PDA

View Full Version : eAssist AC Induction Motor



Nuts & Volts
03 December 2012, 0849
I stumbled across this motor about a week ago. The OSU EcoCAR team has the engine out of the Malibu Hybrid Vehicle that they are using for the next build and I noticed it had the BAS (Belt Alternator Starter) motor on it. I didn't know much about it so I did a bit of research.

The basics

AC Induction 8 pole Radial Flux Inrunner
Water/Glycol Cooled Stator
15-18kW peak
115Vdc Bus
150Nm at 230Arms (in vehicle)
30lbs
4000RPM peak
72 stator windings in wye, 56 aluminum rotor bars
10 pole induction speed sensor


I am hoping to get 50-60kW peak for 5-10sec from this motor. My limitations right now are understanding what speeds I achieve with a specific voltage. I understand how the frequency comes into play, but can't figure out if I can spin the motor above 4000RPM with less than 115Vdc...I will also check the bearings to determine that they are rated to higher speeds.

I am looking at using either a Curtis 1238 Controller or a Sevcon Gen4 Size6 controller. I am concerned that the 300Hz of the Curtis will spin the motor quite a bit slower than the 500Hz of the Sevcon.

Speed = 120*(freq/# of poles)
Speed Curtis = 4500RPM
Speed Sevcon = 7500RPM

What factor does voltage plan thou? It appears that normally you keep the V/Hz ratio the same at all speeds, but this article (http://www.teslamotors.com/blog/induction-versus-dc-brushless-motors) states something about changing voltage to optimize efficiency. The sevcon has field weakening capabilities so that gives it the nod in my book even if it may take a bit of time and money to tune it. I think getting this motor up to 7-8k will make this a killer combo, because losses don't increase as much as a PM motor would with speed.

Thoughts, or Questions? My motor may be here by the end of the week so I can add more data then.

The best bit of information I found was from this video

http://www.youtube.com/watch?v=n2NQ_dO3lMU&feature=g-like

Possible Controllers
http://www.sevcon.com/media/2104/gen4_web.pdf
http://www.curtisinstruments.com/index.cfm?fuseaction=ProductsGrid.ACMotorControlle rs

Motor Sources
http://www.tonkinonlineparts.com/p/__/GENERATOR-Engine-Electrical/6551094/24261691.html
http://www.gmpartseast.com/parts/gm-24261691_generator.html

Thanks to Tony Helms for actually getting me to want to look the motor over more closely :D Probably won't beat a Roadster motor, but I'll try haha

lugnut
03 December 2012, 1150
Thanks, very interesting. The narrator doesn't know much about induction motors. There are 72 teeth in the stator and 56 bars in the rotor. The number of poles is determined by the stator winding. It appears that the coil span would give about 8 poles. And 8 poles would be consistent with your synchronous speed-frequency calculation for the controllers. He also mentions 79 lb.ft. and 20 hp which yields 1330 RPM. Nice looking motor.

Nuts & Volts
03 December 2012, 1238
Thanks, very interesting. The narrator doesn't know much about induction motors. There are 72 teeth in the stator and 56 bars in the rotor. The number of poles is determined by the stator winding. It appears that the coil span would give about 8 poles. And 8 poles would be consistent with your synchronous speed-frequency calculation for the controllers. He also mentions 79 lb.ft. and 20 hp which yields 1330 RPM. Nice looking motor.

Yea I figured he was talking about teeth and bars, but didn't really know what to call them. The paper I linked states that the motor is an 8 pole machine.

This is where I got the peak torque values from http://www.greencarcongress.com/2011/12/eassist-20111215.html

ARC EV Racing
04 December 2012, 0019
The 1238 has field weakening too, just in case that becomes the deciding factor.

Sent from my AZ210A using Tapatalk 2

Nuts & Volts
04 December 2012, 0826
The 1238 has field weakening too, just in case that becomes the deciding factor.

Sent from my AZ210A using Tapatalk 2

Thanks, I am putting together a comparision chart right now of the two controllers.

Is there any difference between the 1238-7601 and the 1238R-7601? I am having trouble figuring that out...

ARC EV Racing
04 December 2012, 0918
I think the 'R' tolerates the higher 96v nominal pack voltage (can be persuaded to work with 115v nominal though). Also it sounds faster :)

frodus
04 December 2012, 1049
The r is a newer Rev. I think its just a name change though. 7xxx series are all 120/130vdc Max depending on the literature you read :rolleyes:. 6xxx are 108 Max.

Nuts & Volts
04 December 2012, 1202
The r is a newer Rev. I think its just a name change though. 7xxx series are all 120/130vdc Max depending on the literature you read :rolleyes:. 6xxx are 108 Max.

Thanks Travis. The more I look at it I am liking the Curtis over the Sevcon. I have a spreadsheet that I will throw up comparing the models.

Still trying to determine the Max speed thing.
Does the field weakening allow the motor to spin faster than the 4500RPM that the Curtis frequency is limited to? That is before the BEMF limits RPM? From my understanding it seems that even with field weakening the Curtis controller will not be able to switch quick enough to spin the motor any faster regardless of the BEMF.

IE a 150V, 300Hz and a 300V,300Hz control will only spin this motor to 4500RPM. Thou the base speed will be doubled on the 300V, which will still double power...

Nuts & Volts
04 December 2012, 1247
Here is the chart I created to compare the controller options.

3870

I am leaning a little towards the Curtis because it has higher power capabilities and will probably be easier to set-up. May try to HV version depending on what the final price and specs look like.

lugnut
04 December 2012, 1251
Still trying to determine the Max speed thing.
Does the field weakening allow the motor to spin faster than the 4500RPM that the Curtis frequency is limited to? That is before the BEMF limits RPM? From my understanding it seems that even with field weakening the Curtis controller will not be able to switch quick enough to spin the motor any faster regardless of the BEMF.

IE a 150V, 300Hz and a 300V,300Hz control will only spin this motor to 4500RPM. Thou the base speed will be doubled on the 300V, which will still double power...

The maximum speed is set by the controller; the max frequency minus a few percent slip. The winding in the motor determines a volts per hertz ratio. The battery voltage determines the frequency at which this ratio can no longer be maintained (base frequency). At frequency above base, field is weakened and torque decreases. But the speed continues to follow the frequency. Eventually a frequency would be reached where there is insufficient torque to maintain speed but that is likely several times the controller limitation.

In your example, doubling the voltage does not change the volts/hertz but doubles the base frequency and would double the maximum power if the V/f ratio was greater than 1:1.

Nuts & Volts
04 December 2012, 1347
Thanks lugnut, so it works in a similiar manner to a PMAC motor and its KV (RPM/V) constant.

I will reread your post a couple of times and work out an example and hopefully it will stick :D

ARC EV Racing
04 December 2012, 1450
The r is a newer Rev. I think its just a name change though. 7xxx series are all 120/130vdc Max depending on the literature you read :rolleyes:. 6xxx are 108 Max.

