[catphish]

I'm new here so I hope this post is appropriate / not too rambling.

Goal

I am aiming to to build my own high voltage AC inverter/controller with the ability to power an EV with a 400V 30kW industrial AC induction motor. I am also aiming to convert a small car for the purpose of testing my controller. My skillset is primarily software development, so the mechanical parts of the project will be the biggest challenge for me.

Progress

I have built the bulk of the electronics for the inverter. This consists of an IGBT block (with a capacitor and cooling), a driver board, and an arduino to control it: http://i.imgur.com/edeGn7y.jpg

I have also written basic software to run on the arduino to run it: https://github.com/catphish/vfd

This setup works well for my low power (85V, 180W) test motor with no load: http://imgur.com/IIbTXl3

What's next

In order to continue work on this, I will need to scale things up. The electronics should be able to handle a 30kW motor, however I will need to buy a 30kW motor and 500V-600V of batteries to power it all.

I will also need a DC-DC converter that can convert my 600V battery voltage to 14V for the normal low voltage systems.

Questions

I have a few immediate questions:

1) As shown in the photo above, my controller requires timing feedback from the motor. At present this is achieved my a rotary encoder connected to the motor shaft with a timing belt. Will this same setup be practical with a large motor in a car? I am worried about how everything will fit together.

2) Related to (1). When people use industrial AC motors, do they use the simple kind with a fan mounted on the rear end of the shaft? Or a different cooling mechanism / different setup? Almost all the motors I see for sale have a fan on one end and shaft on the other. This makes it very difficult to mount the encoder, and I worry about cooling at low speed.

3) I need to test regenerative braking, ideally without having to test it in a real vehicle, so I'd like to hook up some kind of heavy flywheel to my small test motor. Can anyone recommend anything I can buy to achieve this? My test motor has an 11mm shaft.

4) Where do I get started actually putting this all in a vehicle? I don't know what I'm doing with regard to mounting the motor in an engine bay, building an adapter plate and shaft coupler to connect it to a clutch or gearbox. I'm sure there's a lot on this site I can look through, per personal pointers would be lovely. I haven't bought a victim vehicle yet, but considering a Fiat Seicento for low weight / low cost.

Any answers / pointers / comments would be gratefully received. I would be very happy to post any other information, schematics, answer questions, etc. I'm very new to this.

Thanks in advance, and special thanks to jhuebner from whom I have taken a lot of inspiration for this project.

 

[dcb]

1. nice example there! might need to make throttle volatile, and the pwm frequency might be a bit slow by modern standards, but otherwise easy to digest.

2. typically a 5-20hp motor is used, it can be TENF (totally enclosed, no fan), ODP (open drip proof), whatever. ODP tends to have better cooling but can collect junk, but is easier to add a filtered blower too so cooling isn't so rpm dependant.

3. a smaller motor is used (look at the weights of a 30kw 60hz motor!) and since it is rated at 60 hz, we overclock them to maybe 180hz or wherever the magnetic losses start dominating. aluminum housing is good, or thin steel. not those cast iron ship anchors. might want to look into the ac24ls (google ac24ls site:diyelectriccar.com ), easily makes 30kw and quite a few surplus at the moment ($300 or less), threads on here somewhere, w/180v in delta you can get ~50kw peak, 20kw continuous.

4. there are a couple open source ac kits on this site too, i,e, http://www.diyelectriccar.com/forum...er-nissan-leaf-motor-vehicle-test-171442.html http://www.diyelectriccar.com/forums/showthread.php/another-homebrew-ac-controller-45909.html
with more sophisticated control algorithms.

plus lots of ev/hybrid motors/batteries/controllers/chargers/bms/etc are making their way to the scrap yard for affordable prices, http://www.car-part.com/ ebay craigslist etc.

anyway, happy digging

 

[catphish]

Thanks!

the pwm frequency might be a bit slow by modern standards

Not sure if it's clear from my code but the PWM frequency is 15,625 Hz. This seems to work well at the moment, but I can change it if necessary.

typically a 5-20hp motor is used

Thanks, larger industrial motors certainly seemed very heavy, but I want to make sure I have enough output.

might want to look into the ac24ls

I wouldn't find any of these where I am (UK). I've started a new thread specifically about this.

 

[dcb]

Not sure if it's clear from my code but the PWM frequency is 15,625 Hz. This seems to work well at the moment, but I can change it if necessary.

ah, yah I haven't AVR'ed in a while. If you get to it a comment in vfd.c about how that is determined would be helpful, and a high level schematic so we know what is connected to what pin. 15k is plenty for igbt!

sorry you are on your own with the flywheel, I've had plenty of similar thoughts though. Depending on your free space maybe just an 11mm shaft sprocket chained to a car tire would make a good flywheel dyno with the right ratio and some #25 chain

 

[catphish]

If you get to it a comment in vfd.c about how that is determined would be helpful, and a high level schematic so we know what is connected to what pin.

