The One True Alter-motor

[acuteaero]

I've been waiting for a chance at one of these for a while- finally got a lucky break and snapped one up for cheap. It's a "belt alternator/starter" from a GM Chevy Malibu Hybrid, also used on the Saturn Vue and Aura, hybrid models. These are "mild" hybrid systems, just do a couple Kw assist and regen, as well as cranking the engine over for start/stop, all through this alternator-looking thing driving the crankshaft through a high-tension 7 rib ribbed v-belt. The original system uses a 36v NiMh battery.

Sep-ex, brushed :?

Thin, ~.3 mm laminates

I was hoping to get the main case apart and get a good look at the rotor but I was unable to get the tail-shaft to go through the rear bearing and the pulley appears to be impossible to remove in one piece due to the use of a taper-lock bushing on the shaft... (Photo is after removal of usual style nut)

I'm thinking about getting this rigged up on a bike with some variable or at least 2 or 3 stepped field control and do some testing (maybe some racing also ) on it. First I need to work out how to get power out of it- currently I'm leaning towards removing the pulley with an angle grinder and machining some parts to replace it with a nice flange for flat-sprocket mounting. If anyone has any input about this I'd like to hear it very much.

Then need to pick a controller- I'm not particularly interested in going sensorless, I'm imagining with "hall style" commutation sensors a beefy Infineon ebike style controller could probably do the job, sevcon or kelly is a more expensive option, I guess, for higher current. Anyone have experience with the Austria Microsystems family of magnetic absolute encoders, particularly the AS5134, with the UVW style commutation outputs? Or how about the CUI AMT303 module? If this BAS alternator is the same as all the other alternators I've ever seen it should have 6 pole pairs.

 

[Hillhater]

Nice find !
This will be an interesting thread.
I suspect that pulley is locked onto a tapered shaft with the split bush expanding on the shaft to lock the pulley ??
Maybe a few raps on the pulley will shock it loose ?

 

[liveforphysics]

Excited to see this used on a bike! It was cool to see it in person. Thank you for bringing it by.

 

[acuteaero]

The plot thickens....... I got a "bigger hammer" and "convinced" the taper bushing to pop out- I'm glad I didn't have to destroy it to remove it! I don't think I'll want to use it, but if I decide to at least it still exists. Now, for the real business:

Woah, what's going on here? 8 pole pairs instead of 6, and again, what? That's little ferrite magnets snuck in between the claw fingers! Kind of like was discussed over here http://endless-sphere.com/forums/viewtopic.php?f=30&t=16328#p240248 I won't say I know exactly what's going on here, but the more differences I see in this than the usual alternator, the better- because I presume any changes they made are for the purpose of making it a more efficient and power capable motor/generator.

Comparing a standard Delco alternator to the BAS alternator, stator and rotor:

So, the upshot- knowing it's 8 pole pairs means I know the aformentioned austriamicrosystems chip won't work, but the CUI encoder might still be an option. Seeing the way the pulley mounted means I can design an adapter to mount a sprocket.

 

[fechter]

Interesting 8)

It looks like they are using a mini-alternator for the encoder. You may need to replace this with a hall sensor arrangement to use a standard controller. You might be able to simply remove the encoder stator and mount hall sensors in 3 locations and use the existing magnetic rotor.

The magnets wedged between the fingers is also very interesting. Presumably this will reduce the amount of wasted power in the field coil.

 

[Gordo]

Interesting 8)

It looks like they are using a mini-alternator for the encoder. You may need to replace this with a hall sensor arrangement to use a standard controller. You might be able to simply remove the encoder stator and mount hall sensors in 3 locations and use the existing magnetic rotor.

I am puzzled by the suggestion of 3 Halls. When you have 8 pole pairs, won't you be using 2 or 4 Halls?

 

[acuteaero]

Every 3 phase motor will use three hall sensors, corresponding to the three phases. The number of pole pairs determines how many cycles of output states the sensors go through per revolution, hence the number of commutation cycles per revolution. In this motor that'll be eight cycles.

I did a little more reading on the CUI encoder- it looks like a pretty neat little unit, unfortunately it can be set up for six pole pairs or ten, but not eight. So that option is off the table as well. I'm not really digging the idea of cutting little recesses in the middle of the stator for hall sensors as others have done. Unfortunately I don't think the original encoder can be modified, its rotor is simply a small stack of steel laminations cut in an oval outline- there are no magnets. I'm leaning toward building an optical sensor right now.

A little more investigation of the permanent magnets reveals that they are magnetized across the narrow face between the claw teeth- so that they magnetize the claw teeth alternating N/S- so like Fetcher speculated they're just sort of providing a (relatively) moderate magnetism to the claws with no field current- should cut the field current requirement a bit. Also means I'll have to take care to drive the field coil with the correct polarity to match the permanent magnets! I wonder if it was also conceived as a safety feature for rotor coil failure, which causes the back EMF to disappear- which could cause trouble for the attached inverter... although if the inverter is built right with good current limiting it shouldn't cause it to blow up.