I was just going off the datasheets. The one for the 1238 doesn't list the -7xxx variant but the 1238R only lists the -7xxx. I just noticed that my version of the 1238 datasheet is from 2005 though so I guess it's there on the later versions.

Nuts & Volts
05 December 2012, 2008
Just threw the money down for the motor. Hopefully this source will actually be able to ship me a motor. Said 3-5 days so hopefully will have something next week.

A few shots of the motor connected to the GM engine. (I don't have access to use this one unfortunately)
388838893890

Allen_okc
06 December 2012, 0529
is that a coolant line going into the motor???

Allen_okc
06 December 2012, 0535
sorry my bag, i just watched the video, a very interesting concept...

Nuts & Volts
06 December 2012, 1136
Source two ran into the same issue as my buddy at the dealer. It appears that GM really doesn't want anyone using any of these motors. I'll keep snooping around trying to get my hands on one of these.

Also tried Continental AG whom is the manufacterur of the motor. They only sell to OEMs.

Only option will be a crashed vehicle which GM probably snatches up themselves. I have one other source but the chances are very very slim

:( My holy grail AC motor journey will continue. Thou I did get a chance to learn a lot about the Curtis controllers and how awesome they are!

frodus
06 December 2012, 1432
Wtf. That's just stupid. I mean they're still making money!

Yeaw
02 January 2013, 1317
The dealers won't let you order the motors at all? Or they won't let you order them unless you prove that you own the vehicle?

Nuts & Volts
02 January 2013, 1345
Either you have to own the vehicle or I think more likely you have to exchange the old part for the new one.

Athlon
02 January 2013, 1712
:( My holy grail AC motor journey will continue. Thou I did get a chance to learn a lot about the Curtis controllers and how awesome they are!


maybe in the future you will have the chance to use Curtis 1839 not only with induction motors ;) , but I'm not say anything

Yeaw
04 January 2013, 0638
Kyle, thanks for sharing the great idea. I was looking for a good watercooled ACIM to test the control algorithm for my thesis project. This one looks pretty promising.

Used motors are available from auto wreckers. I searched http://car-part.com/ and found one from a 2012 Buick Lacrosse 2.4L eAssist for $155 shipped. There was no core required or core charge. It also comes with a lifetime warranty from the wrecker. As soon as I receive it, I will weigh it and I will hopefully fully test it once my controller is finished.

frodus
04 January 2013, 0943
Thanks dude! Looks like they have a few!

Nuts & Volts
04 January 2013, 1109
Great find yeaw!...decisions...decisions. I'm never going to actually pick an AC system now haha

liveforphysics
04 January 2013, 1503
Keep in mind, "field weakening" is just a legacy term from sep-ex days, and in the case of a PM motor it's nothing more than timing advance (so you switch the voltage to the coil a few degrees before the magnet passes by the coil generating that BEMF that would have made it impossible to continue feeding current into that coil (because BEMF and packV would be equal), so you can continue to rev at the cost of efficiency and torque. In an induction motor, it simply means increasing the slip rate target, which basically does the same thing. It's not like it has a magic trick to make it rev way higher for a given voltage, it's more like a efficiency/heat/torque costing neat trick to get some extra rev range. It's very cool and has it's place in a motor/controller combo for sure, but I just want to make sure you're aware that it's not like it's just free candy or something.

On ES, a guy named Burtie made a little stand-alone box that takes the hall sensor encoder signal, and applies as much timing advance as RPMs climb as you wish to have. It's like $25 IIRC, and it gives any motor controller "field weakening", it's even programmable to set the curve up how you like. He has sold a few dozen to happy customers for ebikes that get another 10-15mph top-speed at the cost of efficiency/heat/torque, but it ONLY costs them the efficiency/heat/torque while they are operating the motor in the speed range they wouldn't have been able to achieve otherwise. To make a stand-alone box to add "field weakening" to any induction motor/controller would be even more simple, as the encoder is generally simply a rate encoder rather than position encoder.

CliC
05 January 2013, 0746
Ah, thanks for explaining the history. I had never heard the term "field weakening" before getting into elmotos. With industrial AC motors, I always thought; of it as just running above synchronous (60 Hz) speed.

Sent from my ADR6425LVW using Tapatalk 2

lugnut
05 January 2013, 0904
Ah, thanks for explaining the history. I had never heard the term "field weakening" before getting into elmotos. With industrial AC motors, I always thought; of it as just running above synchronous (60 Hz) speed.

You mean base frequency (or base speed). If an induction motor runs faster than synchronous speed, it is a generator. But you are correct. When you use a VFD and drive a standard industrial motor above its 60 Hz speed, that is field weakening. And does not need an encoder for most industrial applications.

CliC
05 January 2013, 1212
[QUOTE=lugnut;34722]You mean base frequency (or base speed). If an induction motor runs faster than synchronous speed, it is a generator. But you are correct. When you use a VFD and drive a standard industrial motor above its 60 Hz speed, that is field weakening. And does not need an encoder for most industrial applications.[quote]

Yeah, it's not a generator when you are driving it at over 100% of its across-the-line speed with a VFD (though we rarely do that). But yes, "synchronous" is a somewhat sloppy term for induction motors because of slip. Maybe I should've said "1750 rpm" :)

chris_rms
07 January 2013, 2044
In looking at the video the speed sensor looks to me a lot like a Tamagawa resolver. It's hard to tell from the picture how many poles it is, but I doubt it is 10. If the motor is 8 poles then you wouldn't want to have a 10 pole resolver. I find it a bit odd that they would use a resolver on an induction motor. I wonder if they have planned to have a different rotor that had magnets???

Does the Sevcon unit actually support a resolver (provides excitiation to resolver) or does it support a sin/cos encoder? I have heard that on some models they don't support a true resolver. Since the Curtis doesn't support a resolver you would have to add your own speed sensor.

The motor looks pretty cool. It would be interesting to put it on the dyno and see what it can do with ~ 320V.

Athlon
08 January 2013, 0515
AFIK sevcon have input for 3 hall sensor , sin/cos encoder and a quadrature encoder (A B signal square wave) ,

Curtis only support quadrature encoder as they only support induction motor

frodus
08 January 2013, 1029
I still have not gotten a response from any of the parts dealers about the motor.....

chris_rms
08 January 2013, 1036
AFIK sevcon have input for 3 hall sensor , sin/cos encoder and a quadrature encoder (A B signal square wave) ,

Curtis only support quadrature encoder as they only support induction motor

Then from what I can tell neither the Sevcon or the Curtis would support this motor without changing the speed sensor. A sin/cos encoder is different than a resolver. The RMS drive does support a true resolver as this is what is used on many of the motors we support (EVO, Remy, etc.). If someone does get one it would be interesting to see what it can do.

lugnut
08 January 2013, 1332
I've heard of a device which will excite the resolver and transform the signal into a encoder type pulse train for use in such cases. I have it bookmarked somewhere ;-)

Being an induction motor, it would be easy to run open loop (not that you'd actually want to use it that way on a bike). I don't think Curtis supports V/f or open loop vector. Don't know about Sevcon. Certainly standard industrial drives do.