I've added schematics to my github repo

https://github.com/catphish/vfd/blob/master/arduino.pdf
https://github.com/catphish/vfd/blob/master/driver.pdf

 

[Ivansgarage]

Has anyone ever used a off the self 3-phase industrial motor
in an electric vehicle and it lasted very long?

------------------------------------------------------------------------

 

[dcb]

johannes did, he isn't trying to win any races though, just bumped up the pack voltage to 400+ and shifts gears as appropriate. catphish needs to see your handiwork too! (hint ivan has rewound motors for lower voltage and higher current)

 

[catphish]

Is this johannes' installation? Can anyone explain what is going on in this photo, in particular:

1) What is the motor?
2) Does it have no cooling?
3) Does it have dual shafts?
4) What's it mounted to?
5) Where's the encoder?

Thanks!

ps. I just found Ivan's rewinding tutorials, however since my inverter is only designed for 100A, it would be a bit of a rethink of my whole design if I reduce voltage too much.

 

[brian_]

Can anyone explain what is going on in this photo, in particular:

1) What is the motor?
2) Does it have no cooling?
3) Does it have dual shafts?
4) What's it mounted to?
5) Where's the encoder?

Just from looking at the photo...

1. It appears to be an industrial motor, but no idea beyond that.
2. It has cooling fins, suggesting that it has no forced air (or liquid) cooling - it just radiates heat to the surrounding air.
3. Motors don't actually have dual shafts, but some - such as this one - have the shaft sticking out at both ends.
4. The output face of the motor appears to be bolted to an adapter (machined from a solid block of aluminum) to connect it to the car's original transaxle (transmission with differential); it is also supported by the stock right-side motor mount, through a custom bracket bolted to the top of the motor.

 

[dcb]

it looks similiar to one of these 2 pole 18.5kw motors, with the fan removed and a homemade encoder instead (was originally TEFC, now is dual shaft TENF) . He kept the gearbox so the lower torque of the 2 pole is mitigated.

https://inverterdrive.com/group/Motors-AC/?filter=Power|18kW/18.5kW

I did a little pole comparison on abb motors here: http://www.diyelectriccar.com/forums/showthread.php/number-poles-revisited-again-164921.html

2 pole has more back iron, because the magnetic path is longer. ivan and miz like to start with a 2 pole and rewind it as a 4 pole (to good effect!)

 

[brian_]

it looks similiar to one of these 2 pole 18.5kw motors, with the fan removed...

So it originally had an external fan, on the non-drive end from the drive, blowing along the fins? Interesting.

It seems strange to remove cooling capacity from a motor which is going to pushed past its normal power rating; perhaps the assumption is that the continuous power level will be very low, and peak power will be used very briefly.

 

[brian_]

Is this johannes' installation? Can anyone explain what is going on in this photo, in particular:

...
5) Where's the encoder?

it looks similiar to one of these 2 pole 18.5kw motors, with the fan removed and a homemade encoder instead (was originally TEFC, now is dual shaft TENF) .

Johannes' web page about the conversion includes a video showing the encoder on the non-drive end of the motor shaft, and it appears around 3:02 in the video. I'm sure more information is buried in his thread about the controller, but I haven't gone through all of that. From the controller sections of the same site, the encoder produces a pulse train for the controller, and uses an optical sensor.

 

[catphish]

it looks similiar to one of these 2 pole 18.5kw motors, with the fan removed and a homemade encoder instead (was originally TEFC, now is dual shaft TENF) .

Thanks, this is exactly the information I was looking for (and hoping for)! It occurs to me from looking at the photo that as the motor sits behind the front grill of the car, there's potential to guide the natural airflow over the fins of the motor to cool it. Perhaps Johannes finds this to be adequate in this vehicle if he's removed the fan. This is ideal for me as I was struggling to work out how I'd mount the encoder with the fan taking up the read of the rotor.

I'm also glad you mention 18.5kw as this is around the size I was hoping I'd need. [edit: according to a youtube comment the motor is actually rated at 30kw]

He kept the gearbox so the lower torque of the 2 pole is mitigated.

I realise that the answer probably depends on the choice of gearing, but is it reasonable to assume that more poles will always be better for this application? Should I be considering the 6-pole option when buying new?

 

[dcb]

as you say, "it depends", 4 pole seems most common, 2 pole not uncommon with right gearing, the power density seems to drop off after 4 pole. even the big dc forklift/warp9 motors are 4 pole.

edit fyi: the number of poles is usually a bit more obvious in a dc motor, for reference. Ac you have to look at the speed rating (1500-1800 ish = 4 pole, 3000-3600ish 2 pole) they trade torque for rpm, or on ac you can closely examine the end turns.

