Sanwiching 2x IMA motors together for EV

[Chris(MA)]

Hi Folks, I've not been on the forum for a long time. I sold my insight back in 2007 or so, although I regret selling it the two seats were a bit limiting for us.

I've been hankering for an electric vehicle for a long time, but I've finally decided to go ahead and try and build one. An insight body is the ideal vehicle due to its light weight and aerodynamics, however I don't have the funds to buy one to gut for an EV, the 2 seats is a bit of a limitation plus I can use my EV mainly on slow back roads so aerodynamics is less of a concern.

With that in mind I decided on a VW beetle, they are cheap, spare are easily available new and in breakers yards, lots of folks have already made EV's from them and they are relatively light. They are also mechanically simple to work on and made of steel, so easy to work with and weld in things like battery boxes etc.

To keep my costs low I initially was thinking I would use a multitude of large model aircraft brushless motors, 3 or 4 of them with toothed belts to run them in parallel. However they aren't as robust as real EV motors and there are problems trying to get reliable inverters to power them. Model controllers are designed to run propellers, very low current and torque at low rpm and steadily rising with rpms. A heavy EV like the beetle would pull enormous current from a standstill and overload them.

I started to think of other options and thought about the honda IMA motor, its robust and has a reliable controller, they are available on ebay and craigslist etc. The only issue with them is either faking signals to the MCM to use a simple drive signal like the MIMA, or design a controller that will send a drive signal directly to the MDM.
As a fall back position if I can't get the honda controller to work, I can always buy a brushless EV controller, I looked into it and they should work with the honda IMA.

I don't have a bug yet as I want to get my engine and transmission all up and running in my workshop before I purchase one, it might be next spring before I do so. But I did start a blog on my build process. I have a transmission and just bought 2x honda IMA's and controllers off ebay.
Here's my blog link:

Electric bug build

Thanks!

P.S. How do you change a user ID? I don't live in MA anymore, I'm in CA now..

[retepsnikrep]

Good luck with your build. We shall be watching with interest. Please post regular updates on here. Do a few searches on here as well as there are quite a few discussions and posts on using/stacking the IMA motors. ASFAIK No one has actually done it yet. HughFalls started the most recent discussion.

[Chris(MA)]

Originally Posted by retepsnikrep

Good luck with your build. We shall be watching with interest. Please post regular updates on here. Do a few searches on here as well as there are quite a few discussions and posts on using/stacking the IMA motors. ASFAIK No one has actually done it yet. HughFalls started the most recent discussion.

Thanks, I've already started doing a CAD model/assembly to figure out the fabrication. I don't have a lot of the dimensions yet. Mechanically this doesn't look too difficult, electronically it might be harder to do.

[MN Driver]

This looks to be a very interesting project and you've already got components and some progress. I'll be watching this build. We've had a few different threads lately with interest on putting hybrid components into vehicles, one here with Mike D using the Prius motors, a few here with plans for IMA motor integrated into an EV, another where an IMA motor was integrated with an 1.2 liter VW Lupo TDI engine in an Insight. There is an endless-sphere thread in the Large EV forum section discussing the use of the Prius drivetrain and also mentions Hondas IMA system but I think they have the continuous rating listed higher than what I think these motors can really handle. Maybe one of these different resources might provide some inspiration and some extra information that might help out along the way.

It looks like you've got a great start so far!

[IamIan]

Lead might be cheap $$ ... but ... Lead is heavy... from your page you look to be expecting about ~4 Miles per kwh ... you want a ~30 Mile Range ... that is ~7.5kwh usable ... If you did 100% DoD you would have a had time getting a lead pack under about ~400 lbs ... if you only want ~50% DoD you will have a hard time keeping a lead pack under ~800 Lbs ... on your web site you listed ~25% ... I think ~1,600 lbs of lead is too much weight for the bug.

You might have to either use more than ~25% DoD ...
find some ways to improve that ~4 Miles per Kwh efficiency ...
or save up for something with better energy density wh per lbs than lead acid batteries.

I also have an interest in the IMA motor work ... and look forward to future posts on your progress.

