Go Back   EcoModder Forum > Off-Topic > The Lounge
Register Now
 Register Now
 


Reply  Post New Thread
 
Submit Tools LinkBack Thread Tools
Old 12-10-2020, 11:33 AM   #21 (permalink)
Somewhat crazed
 
Piotrsko's Avatar
 
Join Date: Sep 2013
Location: 1826 miles WSW of Normal
Posts: 4,061
Thanks: 467
Thanked 1,111 Times in 980 Posts
Didn't know it was 3 phase. 94 amps is a lot but they aren't on 100%, duty cycle should be about 16%. Does look like not enough heat sink, should get nice and toasty @ rated.

  Reply With Quote
Alt Today
Popular topics

Other popular topics in this forum...

   
Old 12-10-2020, 03:36 PM   #22 (permalink)
Master EcoModder
 
Stubby79's Avatar
 
Join Date: Sep 2012
Location: Victoria, BC
Posts: 1,747

Firefly EV - '98 Pontiac Firefly EV
90 day: 107.65 mpg (US)

Little Boy Blue - '05 Toyota Echo
90 day: 33.35 mpg (US)

BlueZ - '19 Nissan 370Z Sport
90 day: 17.19 mpg (US)
Thanks: 75
Thanked 576 Times in 426 Posts
How dare you make me think!

You have to have two of the 6 sets of mosfets conducting, or there would be no path to complete the circuit. There are 6 steps, during any one of them one set of mosfets stays ON the whole time, while another uses PWM to send current from the battery while controlling speed and/or limiting current. Soooo...during one step, one set is on, period, and a second set varies it's on-off state.

(I'm saying "set" since they're groups of 4 mosfets on this controller)



Here's what's going on whenever a pulse wave is sent:



But wait! There's more fun!!!!

Because...the phase current just keeps looping around...inductors and all. A third mosfet has the phase current flowing through it's "body diode" during that same step, whenever the battery current is NOT pulsed ON:



Sooooo...no matter what, there are 2 sets always with current flowing through them. 33%, not 16%.

All that assumes there's not a second phase still doing something. I thought I read somewhere it was 66%...no, wait, I think that's the current going through each phase wire. Back and forth, AC. Dang. I wasted the last hour trying to figure out why I had 66% in my head. Oh well. I needed the lesson anyway.

I have a niggling feeling I'm not taking phase overlap in to consideration. Maybe it doesn't. ARRRG.

Final answer: 33%.

(Just don't quote me on it. Where's Paul when you need him?)

Back to our regurgitated programming:


Here's the heat sink:



I half expected more - it's identical to the 3000w controllers, other than being longer to hold bigger mosfets - but if you've ever looked in to a normal sized controller, it's a huge improvement. They just have a 1/2" square bar for the mosfets bolted to it. This will pull heat away much better, spread it several times over...the real question is whether the case can dissipate it after.

Mosfet closeup:



So, yes, they're the 4568's...150v, 171A.

I was concerned about those leads, but look at them! They're huge! Well, compared to the scrawny ones you find on the little mosfets smaller controllers use.

Weak link here? Well...it's still heat. Either build up in the mosfets, or build-up in the huge(but still not big enough, if you ask me) solder traces carrying all the current.
Attached Thumbnails
Click image for larger version

Name:	Mosfet Phase Sequence.jpg
Views:	113
Size:	64.4 KB
ID:	29738   Click image for larger version

Name:	Battery current.jpg
Views:	111
Size:	40.7 KB
ID:	29739   Click image for larger version

Name:	Phase Current.jpg
Views:	109
Size:	36.4 KB
ID:	29740   Click image for larger version

Name:	IMG_4395.JPG
Views:	112
Size:	76.6 KB
ID:	29741   Click image for larger version

Name:	IMG_4396.jpg
Views:	101
Size:	19.8 KB
ID:	29742  


Last edited by Stubby79; 12-10-2020 at 03:44 PM..
  Reply With Quote
Old 12-11-2020, 09:28 AM   #23 (permalink)
Somewhat crazed
 
Piotrsko's Avatar
 
Join Date: Sep 2013
Location: 1826 miles WSW of Normal
Posts: 4,061
Thanks: 467
Thanked 1,111 Times in 980 Posts
Not only will I make you think, I will challenge your thoughts. Riposte.