 

[MitchJi]

Hi,

About how much does it weigh?

These are "mild" hybrid systems, just do a couple Kw assist and regen, as well as cranking the engine over for start/stop, all through this alternator-looking thing driving the crankshaft through a high-tension 7 rib ribbed v-belt.

First I need to work out how to get power out of it- currently I'm leaning towards removing the pulley with an angle grinder and machining some parts to replace it with a nice flange for flat-sprocket mounting. If anyone has any input about this I'd like to hear it very much.

I got a "bigger hammer" and "convinced" the taper bushing to pop out- I'm glad I didn't have to destroy it to remove it! I don't think I'll want to use it, but if I decide to at least it still exists.

I'd need to have some excellent reasons to replace a belt with a chain. If you don't want to use the ribbed v-belt why not use either 5mm htd or 8mm polychain?

 

[grindz145]

way more copper fill than the standard AC delco alternator! Definately interested in these. Did you find one at a junkyard? Might be worth it to see what a brand new one costs.

 

[MitchJi]

Hi,

Check the Gen II Specs:
http://en.wikipedia.org/wiki/BAS_Hybrid#Generation_II_BAS_.28eAssist.29

Generation II BAS (eAssist)

At the LA Auto Show, on November 15, 2010, General Motors announced that it would be releasing an all-new version of the BAS system available in the 2012 Buick LaCrosse.[8] While still a Belted Alternator Starter system, the system is named eAssist and includes a larger more powerful Hitachi-supplied 115 Volt Lithium Ion battery and a 15 kW (20 hp) motor-generator that delivers 79 lb-ft of torque. The new motor is apparently able to contribute more power in assisting the 2.4L engine and able to operate electric-only but only at very low speeds. The eAssist system includes a modified GM 6T40 6-speed automatic FWD transaxle

 

[acuteaero]

Indeed- if the new system runs 115V/15kw it will be a pretty impressive unit! The description doesn't exactly add up though, saying it'll still be belt driven, but that the car will be able to run at low speeds electric only- We'll see. Maybe there will be some nice bits at the dismantlers in a few years, at least.

The motor, with leads weighs a little more than 20 lbs. I got it from an auto dismantler, via ebay. The prices range between about 200 and 600 dollars. I got a real lucky break on this particular one. They're a little too new and rare to be cheap at this point. I can't imagine there's much demand, but that doesn't necessarily set the price a dismantler wants. Might be worth giving them a low-ball offer, though.

I'm not particularly impressed with the copper fill or end-turns- the windings look really similar to the standard alternator- as far as I can see. It does have many more stator teeth, though, and much thinner laminations. People seem to say good things about the Denso SR style, with square wire single file in super narrow slots- very much like the Remy HVH winding style

But this alternator is not one of those. I'd be interested in testing both- experience and quantative testing will go a long way toward establishing what the potential strengths and weaknesses of these alternator-style construction motors are. I saw this thread linked from one of the other altermotor threads http://www.rcgroups.com/forums/showthread.php?t=1123956&highlight=reem ... it's a really good read about some actual testing done on a SR style alternator. It really highlights some of the potential advantages of the seperate field excitation in these motors- if you get it dialed it it could be pretty advantageous- And no chinese RE reliance... 8)

As for the chain vs. belt - I think belts are way preferable in many situations, for a bunch of reasons- my last bike used a single stage belt drive and it was quite nice! For this motor I'm thinking I will want to put it on a barebones bike, modeled a bit off of Luke's current ride- the "death bike" as it were... for racing- robustness but without as much concern for noise, livability or long term longevity. For this application I think a single stage chain drive is the way to go. Looking at the dyno charts on the linked rcgroups thread it looks like there's is likely to be lots of very useful power available under 5k RPM, which should make a single stage chain doable. I think.

 

[jag]

http://en.wikipedia.org/wiki/BAS_Hybrid#Generation_II_BAS_.28eAssist.29

From same wikipedia page:

According to the EPA the 2009 Saturn Vue BAS hybrid garners an improvement of 32% city (19>25mpg) and 24% highway (26>32mpg) making the combined economy improvement 27% (22>28mpg) over the base 4cyl FWD version.[2] )

Obviously a hybrid will cut city mpg a lot (just stopping the ICE to avoid idling at readlights etc gives 10% mpg improvement according to a Scientific American survey).

What I'm puzzled by is how hybrid highway mpg can improve so much over the same ICE. I assumed hwy is mostly steady speed driving, where the hybrid system is mostly inactive. Or is the EPA cycle different? (Note a Prius uses a significantly different ICE than other Toyota's)

 

[acuteaero]

Vehicles with the BAS system use a conventional 4T45-E automatic transmission which has been modified to include a more efficient final-drive ratio

Again, from the same wikipedia page... That may have something to do with it. The 25% highway mileage improvement may be a figment of the test cycle too, and not nearly as dramatic in real life.