Athlon
21 January 2013, 1531
I don't think Curtis supports V/f or open loop vector. Don't know about Sevcon. Certainly standard industrial drives do.


You can use a Curtis in Open loop by using the motor auto tuning menu and by diabling the encoder error detection.

Just go in the motor tuning menu , choose the RPM and the max current and start the autotuning , the motor will run in open loop mode at fixed RPM

Yeaw
22 January 2013, 1240
I received the motor today, somehow my buddy jammed it into a USPS flat rate box, haha. It weighs 30lbs dry. Here are some pictures:
404640444045

liveforphysics
22 January 2013, 1439
Nice little motor package. It's exciting to see hybrid car parts getting directly re-purposed. That's the very highest efficiency form of recycling, direct re-purposing. Cool little motor.

Nuts & Volts
29 January 2013, 1008
I think I was finally able to place an order for 1 of these motors from a salvage yard. Will report back next week sometime when it arrives... Paid about $480 out the door. Thoughts are to purchase a Curtis to test out its functionality. I will also use an insulation resistance tester to determine what type of voltage the motor might be able to handle. If capable of handling ~400V I will look at getting the RMS PM100 tuned to motor. Probably not quite 100kW capable, but 70kW peaks/15kW cont would be killer for this size motor.

Otherwise I am hoping it is a replacement for an AC-20 at about half the weight and size...

Nuts & Volts
03 February 2013, 0907
I received my motor on Friday. It weighs in at 13.15kg (or 29lbs). It measures about 180mm (7.1") diameter and 125mm (4.9") long. Then terminals add about another 2.5" extended block off the rear and cooling ports stick off the side and rear slightly. The mounting will be slightly unconventional for the motor, but won't be too hard.

http://i194.photobucket.com/albums/z160/buckboom00/20130131_174805_zpsaccbd926.jpg

The particular motor I got is a bit damaged (I assume it came out of a wrecker, so kinda expected). Basically one of the cooling tubes a slightly bent and a tab for the terminal cover is cracked. I should be able to weld the tab back on and I may get some custom cooling ports installed later on. Rotor spins freely so the important bits are in shape. As you can see from the pictures some water got into the end and rusted slightly around the speed sensor and other parts there.

http://i194.photobucket.com/albums/z160/buckboom00/20130201_153437_zps12c443ef.jpg

The speed sensor is a 6 wires. It has ten windings with a oval shaft in the center. Anyone know what type this is? I assume some sort of encoder. I think I will just install my own temperature sensor, but will check to see what type is installed in the motor.
http://i194.photobucket.com/albums/z160/buckboom00/20130201_153806_zpsc5a22405.jpg

Still trying to figure out how to get the pulley off the shaft thou, any tips?

Yeaw
04 February 2013, 1106
Congrats on receiving your motor N&V!

As chris_rms said earlier, it is a resolver, not an encoder. They probably used a resolver because it is more tolerant of heat and vibration than an encoder - important for something strapped to the side of an ICE. I am not sure what the model of it is, but I would guess that the black and beige wires are + and - Excitation for the rotor, and the other 4 are the + and - sin and cos signals coming from the stator. Almost all resolvers have 6 leads. We can use voltmeter to ring out the leads and determine which 3 pairs of wire are circuits. Then record the resistance of each circuit. The Cos and Sin circuit will have the same resistance, so the circuit with the different resistance will be our excitation circuit. We will need to figure out what the voltage and frequency the excitation needs, is it labeled with anything (I haven't opened my motor up yet)? Most resolvers are between 4-6V at 3-10KHz.

Have you decided to go with the Curtis for testing? Have you thought about going with the Tumanako setup? You could help our development effort, I am writing a new control algorithm for it after all.

Nuts & Volts
04 February 2013, 1206
Congrats on receiving your motor N&V!

As chris_rms said earlier, it is a resolver, not an encoder. They probably used a resolver because it is more tolerant of heat and vibration than an encoder - important for something strapped to the side of an ICE. I am not sure what the model of it is, but I would guess that the black and beige wires are + and - Excitation for the rotor, and the other 4 are the + and - sin and cos signals coming from the stator. Almost all resolvers have 6 leads. We can use voltmeter to ring out the leads and determine which 3 pairs of wire are circuits. Then record the resistance of each circuit. The Cos and Sin circuit will have the same resistance, so the circuit with the different resistance will be our excitation circuit. We will need to figure out what the voltage and frequency the excitation needs, is it labeled with anything (I haven't opened my motor up yet)? Most resolvers are between 4-6V at 3-10KHz.

Have you decided to go with the Curtis for testing? Have you thought about going with the Tumanako setup? You could help our development effort, I am writing a new control algorithm for it after all.

Yea I've worked with a resolver on the EMRAX motor and understand its basic operation. Started thinking that might be what it was. Unfortunately taking a closer look at my resolver I think some of the windings are destroyed...maybe I can rewind them...or I'm going to get another resolver and custom install it on the shaft. Either way that will wait until after I pull it all apart and inspect everything else. At some point I may try to use an o-scope on the outputs of the resolver, but if mine's damaged.

I haven't decided yet how I'm going to run this. After buying the Rinehart and this motor I don't have the funds for a Curtis (unless I sell my AGNI/Alltrax combo). Come August I start full time, so funds will be good then. I may just stick my Agni and Alltrax back on my R6 and then tinker with this motor off the bike. Not deciding anything until I get this motor apart and can inspect that it will function...

Nuts & Volts
09 February 2013, 1011
Some really good news. I tested the insulation resistance yesterday with a nice Fluke meter. We tested first at 273V (that's just what the meter used) and resistance was somewhere in the 6 mega-ohm range. This was with one clip on the ring terminal you can see in the pictures and the other one around the insulation of a winding. At 1000V we got about 6 mega-ohms as well. Winding to the outer case was like over 100 mega-ohms. Sorry I don't have exact numbers, but I didn't really think it mattered too much.

In my opinion I think we can say this motor can be run on up to 500V systems very safely.

Also something to note is that each tooth has its own winding and the full filled slots of the motor windings when that they never have contact between two windings of two different phases. This means that you need to arc across to short two phases throughout most of the motor.

It also makes sense that the motor can handle more voltage because it is rated for 18,000 RPM, but only does 4000RPM on 115V. So 400V should be well over 10,000RPM. It must have some very nice bearings in there!

Yeaw
09 February 2013, 1340
Great news N&V. I think this is going to prove to be a great motor to run for motorcycle applications.

