New H3548 random cutouts

[kokopelli314]

Crystalyte H3548 + Infineon 40A = Random Cutouts

Both purchased from Grin Technologies. Phase and sensor wires Y-Y G-B B-G as advised by Ben. Y-Y B-B G-G does not work.
This is the second 3548 I have had this problem with. Same behavior with Crystalyte 40A sensored controller.

No cutouts with Crystalyte 40A sensorless but that controller runs very hot with this motor, and cold with a 405 and a 9C
Hopefully someone can give some insight into this.

[youtube]YdKEPCfWWcU[/youtube]

Have tried various Phase / sensor combo

http://www.youtube.com/watch?v=YdKEPCfWWcU

 

[dogman dan]

Mystery to me. I'd suspect the controller a lot more than the motor.

I'm confused why you'd need to try altenative halls and phase combos if all came from grin. Opening the motor, grin would be able to see if stuff inside was wired funny compared to normal.

 

[kokopelli314]

Mystery to me. I'd suspect the controller a lot more than the motor.

I'm confused why you'd need to try altenative halls and phase combos if all came from grin. Opening the motor, grin would be able to see if stuff inside was wired funny compared to normal.

This happens with a crystalyte controller as well. Both controllers work fine with a 405. It was suggested to me that there may be a problem with the wiring, so i tried various combinations. What works is Y-Y B-G G-B

 

[mvly]

Yeah the crystalyte sensor less runs hot on my hs3548. I since then moved to lyen 18FET 4110 and never had a problem. The lyen controller would get warm but never hot.

 

[kokopelli314]

Yeah the crystalyte sensor less runs hot on my hs3548. I since then moved to lyen 18FET 4110 and never had a problem. The lyen controller would get warm but never hot.

That's what I thought too.

 

[dogman dan]

I didn't quite get it that the same thing happens with two different controllers.

I had the same thing eventually develop with two 20 amp sensorless controllers. I returned both to the seller so he could investigate, recieving one new controller in trade. The seller never did let me know what it was, nor would I have understood it. To me, as I rode it felt like a cutout circuit like the LVC gone wonky or something.

No chance it's your battery sagging?

Just curious about one thing, your name is the flute player, but I never saw gumby play the flute.

 

[kokopelli314]

Just curious about one thing, your name is the flute player, but I never saw gumby play the flute.

A bit of a juxtposition. My YT avatar is more akin to the Hopi Fluteplaying Dancer. I do make and play flutes though

I didn't quite get it that the same thing happens with two different controllers.

I had the same thing eventually develop with two 20 amp sensorless controllers. I returned both to the seller so he could investigate, recieving one new controller in trade. The seller never did let me know what it was, nor would I have understood it. To me, as I rode it felt like a cutout circuit like the LVC gone wonky or something.

No chance it's your battery sagging?

I hope you got a controller that worked, considering you sent two.

I'm using 5, 12AH sla's with a 12A charger hooked up. About 69.1V No sagging voltage

Yes, the cutouts are probably related to the way the controller is responding to the motor. It may be possible that with this 23 pole motor, the positioning of the sensors is super-critical and that under load, phase lag is causing misfiring in the controller, almost like it's trying to jump to the previous cycle. There may be a way to correct this by refining sensor angle settings in the controller. Not really sure without a scope though.

I've purchased an 18 FET Lyen sensorless and am waiting for it to arrive. In the meantime Justin is looking into this problem. Grin sent this motor as a replacement for the last H3548, that had exactly the same problem.

 

[dogman dan]

I figured two broken controllers for one new one was a great deal. Both old controllers had over a year's hard use.

 

[kokopelli314]

[youtube]vpMVai081LM[/youtube]
Unlike the $175.00 infineon based controller from ebikes.ca, this $35
controller from ebike-bmsbattery.com G-CO1560 runs the 3548 without
cutouts.
So far I've tested:
Clyte 405 + Crystalyte 40A sensored = pass all speeds and loads
H3548 + Crystalyte Sensored 40A = fail under load at speeds over 35kph
H3548 + Crystalyte Sensorless = pass, but runs extremely hot
H3548 + Infineon 40A sensored= fail under load at all speeds
Clyte 405 + Infineon 40A sensored = pass but runs a little hot
H3548 + G-CO1560 = pass at all speeds and loads

 

[dogman dan]

Weird that it runs the 405 fine, but I think you have some kind of evidence that it's not the motor, it's the controller.

405 works under load, but maybe the 405 can't develop the kind of load it takes to see the cutout? A total guess, but I'm thinking some kind of high amps shutdown.

 

[kokopelli314]

A total guess, but I'm thinking some kind of high amps shutdown.

Possible, but the cutouts can begin at ~ 4A with the Infineon. I was wondering if the Clyte and infineon firmware may need to be upgraded for these motors.
Anyway, Justin has offered a replace/refund, so at least on the sales side things are good.