 

[salty9]

Interesting 8)

It looks like they are using a mini-alternator for the encoder. You may need to replace this with a hall sensor arrangement to use a standard controller. You might be able to simply remove the encoder stator and mount hall sensors in 3 locations and use the existing magnetic rotor.

The magnets wedged between the fingers is also very interesting. Presumably this will reduce the amount of wasted power in the field coil.

Could this "mini-alternator" be part of a selsyn (self-synchronous) and control the rpms of the altermotor? Described in http://openbookproject.net/electricCircuits/AC/AC_13.html

Edit: In back checking this URL I found http://openbookproject.net/electricCircuits/ which may be of interest to some here.

 

[acuteaero]

I'm pretty sure the positional sensor is a type of resolver- I haven't found really clear proof, but the whole coils around oval rotor thing fits:

(from ORNL 2005 "Report on Toyota Prius Motor Thermal Management")

 

[Gordo]

http://en.wikipedia.org/wiki/BAS_Hybrid#Generation_II_BAS_.28eAssist.29

From same wikipedia page:

According to the EPA the 2009 Saturn Vue BAS hybrid garners an improvement of 32% city (19>25mpg) and 24% highway (26>32mpg) making the combined economy improvement 27% (22>28mpg) over the base 4cyl FWD version.[2] )

Obviously a hybrid will cut city mpg a lot (just stopping the ICE to avoid idling at readlights etc gives 10% mpg improvement according to a Scientific American survey).

What I'm puzzled by is how hybrid highway mpg can improve so much over the same ICE. I assumed hwy is mostly steady speed driving, where the hybrid system is mostly inactive. Or is the EPA cycle different? (Note a Prius uses a significantly different ICE than other Toyota's)

From Wiki:
The "highway" program or Highway Fuel Economy Driving Schedule (HWFET) is defined in 40 C.F.R. 600 App I and uses a warmed-up engine and makes no stops, averaging 48 mph (77 km/h) with a top speed of 60 mph (97 km/h) over a 10 mile (16 km) distance. The measurements are then adjusted downward by 10% (city) and 22% (highway) to more accurately reflect real-world results."

 

[flathill]

I'm pretty sure the positional sensor is a type of resolver- I haven't found really clear proof, but the whole coils around oval rotor thing fits:

(from ORNL 2005 "Report on Toyota Prius Motor Thermal Management")

You are correct. The Honda IMA motor has one also. I've heard SEVCON's have sin/cos inputs as an special option but good luck getting it working. You'll probably be better of adding hall sensors. See if the back emf of the motor is trap or sinu

 

[salty9]

The site I posted earlier says a resolver is a type of selsyn and gives some explanation and math that I think i forgot.

http://openbookproject.net/electricCircuits/AC/AC_13.html#xtocid1540153

 

[Leigh Dleedlee]

I'm working on a project and I want to use a BAS to wholly replace the starter; I don't need the the power assist function, so it would just be starter be alternator. Does this motor have enough power for cold cranking and would I be able to use a standard controller?

 

[ShadowNightmares]

I'm working on a project and I want to use a BAS to wholly replace the starter; I don't need the the power assist function, so it would just be starter be alternator. Does this motor have enough power for cold cranking and would I be able to use a standard controller?

Mild hybrid DIY builds aren't that common, so i wonder how will you couple the alternator to the crankshaft without overrevving it as the motor fires up and works in thousands of RPM (normal starters can uncouple with the solenoid-actuated mechanism, AFAIK)

 

[Leigh Dleedlee]

Mild hybrid DIY builds aren't that common, so i wonder how will you couple the alternator to the crankshaft without overrevving it as the motor fires up and works in thousands of RPM (normal starters can uncouple with the solenoid-actuated mechanism, AFAIK)

I think you misunderstood, there's no reason to uncouple it from the crankshaft because I would be using it to function as starter and alternator.

 

[fechter]

It can work as both a starter and an alternator. The tricky part is the controller. For starting, it needs to resemble a bike motor controller. Once it fires up, it needs to switch to regen mode and charge the battery. Most existing setups are running much higher than 12v though. Typically there would be a HV pack and a dc-dc converter to charge the 12v battery.

Do you know what voltage you want to run the starter from?

 

[Leigh Dleedlee]

It can work as both a starter and an alternator. The tricky part is the controller. For starting, it needs to resemble a bike motor controller. Once it fires up, it needs to switch to regen mode and charge the battery. Most existing setups are running much higher than 12v though. Typically there would be a HV pack and a dc-dc converter to charge the 12v battery.

Do you know what voltage you want to run the starter from?

I was thinking of having a single 36V battery, running the alt/starter and ignition coils direct, then having a 12V step down for the electronics.

 

[stevon]

acuteaero,

Any updates on this alternator? What controller are you going with? I'm thinking of converting a water cooled Hybrid Hyundai Alternator, part number 373902G920