Nuts & Volts
21 February 2013, 2047
Brute Force for the win! Finally go the the resolver out of the motor. It's just pressed fit in, but there is not much room for leverage. I manage to pry it out without damaging it too much. I think it may actually still work, but I want to clean the rust out anyways. The bearing looks to be in good shape which is what I was worried about the most. Plenty of room to add an encoder to run with the Curtis (and maybe just keep it for use with HV controllers.

http://i194.photobucket.com/albums/z160/buckboom00/20130221_192800_zps08cc5d88.jpg

http://i194.photobucket.com/albums/z160/buckboom00/20130221_192811_zpsf6c91134.jpg

http://i194.photobucket.com/albums/z160/buckboom00/20130221_192824_zps6b46d96e.jpg

Still working on getting the pulley then front cover off the motor to get it all apart.

Nuts & Volts
23 February 2013, 0741
Alright so I have been looking at Encoders from US Digital. Since I will be using a Curtis controller with this guys first. There are 3 speed limits; firmware 8000RPM, Elec frequency limit (36000/8poles = 4500), and Encoder CPR (600000/CPR=XXXX). What encoder CPR should I get if I know will be wanting to use a new controller in the future? I want at least want CPR=100 max (6000RPM) so that it won't limit anything. Is there any reason to go lower (96, 50, or 32)? I want to try to use the motor at close to 10,000RPM in the future.

Oh I think this is the encoder I will be getting http://www.usdigital.com/products/encoders/incremental/rotary/kit/E2

Looking at the Sevcon Gen4 Size8 the lower encoders allow a higher RPM possible. Not the controller I am using, but I think I will go with the CPR=50

Anyone have any insight? (like the Rinehart PM100DX ideal encoder CPR)

Yeaw
16 March 2013, 0753
N&V have you thought about actually using the resolver? That is my plan, I am planning on using a AD2S1205 R/D Converter on the development board that Analog Devices sells to convert the signal to a 1024 line encoder signal. So I am getting the EVAL-AD2S1205SDZ and EVAL-SDP-CB1Z boards from Digikey.com and then just hooking up the resolver, and then outputting the A, B, and NM signals to my controller. The resolver is much more abuse tolerant than an encoder will ever be. I think it is what Toyota actually uses for their motor speed sensors - and obviously what GM uses.

It took me way to much internet searching, but I found the female half of the plug that connects to the case of the motor and outputs the resolver signals:
Molex MX150 12 pin Female Plug 33472-1201
Molex MX150 18-20AWG Female Connectors for the Female Plug 33012-2002
Molex MX150 Cavity Plugs for the unused connector pins 34345-0001
Both Mouser.com and Digikey.com sell the connectors

Nuts & Volts
16 March 2013, 1033
N&V have you thought about actually using the resolver? ...

Yea a little bit. I actually think it will be a lot better plan to try that. Two issues right now. Testing that the resolver still functions (visual inspection looks good). Know any good ways to test the resolver signals? Maybe I can do some more reading soon.

And then the cost involved in getting those boards is also a concern. Any cheaper options??

Pretty sure my Curtis Controller just arrived so I almost have all the bits that I need to get the bike working. I think time will be very limited in the next few weeks thou...

Nuts & Volts
16 March 2013, 1404
Think this guy would work http://pico-systems.com/resolver.html?

lugnut
16 March 2013, 1954
Think this guy would work http://pico-systems.com/resolver.html?

Is that 4096 ppr? It might be too high for the Curtis. Too high of a ppr and high RPM can get into frequency response problems with some controllers. Yep, just checked the Curtis manual:


The maximum encoder frequency the controller will accept is 10 kHz.
To determine how fast this constraint will allow your motor to spin, use the
equation
Max Motor RPM = 600000 / Encoder Size
(e.g., a motor with a 128-pulse encoder can run up to 4687 rpm).

Ohm6
18 March 2013, 0839
Think this guy would work http://pico-systems.com/resolver.html?

Max RPM of the pico-system is only 1000 and as lugnut pointed out is way too high of a resolution for the Curtis.

I am also thinking about using this motor and wanted to keep the existing resolver. Looking around I came up with two options not yet mentioned.

1. An arduino powered resolver converter which has already been developed here: http://wiki.linuxcnc.org/cgi-bin/wiki.pl?ResolverToQuadratureConverter

2. Tritium sells "motor interface boards" to adapt any motor to their inverter. It looks like type 3 or 7 is just what you need. It's worth a shot to see if they will sell one to you without bying the inverter.
http://tritium.com.au/products/wavesculptor200-motor-inverter/
(near the bottom)

Both of these are probably cheaper than the development boards.

Nuts & Volts
18 March 2013, 1030
Max RPM of the pico-system is only 1000 and as lugnut pointed out is way too high of a resolution for the Curtis.

I am also thinking about using this motor and wanted to keep the existing resolver. Looking around I came up with two options not yet mentioned.

1. An arduino powered resolver converter which has already been developed here: http://wiki.linuxcnc.org/cgi-bin/wiki.pl?ResolverToQuadratureConverter

2. Tritium sells "motor interface boards" to adapt any motor to their inverter. It looks like type 3 or 7 is just what you need. It's worth a shot to see if they will sell one to you without bying the inverter.
http://tritium.com.au/products/wavesculptor200-motor-inverter/
(near the bottom)

Both of these are probably cheaper than the development boards.

Thanks for the info. I have a feeling the Arduino solution won't be any better than the Pico.

Funny thing is I actually have experience with the Tritium Controller and the type 3 interface board. We also happen to have an extra board sitting around. I sent an email off to Tritium to ask whether the board could be used without the Tritium WS200 controller. I will report back.

Ohm6
18 March 2013, 1642
Thanks for the info. I have a feeling the Arduino solution won't be any better than the Pico.

Funny thing is I actually have experience with the Tritium Controller and the type 3 interface board. We also happen to have an extra board sitting around. I sent an email off to Tritium to ask whether the board could be used without the Tritium WS200 controller. I will report back.

That is too perfect with the Tritium board. What are the chances of just having an extra.


It handles 3 resolver channels and has been tested to 1600 rpm (the
max speed of my cordless drill). Bench testing indicates that the
quadrature pulses should keep up up to 3750rpm. For higher speeds it
would be necessary to run fewer resolver channels.

Apparently, according to the developer, the little arduino can do pretty well for itself. With only the one channel rather than 3 it might just keep up.

Nuts & Volts
19 March 2013, 1306
Yea the spare one is because I wired it backwards the first time around and messed up the temperature sensor side of board (which conveniently isn't needed for this mod!).

Got a reply from James at Tritium about using the board in another application. This is the reply I got

"It certainly does perform that function, with a 1024ppm encoder emulation.

But there's a bit more to it than that, it does some 'smart' stuff during startup which might cause trouble for your other controller if it's not expecting it. Basically, on startup, it generates an index pulse (no matter where the motor is) and then enough A-B steps to make it look like the encoder has rotated around to match the physical position of where the resolver is. So this way, your encoder position count register will be correct, and roll over in line with the index pulse in future revolutions of the motor.