Umm your flow is wrong. The back flow through the circuits is through the diode. I don't recall mosfets conducting in reverse bias. Same deal with IGBTs which is why you need 2 to do AC. So there is no voltage applied to the corresponding gate and the mosfet is not on. You said so yourself right after "but wait".

Last edited by Piotrsko; 12-11-2020 at 09:37 AM..
  Reply With Quote
Old 12-11-2020, 10:25 AM   #24 (permalink)
Master EcoModder
 
Stubby79's Avatar
 
Join Date: Sep 2012
Location: Victoria, BC
Posts: 1,747

Firefly EV - '98 Pontiac Firefly EV
90 day: 107.65 mpg (US)

Little Boy Blue - '05 Toyota Echo
90 day: 33.35 mpg (US)

BlueZ - '19 Nissan 370Z Sport
90 day: 17.19 mpg (US)
Thanks: 75
Thanked 576 Times in 426 Posts
Not my image. I just stole it for reference! (And noticed the same thing you did)

The body diode still needs to conduct, the mosfet - as a whole - still has current flowing through it.

And electrons flow the other way out of the battery, so the whole thing is inaccurate!

You still have two sets with current flowing through them. Just not necessarily through the gate/battery.
  Reply With Quote
Old 01-30-2021, 02:52 PM   #25 (permalink)
Master EcoModder
 
Stubby79's Avatar
 
Join Date: Sep 2012
Location: Victoria, BC
Posts: 1,747

Firefly EV - '98 Pontiac Firefly EV
90 day: 107.65 mpg (US)

Little Boy Blue - '05 Toyota Echo
90 day: 33.35 mpg (US)

BlueZ - '19 Nissan 370Z Sport
90 day: 17.19 mpg (US)
Thanks: 75
Thanked 576 Times in 426 Posts
New victim came in the mail yesterday:



It's a starter/generator "altermotor" from a hybrid. Rated at 10kw! @200-odd volts, so good for about 50 amps (at whatever voltage up to 200 that you can supply) continuous. Brushless DC. Water cooled. Weighs in at all of 27lbs. It's a couple of inches bigger in each dimensions than a standard, full-size alternator.

Found dirt cheap (for these numbers) on ebay. Once I found out that it was a BLDC motor, I had to have it. The little bit of cogging as I turned the pulley confirmed it as such.

Spun it with my drill, hooked up to a rectifier, and it put out 40v @1350 rpm, ~33 rpm/V. Lower than most of my reasonably high-powered DC motors. NICE!

No hall effect sensors, unfortunately:



But hey, I have ebike controllers that can run sensorless. Sooooo...time to hack together a bench test:



Results:



Best part:



It was only drawing 0.81 amps, even without sensors. That's as good as anything I've hooked up with sensors. Most of them, without sensors, end up pulling ~1 amp for every 10 volts thrown out. Gets quite wasteful when you're throwing 80 or 90 volts at them. I expected it to be efficient, but not without sensors.

This is nice. I can put it to work without messing around with hooking up hall sensors for now. Don't have to worry about over-revving it (might have to worry about the controller, but not the motor). Totally thrilled with it, so I ordered another!

Ok, back to your regularly misguided programming.
Attached Thumbnails
Click image for larger version

Name:	IMG_4640.JPG
Views:	104
Size:	114.5 KB
ID:	30086   Click image for larger version

Name:	IMG_4648.JPG
Views:	100
Size:	126.3 KB
ID:	30087   Click image for larger version

Name:	IMG_4641.JPG
Views:	94
Size:	132.6 KB
ID:	30088   Click image for larger version

Name:	IMG_4647.jpg
Views:	89
Size:	37.7 KB
ID:	30089  
  Reply With Quote
Old 01-30-2021, 03:43 PM   #26 (permalink)
Master EcoModder
 
freebeard's Avatar
 
Join Date: Aug 2012
Location: northwest of normal
Posts: 27,665
Thanks: 7,767
Thanked 8,575 Times in 7,061 Posts
Nice. Do you know what the donor vehicle was?

The GM eAssist supposedly has 15kW rating with no water cooling.