4 pole warp 9

 

[catphish]

Thanks everyone for the assistance in my various threads on this subject. I have made some decisions and ordered some bits. These are not the ideal, but I hope they will be appropriate for some initial low-budget testing of my inverter.

2) 30kW 2-pole 400V 3-phase motor in a 160 aluminium frame with a B5 flange and no fan. £500
[edit: I only have the motor, decided against a tiny car]

My intention is to run a battery pack at approx 288v (24 x 12v lead acid, or an old prius pack). This will supply 204V to the motor, and my rough calculations indicate this will provide 100Nm of torque up to 1400RPM, and 15KW of power at any speed above that. This will be a needlessly heavy way to get 15KW, but should be tolerant of my abuse.

Please let me know if I've made any incorrect assumptions with the above calculations.

It is my hope that this motor can also be rewound for 200V 4-pole operation in the future if I need more power. I'll have to ask Ivan.

This battery voltage should provide a sensible balance, allowing me to to use more off-the-shelf parts, particularly the DC-DC converter for the 12V systems.

I will try to produce a complete diagram of my intended setup.

My main interest is still in testing and developing my inverter. This seems to be going well and I'm looking forward to testing it with something more than a 180W motor with zero load, and locking the shaft with my hand.

Things still to consider:

* Everything mechanical, adapter plate, shaft coupler, etc
* Contactors, fuses, pre-charging and safety cut-outs

I'm sure I'll be back with more newbie threads on these topics shortly. Thanks again!

 

[dcb]

sounds about right. got a link for the motor?

fwiw, just pwming a sine wave will get you a rms voltage of ~ .56*pack voltage (161v). if you sort out sVpwm you can get up to .707*pack voltage (208v)

https://www.ijareeie.com/upload/2014/may/25_Comparative_Study_of_SPWM_and_SVPWM.pdf also discussed in johannes thread and elsewhere.

also, you might be able to easily switch it to delta from wye, the voltage requirements go down by sqrt(3) (161vrms) but the current will go up by sqrt(3) (127arms). but surplus igbt are cheap if you need to bump up your switches. with those two mods and a 288v pack you might get 26kw.

 

[catphish]

Thanks dcb!

Unfortunately the motor (like almost all European 3-phase motors) is rated for 400V/690V, so when I mention 400V, this is unfortunately already its minimum possible voltage (in delta configuration).

Similarly, my inverter is already running space vector modulation (SVPWM), so I should be getting approx 200V, so I don't think there's any way to squeeze any more power out without a rewind.

Of course all of this assumes that my inverter is perfect, and that both the motor and the inverter are 100% efficient, which will clearly not be the case. I hope I can achieve a decent portion of the theoretical max 15KW though.

Is there some information around about how much torque and what RPM I actually need in various configurations (normal manual gears, fixed gearing, direct drive)?

Compared to an ICE, 100Nm seems like a reasonable amount of torque, but I'll have to work out what gearing makes sense, particularly as I will only get this torque up to 1400RPM which seems very low compared to ICE.

 

[catphish]

got a link for the motor?

http://www.ebay.co.uk/itm/Electric-Motor-3-phase-30kW-Free-P-amp-P-NEW-/141708339034?rmvSB=true&clk_rvr_id=1275472156044&nma=true&si=bpXFeu4P2wzhORDLG8bwciAj2q8%253D&orig_cvip=true&rt=nc&_trksid=p2047675.l2557

 

[sprocketman]

Thats a lucky find they dont usualy do a 30kW 2pole motor in a 160 frame. I think they are normally a 200 frame and cast iron and weigh about 180kg?
If its 30kW you should have 97Nm up to 3000rpm but they can give 3 times the torque for short periods. How much does this one weigh?

 

[catphish]

Thats a lucky find they dont usualy do a 30kW 2pole motor in a 160 frame. I think they are normally a 200 frame and cast iron and weigh about 180kg?

This did look like excellent value and great size / power ratio. You can get 30kW in an aluminium frame, but it is usually much larger as you say. I'll see what arrives. hopefully the description is correct.

If its 30kW you should have 97Nm up to 3000rpm but they can give 3 times the torque for short periods.

To begin with, I'm planning to run this at with a 288V pack (200VAC). This should permit rated torque (100Nm, 55A) up to 1500RPM.

My current inverter will work up to 70A, so I should be able to push the motor to 125Nm. In the future I will look at rewinding the motor, and upgrading my inverter to 140A.

How much does this one weigh?

No idea yet, I asked the seller and they said the thought it was about 35kg! I'll have to see what it weighs when it arrives, I suspect it'll be more like 90-100.

There's still lots of work to do before I can test anything, mostly mechanical, deciding what batteries to use, and building battery packs / BMS.