None of which you probably care about for an induction motor, since normally you're just looking for speed, not position (unlike an IPM motor, where you care about both). But the 'burst' of pulses at startup might confuse a controller that isn't expecting it."

Unfortunately I don't have much free time (or focus in my final 5 weeks of college). So I may not get to testing something like this out until late April or June. But it seems like a good option.

EDIT: Crap it sounds like the Tritium board won't work well either because the 1024ppm is way to high for the Curtis. Darn

Ohm6
19 March 2013, 1806
Yea the spare one is because I wired it backwards the first time around and messed up the temperature sensor side of board (which conveniently isn't needed for this mod!).

Got a reply from James at Tritium about using the board in another application. This is the reply I got

"It certainly does perform that function, with a 1024ppm encoder emulation.

But there's a bit more to it than that, it does some 'smart' stuff during startup which might cause trouble for your other controller if it's not expecting it. Basically, on startup, it generates an index pulse (no matter where the motor is) and then enough A-B steps to make it look like the encoder has rotated around to match the physical position of where the resolver is. So this way, your encoder position count register will be correct, and roll over in line with the index pulse in future revolutions of the motor.

None of which you probably care about for an induction motor, since normally you're just looking for speed, not position (unlike an IPM motor, where you care about both). But the 'burst' of pulses at startup might confuse a controller that isn't expecting it."

Unfortunately I don't have much free time (or focus in my final 5 weeks of college). So I may not get to testing something like this out until late April or June. But it seems like a good option.

EDIT: Crap it sounds like the Tritium board won't work well either because the 1024ppm is way to high for the Curtis. Darn

I think you could use a digital counter on each quadrature channel as a frequency divider to bring the output in line with what the Curtis can handle.

Something like a DM7490 can divide the number of pulse by 10 giving a manageable pulses per revolution of 102.4. With the Curtis limited to 10kHz, that gives you a max RPM of 5860 (right?).

I have never used a digital counter so I may be way off on it actual capabilities but it’s data sheet says it is good up to 42MHz.


EDIT: Having a fraction of a pulse per revolution probably isn’t the best idea. My new suggestion would be to use a 74LS93 4-bit serial counter to divide the frequency by 16 giving 64 pulses per revolution.

Yeaw
21 March 2013, 1103
I removed the pulley today. It is a 7 rib belt pulley and the ID is tapered. There is a locking ring that fits between the pulley and the shaft with a tapered OD and I think 17mm inside diameter (I don't have micrometers handy to measure it). The pulley is held on by a nut that I removed with an impact wrench and then I used a large puller and a hammer to remove the pulley. The correct tool would a small puller that grabs the locking ring with the puller arms.

One downside is that the shaft does not have a keyway in it, it is flat all the way around. So to get a chain sprocket on it I think I am going to have to bring it to a machine shop and get a keyway cut in it. I can't think of any other ideas, besides making my own ghetto keyway but I don't think that will be worth my time and definitely won't come out good.

Diethelm Rust
21 March 2013, 1133
I removed the pulley today. It is a 7 rib belt pulley and the ID is tapered. There is a locking ring that fits between the pulley and the shaft with a tapered OD and I think 17mm inside diameter (I don't have micrometers handy to measure it). The pulley is held on by a nut that I removed with an impact wrench and then I used a large puller and a hammer to remove the pulley. The correct tool would a small puller that grabs the locking ring with the puller arms.

One downside is that the shaft does not have a keyway in it, it is flat all the way around. So to get a chain sprocket on it I think I am going to have to bring it to a machine shop and get a keyway cut in it. I can't think of any other ideas, besides making my own ghetto keyway but I don't think that will be worth my time and definitely won't come out good.

There are pulleys available called "taperlock" pulleys. They are designed to mount on plain shafts. Google will explain further.

Yeaw
21 March 2013, 1145
There are pulleys available called "taperlock" pulleys. They are designed to mount on plain shafts. Google will explain further.

Yes, I have seen those. The problem is I don't want to fit a pulley, I would like a sprocket to hook to the rear wheel via a chain. I haven't seen taperlock sprockets.

Diethelm Rust
21 March 2013, 1212
Yes, I have seen those. The problem is I don't want to fit a pulley, I would like a sprocket to hook to the rear wheel via a chain. I haven't seen taperlock sprockets.

Ok, one option would be getting a taperlock bolt on flange to which you could mount the sprocket... taperlock sprockets are available. But then again, getting a keyway cut, as you suggested, is probably way cheaper and the best option.

Nuts & Volts
21 March 2013, 1350
Use one of these and machine out a sprocket bore

http://www.fennerdrives.com/trantorque/_/Trantorque-OE-17mm/?s=cHw1
http://www.fennerdrives.com/trantorque/_/Trantorque-GT-17mm/?s=cHwyOmZ8MjMxOA%3d%3d

Might not work with my Curtis if I drive 300Arms+ which will get close to the ~200Nm limit

Yeaw
21 March 2013, 1417
Use one of these and machine out a sprocket bore

http://www.fennerdrives.com/trantorque/_/Trantorque-OE-17mm/?s=cHw1

So it looks like if I got one of these (http://www.globalindustrial.com/p/motors/sprockets/finished-bore-sprockets/sprocket-08bs20hx32-metric-12-pitch-32mm-finished-bore-20-teeth) 32mm bore sprockets, I wouldn't have to do any modifications at all. That definitely looks like it may be a pretty good option. I think finding a machine shop where I am right now is going to be challenging.

Nuts & Volts
21 March 2013, 1435
Yea that would work out well without needing any machining work. 1/2" pitch is a #40 chain 5/8" pitch is #50 which is the same as 530 motorcycle chain

Could also go with one of these if you really wanna pull some torque! edit: nevermind units were wrong its only 153 ft-lbs (207Nm)
http://www.grainger.com/Grainger/CLIMAX-Keyless-Bushing-38G684
http://catalog.climaxmetal.com/item/locking-assemblies/metric-locking-assembly-series-c192/c192m-17x38?

Ohm6
21 March 2013, 1846
Use one of these and machine out a sprocket bore

http://www.fennerdrives.com/trantorque/_/Trantorque-OE-17mm/?s=cHw1
http://www.fennerdrives.com/trantorque/_/Trantorque-GT-17mm/?s=cHwyOmZ8MjMxOA%3d%3d

Might not work with my Curtis if I drive 300Arms+ which will get close to the ~200Nm limit

The published 150Nm number is ďengine torqueĒ. This means it gets measured after the motor/crankshaft pulley reduction. Looking at the sizes of the pulleys I would guess itís about a 2.5:1 reduction (anyone capable of measuring the crank pulley?). This would mean that the motor is only actually putting out about 60Nm of torque at 230 Amps. Crank it up to say 400 Amps and assume your efficiency remains the same (not a safe assumption just throwing it out for quick math) and you would be up to 104 Nm, still giving you a safety factor of 2 for the Trantorque.