(I haven't forgotten your $20CDN, I just haven't done it)
__________________
.
.
Without freedom of speech we wouldn't know who all the idiots are. -- anonymous poster

____________________
.
.
"We're deeply sorry." -- Pfizer
  Reply With Quote
Old 01-31-2021, 12:03 AM   #27 (permalink)
Master EcoModder
 
Stubby79's Avatar
 
Join Date: Sep 2012
Location: Victoria, BC
Posts: 1,747

Firefly EV - '98 Pontiac Firefly EV
90 day: 107.65 mpg (US)

Little Boy Blue - '05 Toyota Echo
90 day: 33.35 mpg (US)

BlueZ - '19 Nissan 370Z Sport
90 day: 17.19 mpg (US)
Thanks: 75
Thanked 576 Times in 426 Posts
15hp, not 15kw, and they're liquid cooled. Not sure if they're brushless DC or if they're ac induction motors, thanks to this guy calling them AC:



Anyway, I have looked often on ebay, and never found an affordable one going to Canada. When I did see them, they were at least 3x the price of this one.

It's from an '11-16 Hyundai Sonata.
  Reply With Quote
Old 01-31-2021, 12:57 AM   #28 (permalink)
Master EcoModder
 
freebeard's Avatar
 
Join Date: Aug 2012
Location: northwest of normal
Posts: 27,665
Thanks: 7,767
Thanked 8,575 Times in 7,061 Posts
Is it an Hitachi part?

Quote:
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[citation needed] 115 Volt Lithium Ion battery and a 15 kW (20 hp) motor-generator that delivers 79 lb·ft (107 N·m) of torque.
ecomodder.com: Controller mods or build for E-assist altermotor
__________________
.
.
Without freedom of speech we wouldn't know who all the idiots are. -- anonymous poster

____________________
.
.
"We're deeply sorry." -- Pfizer
  Reply With Quote
Old 01-31-2021, 03:20 PM   #29 (permalink)
Master EcoModder
 
Stubby79's Avatar
 
Join Date: Sep 2012
Location: Victoria, BC
Posts: 1,747

Firefly EV - '98 Pontiac Firefly EV
90 day: 107.65 mpg (US)

Little Boy Blue - '05 Toyota Echo
90 day: 33.35 mpg (US)

BlueZ - '19 Nissan 370Z Sport
90 day: 17.19 mpg (US)
Thanks: 75
Thanked 576 Times in 426 Posts
Not hitachi, unless the guts are. Which could well be. Has Hyundai embossed on it. Not the same though, as this appears to have finer windings compared to what they look like in the photos of the E-assist.

For the record:

It's body is 6" in diameter, and 6" long; mounts and pulley stick out from that, of course.

The battery voltage of the hyundai hybrids is actually 270v.

@56VDC, I was getting 1500rpm on the nose. So more like 27 rpm/v as a motor. Based on that and the original battery voltage, I worked it out to a bit over 12 ft/lbs @ 50 amps. I couldn't stop the pulley turning by hand at 12amps in to the controller(I'll have to try measuring the phase current under the same conditions).

Ok, so I went and checked. The current going through one phase wire was a bit above battery current. Around 10% over. Interesting...however the three of them work out with three of them compared to the input. Put on a better glove for the job and it went off the 20-amp scale of my multimeter trying to stop it, which I did not manage to.

Took a pic of how fine the windings are, while I was at it:



Which makes sense, considering the rated battery voltage.

It came with the plug with the wires cut off. They appear to be around 10 gauge, which was my initial indicator of it's current capacity. 10 gauge translates to 30 amps, but since there's three and current is always passing through 2 at any one moment, it would be *3/2 = 45 amps, which corresponds to 10kw @ 270v with expected losses.

Hmm.

Well, the fine windings explain why the ebike controller is happy to run it. You'll find similar thickness wires feeding my 3kw hub motor, which runs at ~40 amps continuous for rated power. More turns per winding = more torque per amp, less RPM per volt, again pointing to the high input voltage of the hybrid system.

More turns might explain why it's easier for the ebike controller to decide the position of the motor without using sensors. The two other motors I tested sensorless with had fewer turns...one being an actual 12V alternator, the other being a motor meant for an electric skateboard. They both pulled exponentially more amps spinning unloaded the higher the rpm, and would not reach full rpm, and the alternator would even spaz out past a certain rpm; the skateboard motor ran off little current and reached full rpm once I hooked up the sensors. I also have a large BLDC motor with super thick windings, which the ebike controllers would not touch...way too little resistance, took too much current to get it to start, so the controller would shut down to protect itself first. So those fine windings are probably key to my current "success" with it.

