[evforme?]

Finally a current thread on the system I want to convert an old bobbed kawi 454 ltd I have in the back yard to electric, I can pick the 06-11 gen IMA civic motor on ebay for 180$ (CAN) and I got a bunch of free gel cell 12v batteries from soem UPS's. I'm thinking 15kw is pleanty, it's not going to be a racer.

So I'll be looking forward to see some controller and mounting solutions, I live on a hill so I'd like to see some regen one day.

 

[EV-4-Me]

I realized I never posted pictures back here after getting an email notification of a post in the thread. Here they are:

 

[EV-4-Me]

You can see that the top board kind of floats directly on top of the IGBT's and their drivers. The white rectangles in picture two are all the connections to the drivers from the top board. There are 11 in all. I haven't identified all of them, but four are +-17V and one a constant 5V.

 

[EV-4-Me]

These are old and life has gotten in the way, but I still have plans for this. 1998 CR250, originally planned on running NiMH hybrid cells since they were the cheapest option at the time. Leaf and Volt cells have now made those a pretty poor bargain.

 

[Tomdb]

That might work, I bet and plate to mount where the motor was would be quite pricey to get made. You could get away with a single sided bearing setup with two bearings.

Could you explain what it took to get the top board off?
Thinking about getting some more documentation to together, so powering it up and measuring all the voltages for the interfaces, to the drives and sensors.

 

[EV-4-Me]

You're not going to like my answer to that one...

I just pulled it and wobbled it around until it came apart. I don't see any easy way to desolder and remove the board nicely. You have to break those wires apart as far as I can tell.

I'll be designing and fabricating an entire motor enclosure. I own a machine shop and have a degree in ME, so no biggie. I am going to try to do it in a way that will be fairly inexpensive for others to copy and that will allow motors to be stacked. The stock housing is nice, but is very bulky and not great for liquid cooling.

 

[evforme?]

Would that cover plate need to be non ferrous? Have you figured out your gearing yet?

 

[EV-4-Me]

As far as it being non-ferrous, I wouldn't think it matters. Mine will be aluminum just for weight savings.

No idea on gearing, yet. It will likely have to run a jackshaft (maybe through swingarm pivot with some mods?) I'm still in the, "get the motor finished and turning" phase of this project. I'll also have to develop an encoder solution for it. As far as I know, neither the Lebowski brain or Eldis' solution is capable of taking resolver input yet.

 

[Tomdb]

You're not going to like my answer to that one...

I just pulled it and wobbled it around until it came apart. I don't see any easy way to desolder and remove the board nicely. You have to break those wires apart as far as I can tell.

I'll be designing and fabricating an entire motor enclosure. I own a machine shop and have a degree in ME, so no biggie. I am going to try to do it in a way that will be fairly inexpensive for others to copy and that will allow motors to be stacked. The stock housing is nice, but is very bulky and not great for liquid cooling.

So the main board to the, gate drivers you mean?
I see that the current sensors are soldered, this will be the first thing im going to do after taking initial measurements.

 

[EV-4-Me]

Yessir, that is correct. You have to desolder the current sensors first.

 

[Tomdb]

how does the back of the big board look?

So how is the connection made between the main board and IGBT, also soldered, the plastic makes me think some kind of spring connection?

 

[EV-4-Me]

The black plastic pieces are floating on top of each other with springy, but solid connections. I think these may be put on the IGBT boards first, then the top board placed on and the connections soldered. There's no way to get the black pieces apart without breaking them. You may be able to get the top board off by desoldering all the pins from it.

 

[Tomdb]

Sweet picture.

Made a pdf with my findings. I shall measure the gate drive signals with my scope now.

pins 6-9 in my labeling.

Could you post a picture of the back of the gate driver board?

 

[EV-4-Me]

Sure, give me a bit to get one off again. I don't think I took any pictures when I had it apart.

 

[Tomdb]

Hmm, this puppy is telling all its secrets already. Now lets ope it does not get scared when it decide to pwm one of the gate driver legs.

Edit: looked at the gate driver schematic. Almost clear how it works. Going to look if I can get it switching and measure its response to determine if my assumptions are correct. In the meantime I update my document, please give me some feedback if you have a different idea.

 

[Tomdb]

Between Inverter power on and 5 volts being on takes roughly 100ms.

Time to hook up the arduino and take a look if i can just drive one of the two sides.

 

[Tomdb]

ok i got the igbt working.

Arduino pushing a 5v pwm and we are in like flynn, as my favorite electronics youtuber might say.

Next step might be actually hooking up a brainboard from Johannes, and see how we do.

I am not 100% sure yet if the controls are directly linked to driving the and low section or there is one pwm duty signal and one high low signal.

First pic shows me driving a nearly 50/50 duty cycle, rough arduino code bashing. With 10us of forced delay, visable on turn on of the "high" side, however when driving "low high" the delay is gone. I wonder how this is happening.

Second pic shows 25% duty cycle for the High and low signal. Again one edge has the duty cycle missing. Got to test it with load next, because maybe these is some capacitance in there.

Last pic is my measuring setup. I will need to attack the signal application a little more structured so i can derive what pin does what.

Right now I am just using; Pin 7 as a ground pin 6 and 8 as driving signals.

So keep tuned for further testing to come, including mapping the current sensors, lets hope these puppies are linear.

 

[EV-4-Me]

That's great progress!!

Where did you get a gate drive schematic?? I took some closer pictures of the top of the drivers to hopefully help out. I looked up the number of the driver chip itself and called around everywhere I could...Probably put 10 hours into finding a datasheet and it was apparently proprietary to Honda so nothing is available.

I also realized why I never took a picture of the back side...

Edit: Realized I misunderstood you about the gate drive. Looks great so far!

 

[Tomdb]

I did some analysis of the pcb layout to come up with the workings of the opto couplers

I do not have any time today to fool around with different driving signals, will do that tommorow.

 

[Tomdb]

Got it hooked up and the appropriate signals broken out.
Now i just have to figure out if I am correct that the signals are high and low side. Maybe I am wrong and it runs on a with a different signal set.

So more experimenting will be needed.