Nuts & Volts
21 March 2013, 1905
The published 150Nm number is “engine torque”. This means it gets measured after the motor/crankshaft pulley reduction. Looking at the sizes of the pulleys I would guess it’s about a 2.5:1 reduction (anyone capable of measuring the crank pulley?). This would mean that the motor is only actually putting out about 60Nm of torque at 230 Amps. Crank it up to say 400 Amps and assume your efficiency remains the same (not a safe assumption just throwing it out for quick math) and you would be up to 104 Nm, still giving you a safety factor of 2 for the Trantorque.

Wow you're totally right here. I have a torque speed curve of the eAssist but never thought that it would be after the reduction. So this means it might spin up to 10k rpm with only 115V. This changes my gearing all around, thanks for pointing that out.

I can measure the pulley diameters tomorrow.

The document I have says the motor is good for >60Nm. So that transtorque will work great like you said.

Ohm6
21 March 2013, 2005
Wow you're totally right here. I have a torque speed curve of the eAssist but never thought that it would be after the reduction. So this means it might spin up to 10k rpm with only 115V. This changes my gearing all around, thanks for pointing that out.

I can measure the pulley diameters tomorrow.

The document I have says the motor is good for >60Nm. So that transtorque will work great like you said.

A couple of things to consider:

1. With 64 pulse per revolution (1024 from the Tritium board divided by 16) the Curtis reaches itís 10 kHz limit at 9375 RPM.

2. This being an 8 pole motor and the Curtis having a max 300Hz frequency you max out at 4500 RPM and then go into that whole field weakening thing (this thread being the first Iíve ever heard of it, sorry I canít offer any help). What does that do to your efficiency as you continue up in speed? I donít think it is meant to take you past twice the normal operating speed, but I could be wrong (and I hope I am).


I think 10,000 rpm may be too high a speed with the present hardware. Iím hoping to get 6,000 out of it without going higher voltage and frequency. Assuming you can get 100Nm out of it, I was thinking a 14 tooth front sprocket and a 52 tooth rear sprocket would be a good combination. That gives you a 3.71 gear reduction for a peak torque of 280 ft*lb and still able to go 116 mph at 6000 rpm. I havenít put together a road load simulation yet to know if 116 is doable but if a 23.5 kW (crank) ninja 250 can do 100, it should be close.

lugnut
21 March 2013, 2142
2. This being an 8 pole motor and the Curtis having a max 300Hz frequency you max out at 4500 RPM and then go into that whole field weakening thing (this thread being the first Iíve ever heard of it, sorry I canít offer any help). What does that do to your efficiency as you continue up in speed? I donít think it is meant to take you past twice the normal operating speed, but I could be wrong (and I hope I am).


Field weakening does not affect the frequency/RPM relationship. The maximum with the 300 Hz limit will be 4500 RPM.

Ohm6
22 March 2013, 0439
Field weakening does not affect the frequency/RPM relationship. The maximum with the 300 Hz limit will be 4500 RPM.

Ah, I misunderstood. I thought it was some electrical black magic to simulate a higher frequency. I see your post back on page 1, that makes sense.
Since AC Induction motors have a steep decline in torque as you approach synchronous speed, what do you think top speed should be geared at, 4200 RPM?

Nuts & Volts
22 March 2013, 0505
Yea I meant that you could get 10,000 RPM at 115V if you had the frequency ability (the BEMF won't be the limit). I am still aiming for a ~4500RPM as the top speed I can run with the Curtis due to the 300Hz and 8 pole parameters.

The PM100DX may be able to get the motor to a much higher RPM with the combination of a higher voltage and a higher frequency.

lugnut
22 March 2013, 0744
Since AC Induction motors have a steep decline in torque as you approach synchronous speed, what do you think top speed should be geared at, 4200 RPM?

Strange qualifier to the question. You're always running at slip below synchronous if you're doing any work with the motor so that doesn't enter into gearing choice. As far as gearing this motor, it remains to be seen what it can actually do. Best case would be a 12 to 15,000 RPM maximum with base speed at around 8 or 9000 and 4 to 500 amp capability. There you'd have to use a primary reducer with the chain as the secondary. But that would be a sweet little motor package.

Nuts & Volts
22 March 2013, 0905
Just measured the pulleys on the eAssist motor. The motor has a 3" pulley and the crankshaft has a 6.5" pulley. So at 150Nm crank that means the eAssist electric motor is putting out about 69.23Nm lets call it 70Nm. So 70Nm at 230Arms peak (assumption based). That would mean 167Nm (ideal) at the 550Arms of my Curtis. Don't plan to try to run it that high maybe even, but I will work my way up.

Also base speed is at about 1100RPM of the crank and peak is 3500RPM crank. So this means at 115V (@230Arms) the base speed is ~2380RPM and peak is at ~7580RPM as it is used in the eAssist. Theoretically at 350V this would be a base speed at ~7200RPM with the appropriate frequency.

lugnut
22 March 2013, 2111
Just measured the pulleys on the eAssist motor. The motor has a 3" pulley and the crankshaft has a 6.5" pulley.--snip-- Theoretically at 350V this would be a base speed at ~7200RPM with the appropriate frequency.

Just a thought. If you can control this motor and use 350V, you may want to keep the pulleys to use as the primary reduction on the bike, assuming you could get a shorter belt.

Fab man
23 March 2013, 1226
Just measured the pulleys on the eAssist motor. The motor has a 3" pulley and the crankshaft has a 6.5" pulley. So at 150Nm crank that means the eAssist electric motor is putting out about 69.23Nm lets call it 70Nm. So 70Nm at 230Arms peak (assumption based). That would mean 167Nm (ideal) at the 550Arms of my Curtis. Don't plan to try to run it that high maybe even, but I will work my way up.

Also base speed is at about 1100RPM of the crank and peak is 3500RPM crank. So this means at 115V (@230Arms) the base speed is ~2380RPM and peak is at ~7580RPM as it is used in the eAssist. Theoretically at 350V this would be a base speed at ~7200RPM with the appropriate frequency.

Those RPMs would yield too high of a chain speed for practical use, so a belt drive(at least for the primary) is a better option. It means you'll have to set up a jack shaft, unless you can find a huge rear sprocket like the Zero production bikes, and figure out some way to keep near constant tension on the belt(s) .

As for as mounting a sprocket on the motor shaft, are you sure about the 17mm diameter? One of the lowest cost option, stock Taper-Lock bushings, come in 14, 16, 18, 19mm bores. In some applications, using the Taper-Locks, I've gotten away with not using a shaft key. Your motor shaft is on the small side(most motors this size have at least a 7/8", ~22mm, diameter shaft), so you'd have to confirm if you could do likewise.

Another option would be to machine registers and threaded holes in the existing pulley to mount a separate sprocket(belt or otherwise).

