the torque is low.

Hm... What if you would use a single throttle with split filtering and setup the response so you would always have zero throttle on MG1 before MG2. You would make R bridge so to have MG1 throttle always lower than MG2.
When throttle would be pressed MG1 would idle up to some level when MG2 would develop real torque, then it would help on demand.

[johnspark]

its definitely worth trying because it is a relatively simple control scheme, that can be adjusted during testing...

[johnspark]

its definitely worth trying because it is a relatively simple control scheme, that can be adjusted during testing...

[arber333]

Schematic for the V2 board now done.

Hi JB

I am working on my Prius inverter to charger conversion.
BTW i used your board outline as a measure and will credit you of course .
Now i am observing 50p control cable and others. Can you answer a question or two

- What is the purpose of conn9 pin HVIL? Is this a live HV+ signal or is it opto isolated analog signal. Or even further, digital data for HV value?
- What is the Fault signal like? Is it Active Low or Active High? It is the same for all PS, no?
- What is the characteristic of coolant sensor? Is it NTC or digital? Need to make a values table and i dont want to mess up my workshop . Do you have
the table?

EDIT:
- Also, what is the polarity of IGBT signals. Are they direct or invered?

tnx

Arber

[arber333]

Schematic for the V2 board now done.

Hi JB

I am working on my Prius inverter to charger conversion.
BTW i used your board outline as a measure and will credit you of course .
Now i am observing 50p control cable and others. Can you answer a question or two

- What is the purpose of conn9 pin HVIL? Is this a live HV+ signal or is it opto isolated analog signal. Or even further, digital data for HV value?
- What is the Fault signal like? Is it Active Low or Active High? It is the same for all PS, no?
- What is the characteristic of coolant sensor? Is it NTC or digital? Need to make a values table and i dont want to mess up my workshop . Do you have
the table?

EDIT:
- Also, what is the polarity of IGBT signals. Are they direct or invered?

tnx

Arber

OK I found some more info on one of the posts on DIYELECTRICCAR
- Current sensors are 5v centered and react at 10mV/A - good to know!
- IGBT drive lines are active high 5v - good!
- 7.45khz for the boost converter - This will be noisy!

So, can you fill in the missing info please?
1. What is on the conn9? Is this direct HV or 0V - 5V signal, or anything else?
2. What is the Fault signal? Is it Active high or Low?
3. Coolant sensor curve?

tnx

[arber333]

Damn i just found out HVIL connector is in fact part of an interlock circuit that shuts inverter off if we remove the black cover from the upper part.
So now i can actually use whatever cionnector i want to supply HV to my charger brain.

[Jack Bauer]

Schematic for the V2 board now done.

Hi JB

I am working on my Prius inverter to charger conversion.
BTW i used your board outline as a measure and will credit you of course .
Now i am observing 50p control cable and others. Can you answer a question or two

- What is the purpose of conn9 pin HVIL? Is this a live HV+ signal or is it opto isolated analog signal. Or even further, digital data for HV value?
HVIL = High Voltage Interlock. It is a loop used to detect if hv connectors are properly seated, cover is in place etc. In this case I loop it into the NAND gate and have a jumper to disable for testing.

- What is the Fault signal like? Is it Active Low or Active High? It is the same for all PS, no?
Fault signals are 5v when not faulted and pull low when a fault is detected. It is worth noting that even if you ignore fault signals the igbt drivers shut down anyway:)

- What is the characteristic of coolant sensor? Is it NTC or digital? Need to make a values table and i dont want to mess up my workshop . Do you have
the table?
I have not got that far as yet but they do look like ntc. The igbt temp sensors seem to exactly match the semikron option in tmphs.

EDIT:
- Also, what is the polarity of IGBT signals. Are they direct or invered?
The igbt drive lines must be pulled low via a 100R resistor in order to activate the associated gate.

tnx

Arber

OK I found some more info on one of the posts on DIYELECTRICCAR
- Current sensors are 5v centered and react at 10mV/A - good to know!
- IGBT drive lines are active high 5v - good!
- 7.45khz for the boost converter - This will be noisy!

So, can you fill in the missing info please?
1. What is on the conn9? Is this direct HV or 0V - 5V signal, or anything else?
2. What is the Fault signal? Is it Active high or Low?
3. Coolant sensor curve?

tnx

[arber333]

Ok JB, can you confirm following?

- i will only use PWM on BOOST top driver and pull BOOST low driver up to 5V. This will give me Buck converter from 600Vdc to 380Vdc etc...
- If i dont want to use MG1 phase contacts i need to pull MG1 driver tracks up to 5V?
- If i want to use MG2 phase contacts as input for 3phase i need to pull its driver tracks to 5V. This way i will only use reverse diodes to rectify 3phase.
- Do you think i can leave MG1 and MG2 fault pins floating. Only BOOST fault pin will matter for charging anyway.

What kind of signal is DCBUS 1 and 2? Is it analog 0V - 5V? I cant test because i already took my Prius board apart.

tnx

A

[Jack Bauer]

Ok JB, can you confirm following?

- i will only use PWM on BOOST top driver and pull BOOST low driver up to 5V. This will give me Buck converter from 600Vdc to 380Vdc etc...
they have their own pullups
- If i dont want to use MG1 phase contacts i need to pull MG1 driver tracks up to 5V?
they have their own pullups
- If i want to use MG2 phase contacts as input for 3phase i need to pull its driver tracks to 5V. This way i will only use reverse diodes to rectify 3phase.
they have their own pullups
- Do you think i can leave MG1 and MG2 fault pins floating. Only BOOST fault pin will matter for charging anyway.
yes that's fine.
What kind of signal is DCBUS 1 and 2? Is it analog 0V - 5V? I cant test because i already took my Prius board apart.
Its in my notes. Will lookup.

tnx

A

[arber333]

Tnx Damien.
Theese days i was working couple of times late at night and comparing notes suffers somewhat...
So i will summarize . Feel free to correct me.

- 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.

I decided i will not use original current sensors. Thay 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 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 a NTC resistor across DC contactor to keep DC link precharged. I always use diode like VS-RA160FA120. It could be mounted directly to the car chassis to dissipate heat.

I will now make a new thread.

[arber333]

My new thread... viewtopic.php?f=14&t=257

Say JB, can you tell me if you actually tried to run Buck/Boost converter upwards of 300Vdc on the battery side? Today when i tried to run DCDC again i got a sick feeling that Toyota put a brake to my developement on the 200V battery side. You were talking about some ASIC protection. Do you think it would work to buck from 600Vdc on the motor side down to 360Vdc on the battery side? I think system wouldnt care since transistors have to be 1200V rated anyways. What do you think?