The E-assist has large conductors in it's windings, and appears to makes up for it with considerably more stators(seperate windgins) and poles, at least looking at it in the youtube video breakdown. Thicker windings, fewer turns, less torque, higher mechanical rpm...more stators, more torque, lower mechanical rpm. Results? Able to take a stupid amount of current, put out a lot of torque (since each winding is only trying to push/pull the rotor a much smaller distance), at the cost of running at a much higher electrical rpm - more switching - and having to use more robust transistors to handle the current that would be drawn by the much lower winding resistance. (Same issue I have trying to drive the aforementioned motor that the ebike controllers won't touch)

So, where does that leave us? E-assist for ultimate power, Hyundai for ease of controller/component cost? Something like that.

I can run this $63(US, plus shipping) motor with a $50 ebike controller (that runs in 10s of amps), rather than spend hundreds on the motor(no $100 units that will ship to Canada, or I'd have one already!) and hundreds more on a controller(that would need to run in 100s of amps, just to handle starting it...correct me if I'm wrong!).

Of course, there are drawbacks...if you were to do your own mild hybrid with the motor connected to your engine via the pulley, the engine revving up would raise the voltage going back in to the controller to the point of quickly frying an ebike controller. I'm still trying to figure out how my insight doesn't push a higher voltage in to the battery than it can handle...if it's 144v at 3000rpm, it ought to be heading towards 288v when you hit the rev limiter. Maybe I'm looking at it wrong and the IGBT can actually stop it flowing back? I'd think the body diodes would just make it act as a rectifier and as soon as the motor votlage is higher than the battery voltage, it would start pushing it back in to the battery (charging) until it overloaded it. BOOM goes the dynamite-shaped battery sticks!

Huh. Err. Gonna have to test that theory. Maybe all you need are mosfets that can handle the motor's back emf, not a whole system that runs higher. Dah.

So, yeah...you want a motor for something that isn't going to over-rev, like an all electric small vehicle (motorcycle, whatever else), no problem. Or forget regeneration and throw in a one-way clutch. Or otherwise have it connect directly to a wheel, where it will only over-rev if you exceed the speed you designed it to max out at. VHATEVA YA VANT!

It's a compact, cheap, easy to run motor for it's price, with potential to make lots of power, if run right...just what I was looking for, for some of my projects.

Edit: Just tested it, by driving the motor with my drill and checking the coltage at the controller's battery wires(not hooked up to a battery)...yup! 42 volts. Good to know. Maybe IGBT's work differently? Maybe it's different when it's powered up and can turn them on and off? (doubt it. Ever seen an e-bike rev up past its max (battery) rpm going down a hill? Me neither...you hit a wall!). HmmMMMmmmm...I suppose they might just disconnect the battery via the contactor when the voltage gets too high. Dunno. DUH-NO.
Attached Thumbnails
Click image for larger version

Name:	IMG_4657.JPG
Views:	64
Size:	94.1 KB
ID:	30090  

Last edited by Stubby79; 01-31-2021 at 03:42 PM..
  Reply With Quote
Old 01-31-2021, 03:49 PM   #30 (permalink)
Master EcoModder
 
freebeard's Avatar
 
Join Date: Aug 2012
Location: northwest of normal
Posts: 27,665
Thanks: 7,767
Thanked 8,575 Times in 7,061 Posts
Apparently, it's all over my head. I used to follow the Open Revolt controller thread, and I have a 68hp Lexus axle that wants 400V or something.

Maybe you'd find some answers there: ecomodder.com/wiki/Open_ReVolt

No doubt it lacks a through-shaft for the cooling fan, so it's not what's needed for a Beetle engine?

__________________
.
.
Without freedom of speech we wouldn't know who all the idiots are. -- anonymous poster

____________________
.
.
"We're deeply sorry." -- Pfizer
  Reply With Quote
Reply  Post New Thread






Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.
Content Relevant URLs by vBSEO 3.5.2
All content copyright EcoModder.com