Yeaw
23 March 2013, 1325
As for as mounting a sprocket on the motor shaft, are you sure about the 17mm diameter?

I measured the shaft with calipers today, it is definitely 17mm.

I think using the belt is definitely an option, but if you look at the video on N&V's first post of this thread, the tension spring for the stock pulley tensioner is huge. We may also have problems finding short 7 rib belts, but it is definitely possible.

Ohm6
25 March 2013, 1542
I had a successful day at the local auto recycler. Picked up a used "alternator" for $80 + a $25 core charge, yeah book pricing :cool: . Less than 6,000 miles and they just cut the wires so I have a connector that I can simply re-pin, and the motor stayed sealed to the elements.

http://i187.photobucket.com/albums/x215/2002noma/eassist/5e1c5a18-5107-4e63-9a5a-5289c1e486c9_zpsc457f9e7.jpg

Nuts & Volts
25 March 2013, 1632
I had a successful day at the local auto recycler. Picked up a used "alternator" for $80 + a $25 core charge, yeah book pricing :cool: . Less than 6,000 miles and they just cut the wires so I have a connector that I can simply re-pin, and the motor stayed sealed to the elements.

http://i187.photobucket.com/albums/x215/2002noma/eassist/5e1c5a18-5107-4e63-9a5a-5289c1e486c9_zpsc457f9e7.jpg

You son of a b1tch!

Did they have anymore?

Ohm6
27 March 2013, 1615
Ö Know any good ways to test the resolver signals?...



I finally had a chance to open the motor up a little and found a few things worth mentioning. The HV cables going to this motor do not seem very large and it looks like it will be difficult to fit much larger ones through the back without removing some aluminum.

I checked resistance across all the pins and came up with the following pairs:
1-5 Resolver Phase 1 (35 Ohms)
2-6 Resolver Phase 2 (35 Ohms)
3-4 Resolver Excitation (20 Ohms)
7-8 Thermistor (65 kOhms @~40F, 50 kOhms @~58F) temperatures are best guesses
9-10 No pins
11-12 Safety Switch (3 kOhms with HV cover plate, open circuit without HV cover plate)

I donít have a way to excite the resolver or a scope to measure the outputs so it may have to be trial and error from here to find polarity.

4275

Yeaw
31 March 2013, 0730
Great leg work on the pins, that saved me some time.

I believe the pins are as follows:
Pin 5 (red): Cos+
Pin 1 (black): Cos-
Pin 2 (pink): Sin+
Pin 6 (purple): Sin-
Pin 3 (white): Exc+
Pin 4 (yellow): Exc-

I hooked up the pins like that to my EVAL board and it worked first try and output the position and speed to my computer using USB and the Analog Devices software.

magudaman
25 June 2013, 2328
Anyone get one of these motor up and running yet? I didn't realize they were so available through wreckers; I was looking for months trying to find one before seeing this thread. Time to start calling around.

Nuts & Volts
26 June 2013, 0600
Anyone get one of these motor up and running yet? I didn't realize they were so available through wreckers; I was looking for months trying to find one before seeing this thread. Time to start calling around.

Im getting there. Mine is mounted and chained to the rear wheel. I have a buddy working on a PCB layout for a resolver to encoder board so that I can run it with my Curtis 1238. Itll probably be about two more weeks before I get it running.

Im a bit angry that the motors are more and more available only because I paid $420 for mine. Im a bit bitter haha

yankee1919
26 June 2013, 0801
Nuts & Volts,

Thanks for the quick update. Can't wait to see your Ev in action.

Spaceweasel
26 June 2013, 0938
I just read through this thread - really interesting stuff, if a bit over my head.

Nuts & Volts
02 July 2013, 1300
I think I figured out what resolver is used on this motor. I was reading up about the Prius motors and noticed that someone had determined what resolver was used (http://forums.aeva.asn.au/forums/rx400h-transaxle_topic2435_post33440.html#33440) on that so I searched the catalog and sure enough I found one that matches the physical dimensions of the resolver I have.

Look at TS2224N1012E199 on page 7 of this PDF
http://www.tamagawa-seiki.com/pdf/download/1570N11EJ.pdf

vehiculeselectriques
25 March 2014, 1115
Hi,

no more news about this motor ?

Nuts & Volts
25 March 2014, 1824
Nope pretty much gave it up so that I could have a working bike using a zero motor and sevcon. I still have the motor sitting in my room. It may get used one day

Sent from my SCH-I605 using Tapatalk

vehiculeselectriques
26 March 2014, 0740
too bad,
it seems a very good candidate for E-motorbikes

Nuts & Volts
26 March 2014, 0811
too bad,
it seems a very good candidate for E-motorbikes

Yea, I'm going to chat with a buddy of mine this weekend about making a PM rotor for the motor. Would improve efficiency and could allow us to lower the commutation frequency requirements for higher RPM. This project is last on my list of things to do so I don't think it will really go anywhere any time soon.

ronbo
23 April 2014, 0054
UGH!... a dead thread.. what a shame! I found it a great & hopeful resource to utilize one of these... they're getting VERY reasonable now. I'm probably going to grab it anyway, and see if I can have them throw the LiIon pack in for a few dollars more, so they don't have to worry about "proper disposal" of it... ;-)

UltraCap
22 October 2014, 2111
Has anyone thought about buying the whole controller assembly and battery pack? They can be had for about $500 and are a lot cheaper than the other controllers discussed. Plus you get 0.5 kWh of "free" battery.

e-vektor
25 December 2016, 1329
I am starting an e-motorcycle project with this moto-alternator
I expected to have 150 N*m at 230 Arms but I only got 70 N*m. then I read this post and I discovered that it was measured at ICE engine
This will make me change the transmission using a sprocket of 80 or 90 teeth at wheel, but I think it will work

I am using Curtis 1236 80V 350A Controller and 75V battery only for the dyno. With this setup I get only 8 HP
Motor has to be modified from WYE to DELTA, otherwise voltages are VERY high.

For the motorcycle I am going to use 7 packs of 6S LIPO cells to get 155V (175V full load).
I think max power will be about 50HP.

For higher power voltages start to be very high

here a video of first dyno start:

https://www.youtube.com/watch?v=PES5hT-3nF0&feature=youtu.be

7240

here it can be seen clearly how the "field weakening" starts at 850 rpm (with 75V battery)

e-vektor
31 December 2016, 0259
a quick update: by now I recalibrated the dyno, now it shows +15% more torque, this makes that efficiency is now between 60% and 65% in most cases. And some of them (I cannot repeat) close to 70%, but no more. I had expected to have at least 80% in wide areas, but it is 60%.
I did lots of tests with different currents, different slip gains (mainly 9% and 15%), different speeds, and the best ones are always between 60% and 65%

I also did a quick max HP test and it gave 9.5 HP with 75V. If voltage was 115V (as in car) I should have 9.5 / 75 * 115 = 14.5 HP, which is very close to the official value (15 HP for traction)

Next step is reconnecting the motor, then 115V (star) is equivalent to 66V (delta), so I think I will be able to test it more propertly. Indeed in some efficiency maps I saw on the Internet, max efficiency only happens at higher speeds when high field weakening is being active.
In some way, probably I am trying to measure efficiency in a worst scenario, similar to have a 230 vac motor connected to a battery inverter. As the map is quite complex, probably I am reaching bad eff points at all corners

udaykishanr
16 June 2017, 0902
Hello Everyone,

I need your help in using Curtis AC Motor Controller 1236E series. Basic support is appreciated, like how to connect Curtis device and talk to the Induction Motor.