[Chris(MA)]

Originally Posted by IamIan

Lead might be cheap $$ ... but ... Lead is heavy... from your page you look to be expecting about ~4 Miles per kwh ... you want a ~30 Mile Range ... that is ~7.5kwh usable ... If you did 100% DoD you would have a had time getting a lead pack under about ~400 lbs ... if you only want ~50% DoD you will have a hard time keeping a lead pack under ~800 Lbs ... on your web site you listed ~25% ... I think ~1,600 lbs of lead is too much weight for the bug.

You might have to either use more than ~25% DoD ...
find some ways to improve that ~4 Miles per Kwh efficiency ...
or save up for something with better energy density wh per lbs than lead acid batteries.

I also have an interest in the IMA motor work ... and look forward to future posts on your progress.

My calcs are based around driving around 35mph, my route to work can either be taken via the freeway or back roads, as the bug is not an aerodynamic car by any means taking the freeway would be heavy on kwh. If I had to drive on the freeway a lot the bug wouldn't make sense.
Other errands around town can be done around 35mph, I might venture on the freeway occasionally but only for a few miles and I'll stick to 55-60mph.

My parts just arrived and I'll take some pics and put them on my blog soon, I've been researching 3 phase motor control, it seems vector control is the way to go but requires more computing horsepower than I anticipated, so I'm looking at another microcontroller to do this.

The odd thing is the hitachi MPU in the MCM of the insight does not seem to have much in the way of performance compared to a modern MPU, I'm wondering if the insight uses a simpler control method instead?

If someone hooked a scope up to read the current traces from the current sensing transformer(s) on the MDM they could probably tell immediately, vector control gives a smoothish sine wave looking trace. I don't have an insight unfortunately.
Anything cruder will tend to look more triangular or square wave.

[retepsnikrep]

Originally Posted by Chris(MA)

If someone hooked a scope up to read the current traces from the current sensing transformer(s) on the MDM they could probably tell immediately, vector control gives a smoothish sine wave looking trace. I don't have an insight unfortunately.
Anything cruder will tend to look more triangular or square wave.

If you look thorugh the advanced diagnostics literature on this site you will find several waveform pictures which may or may not relate to the phase current sensors. They may answer your question.

Download 2000-2006 Insight Service Manual PDF and Others

I spent a good day reading all the blurb recently. Very useful.

[Chris(MA)]

Originally Posted by retepsnikrep

If you look thorugh the advanced diagnostics literature on this site you will find several waveform pictures which may or may not relate to the phase current sensors. They may answer your question.

Download 2000-2006 Insight Service Manual PDF and Others

I spent a good day reading all the blurb recently. Very useful.

Great tip, its vector controlled, here's an image of the current waveforms, slightly distorted sine waves that sum to zero.
Still wondering how they seem to do this with little computing power..hmmm

[retepsnikrep]

Probably/possibly programmed in assembler for maximum speed! Yikes!!! Well out of my league.

There is also fair bit going on inside the inverter as there is another control module in there. I did have a look inside it but can't remember if i took any pics. I'll have a look.

[Chris(MA)]

Originally Posted by retepsnikrep

Probably/possibly programmed in assembler for maximum speed! Yikes!!! Well out of my league.

There is also fair bit going on inside the inverter as there is another control module in there. I did have a look inside it but can't remember if i took any pics. I'll have a look.

I am not overly concerned about microprocessor performance for driving my MDM's, these days there is more than enough for an embedded system like this. I'm leaning towards a microchip pic at this time. Chatting to a guy who does C a lot at work and he says microchip have a lot of libraries etc and community for their products.
They also have a BLDC demo board for $130 with support software, I am sure they have a good selection of library files to simplify programming for motor control.

I got delivery of my motors and parts last night, one of the stators was damaged but the seller has already shipped a replacement which is nice.

A couple of things that look a little more complex than I imagined, the commutator sensors don't work from the rotor, they use the flywheel, can anyone confirm this? I won't have an insight flywheel so they will be useless.
I can embed some hall sensors and I also plan on embedding some temperature sensors in the coils to keep track of coil temps.

The other niggle is the gap (or lack of it) on the back side of the stator housing, I was planning on sandwiching the two housing together but this requires enough room to put in a coupling plate to the two rotors. There doesn't seem to be enough room so I may have to add a plate to space them out.
I am trying to get my seller to sell me the outer aluminum housing on the busted stator, it will make a neat and cheap spacer.