To start with, I am using BAS+ eAssist for one of the applications. I am interested to build a starter/generator for a Vehicle, now then, I am intended to control torque of this motor and have to make it operate at its peak value(Pullout Torque). I found Curtis AC Motor controllers interesting for this job. But I am totally unaware of how to use these controllers. I need help in using these Curtis Controllers to establish a communication between Inverter and Motor. At this point, I don't require any communication with CAN bus. Like I said, my interest is to maintain maximum torque value for some time period. Any input is appreciated.

Thanks in advance,
Uday

Matej
17 July 2017, 0001
Has anyone else made any progress as far as controlling this motor goes?

This motor would be perfect for motorcycle conversions once someone figures out how to control it with a readily available controller or builds an affordable controller for it.

ChazFisher
04 September 2017, 1708
I found this thread, and was intrigued. These motors are on eBay now, for $150 or so. If they could be made to work in a motorcycle application, that would be a great price point. But I think they'd be difficult to make work, at least without keeping the transmission. Here's why...

Looking at the specs I saw in various press releases, they were pretty consistent at 15hp, but torque ranged from 79 ft-lb to 110 ft-lb. It was pretty clear, however, that those were the numbers at the engine crankshaft. I expect to see the torque drop off pretty dramatically once the power peak is reached for an induction motor. (Look at the curves for, say, an AC-12 motor at 48 V. The torque fall-off is pretty dramatic.) If I assume that torque/power point represents the knee of the curve, it's reaching peak power at somewhere between 700 and 1000 RPM at the crankshaft. With the 6.5"/3" pulleys that Nuts & Volts cited, that's between 1550 and 2200 RPM at the motor/gen. The actual motor torque will be 36-51 ft-lbs - falling off very quickly over 2000 RPM. So gear it high enough to pull off the line, and it will run out of power at 20 mph. Need to keep the extra weight and complexity of the clutch and transmission to even make it worth trying. And at best, with only 15 hp, you've got an urban commuter bike.

This makes sense, given the "mission requirements" in the original application. This motor is intended to quickly restart the ICE from a stoplight, to regenerate a little power when coming to a stop, and to add just enough power to avoid the transmission having to "kick down" a gear when passing or going up a hill. So decent torque at low RPM is what's needed, and spinning up to higher speeds isn't important.

Theoretically, you could up the battery voltage, hold the max current constant, and push that "knee" to higher RPMs. If it can hold 50 ft-lbs to 4000 RPM, you've got a winner! But would GM have overrated the insulation system by so much that the motor can take the higher voltage? My guess is the dV/dt from switching at the higher voltage would break down the insulation pretty quickly. So I think you're stuck with the 115V of the original battery pack. If a top speed of 25-30 mph is good for your application - a city commuter bike - then this might work. My commute is mostly a 45 mph speed limit, actual traffic up to 60 mph. it won't work for me.

Warren
04 September 2017, 1958
e-vektor has made some progress.

https://endless-sphere.com/forums/viewtopic.php?f=30&t=85514&start=50#p1316758

Electro Flyers
04 September 2017, 2006
I think the insulation would be fine for the higher voltage. You could get the stator redipped at a motor shop or use some of the rattle can insulating varnishes to add some voltage capacity. I don't know much enough about induction motors. Can the torque be controlled by programing the timing and amount of asynchronous slip of the phases(if that's what's needed)? And what controllers are out there that would work with motor?

The motor is pretty small and I think GM stopped offering it as an option, but somebody should try the darn thing.

ChazFisher
09 September 2017, 0810
e-vektor has made some progress.

https://endless-sphere.com/forums/viewtopic.php?f=30&t=85514&start=50#p1316758

I've got to admit, that looks a lot better than I thought it would. Torque doesn't start to drop sharply till above 3000 rpm. I stand corrected!

Stevo
09 September 2017, 0859
I've been patiently waiting for an update on this thread:cool:
"Efficiency at the power peak was about 75% (induction motors are more efficient at parcial loads). With a magnets motor like ME1304, controller's delivered power could be the same, but they should achieve 90 or 92% efficiency, and thus power at shaft would be may be close to 45 HP, but I don't think it can be more at this voltage."
My thoughts exactly! It makes logical sense that a PMAC would be more efficient because a permanent magnet has a stored potential energy already "built-in" to the motor.
7419 < is looking good for next year's upgrade!

ChazFisher
09 September 2017, 1047
So after reading the whole thread on endless-sphere, I do have a couple more thoughts:
1. e-vektor did some excellent reverse engineering. That was much, much more than a simple, "Throw a likely controller at it and see what sticks." Bravo!
2. The conversion from wye (star) to delta wiring made a big difference. Thinking about its use as an alternator, the wye configuration makes sense. You'd want to make enough voltage to be able to charge the 115 V battery pack down to low rpms, as the car is slowing to a stop. Changing it to delta for motor use allowed him to push more current - and therefore make more torque - with a reasonable battery pack voltage.

I'm still pleasantly surprised that the torque curve drop-off is as high as it is.

e-vektor
04 December 2017, 1409
So after reading the whole thread on endless-sphere, I do have a couple more thoughts:
1. e-vektor did some excellent reverse engineering. That was much, much more than a simple, "Throw a likely controller at it and see what sticks." Bravo!
2. The conversion from wye (star) to delta wiring made a big difference. Thinking about its use as an alternator, the wye configuration makes sense. You'd want to make enough voltage to be able to charge the 115 V battery pack down to low rpms, as the car is slowing to a stop. Changing it to delta for motor use allowed him to push more current - and therefore make more torque - with a reasonable battery pack voltage.

I'm still pleasantly surprised that the torque curve drop-off is as high as it is.

here is the most I could get from this motor: 38 HP and 95 N*m on electric kart

https://www.youtube.com/watch?v=V-M_RFaY3aM

thunder
28 February 2021, 0951
Interesting thread, I am looking to put it on a 1988 KDX Kawasaki woods bike, got a few questions
should I use version1 or 2 of the motor I seem to believe there are major differences in the two both in voltage and rotor ( I believe ver1 rotor has PMs embedded in an aluminum core)
15hp would be equivalent to the old two-stroke engine could anyone recommend a voltage for the battery pack and controller options...

Thanks all