PDA

View Full Version : Kelly KLS8080 Auto Identify Help



ldopa_chicago
23 March 2017, 0733
Hey all, I have a 96V, 300A Kelly KLS8080IPS that I'm trying to get running. I'm in contact with Kelly but so far they haven't been able to resolve my issue.

As far as I know (and according to Kelly) all my motor and sensor parameters are set correctly. I write "170" to the identification angle field, wait, power cycle the controller, and then it begins its routine. It appears to simply be stepping through a full 6 commutations. After that, you're supposed to get a "reset error" in the monitor window, but this never comes. I've waited a long time and it never goes to "85" nor throws any error or anything.

Has anyone experienced this and resolved it? Is there a way to manually configure the identification angle? Is there CLI that I can access and get more information as to what's going on?

Thanks.

cmcnall
23 March 2017, 0757
After Reset, If you turn the throttle does the motor spin?
If that's the case then i wouldn't sweat it unless you find that your efficiency is terrible.

Also if you are using the app try using a laptop or vice versa?
I have found tablets to be more communication stable then using an android phone

good luck.

ldopa_chicago
23 March 2017, 0856
After Reset, If you turn the throttle does the motor spin?
If that's the case then i wouldn't sweat it unless you find that your efficiency is terrible.

It will twitch and/or whine (I watch the current to make sure it's not doing anything bad), but not spin. I was able to get it to spin if I turn the field drive off (rotor un-magnetized), but I think in that case it's sort of behaving like an induction motor. If I apply any load whatsoever it stops, naturally.



Also if you are using the app try using a laptop or vice versa?
I have found tablets to be more communication stable then using an android phone

good luck.
I'm using a laptop. Though sometimes it has trouble connecting initially, I do see throttle input reads and stuff like that so I know I'm talking to the controller.

Thanks.

__Tango
23 March 2017, 1408
It's been a while since I setup my KLS-8080I on my bike, so unfortunately, I don't remember much of the specifics now.

however, when I did it, I made a video. I'm not sure if it'll help, but maybe there will be something.


https://www.youtube.com/watch?v=pnVC-RDTIyM

Good Luck.

ldopa_chicago
23 March 2017, 1425
Ha, thanks, I've watched your video twice now looking for clues/differences. One is obviously the I vs IPS distinction but beyond that I'm doing what you're doing and my motor behaves just like yours during the "identify" routine.

Thanks, your videos are helpful.

__Tango
23 March 2017, 1448
Ha, thanks, I've watched your video twice now looking for clues/differences. One is obviously the I vs IPS distinction but beyond that I'm doing what you're doing and my motor behaves just like yours during the "identify" routine.

Thanks, your videos are helpful.

Bummer. It was a long shot.

From what you mention though, I don't think I ever received a "reset error" in the window either. I just reset the controller and re-read the config.

Have you tried swapping the phase wires and re-trying the autoconfig? (I'm absolutely grasping at straws here, but since the magic doesn't seem to be working, this is something i'd try).

ldopa_chicago
23 March 2017, 1458
No, I haven't swapped phase wires yet, but that's definitely worth trying. Theoretically it shouldn't matter but...I have nothing to lose!

__Tango
23 March 2017, 1714
yes, exactly.

maybe also make sure the hall sensor wires are correct and/or swap them up too.

markcycle
23 March 2017, 1909
Try raising the CPU power to 24 volts I had trouble at 12 volts very similar to what you're getting

__Tango
23 March 2017, 2316
Try raising the CPU power to 24 volts I had trouble at 12 volts very similar to what you're getting

Right! I run mine at 24V on @markcycle's recommendation (well from running a my original Kelly KBL controller).

ldopa_chicago
24 March 2017, 0541
yes, exactly.

maybe also make sure the hall sensor wires are correct and/or swap them up too.

My motor has a sin/cosine sensor (the difference between the I and IPS model controllers). I've tried swapping the position of the sine and cosine sensors to no avail. There are so many permutations of variables...

24V, huh? That's too bad because I have a 12V dc/dc on the bike, but I have another isolated bench supply that will do 24V so I'll definitely try that too.

ldopa_chicago
25 March 2017, 1047
A little update. I tried it at 24V, also no effect.

I realized that in the monitor window, the Hall A/ Hall B fields (which correspond to sine and cosine on my controller) were not changing with rotation. I initially assumed that because those were the "hall" fields that maybe they didn't apply to my version of the controller but I asked Fany and apparently I should see something there. Connecting either to 5V yields nothing which it should, so it's possible that my controller is defective or has an issue. I haven't connected anything stupidly to the 5V or those inputs so I'm not sure why/how this happened. I guess the next step is initiating service or return with Kelly. We'll see how that goes.

ldopa_chicago
28 March 2017, 0803
So Kelly has asked me to purchase and try the sine/cosine sensor that they use in the Motenergy ME1114/1115. Though they don't state anywhere that one must use this specific sensor, I don't think they've tested with anything else. So I have some of those on the way. I'll need to redesign how my sensor is mounted to accommodate the RM29 but it's doable, just a pain in the ass.

In the meantime I've asked them to in turn buy my sensor and test with that. We'll see if they reciprocate.

I wish there was more information relayed from the GUI and also that it was possible to do a manual calibration instead of the auto detect. Automation is nice until it doesn't work! It's frustrating because I could easily do the math and calibrate it myself if I could get more feedback from the controller.

ldopa_chicago
29 March 2017, 0923
I may have just figured it out. The sensor that Kelly did development with is a 1 electrical cycle/rotation part. My ADI part is 2 cycles/rotation. This would definitely mess up the auto cal. I believe this means that I can just lie to the controller about the number of poles- in my case telling it there are 24 instead of 48. This would have the unintended side effect of reported RPMs being off by a factor of two as well, so I need to make sure my RPM limit is set correctly.

Thankfully I have nothing going on tonight so I'm going to try this as soon as I get home. Fingers crossed...

ldopa_chicago
30 March 2017, 0735
No dice. Kelly's support continues to be unable to resolve the issue. I think this controller would be a fine choice for someone using off the shelf parts, but maybe not for me.

I'm strongly considering buying a Sevcon Gen4 instead. I know people complain about all the programming and excess of options, but that's kind of what I want/need!

Anyone with a Motenergy 1115 or similar want to buy a 96V 500A Kelly? :D

ldopa_chicago
03 April 2017, 0521
Ok, resolved!

It was partially my foolishness and partially the issue with the electrical differences between the sensor I'm using and the sensor Kelly tested with. Problem #1: despite know better, I was entering the number of stator coils as motor poles. My motor has 48 stator coils and 16 motor poles. Problem #2: the controller expects 1 cycle of a sine wave per rotation. My sensor puts out 1 cycle per 180 degrees. Solution: tell the controller there are half as many poles as there are. In this case, 8 poles. Ironically, this is the number that the controller defaults. It's now running and though I may tweak some parameters once I ride it, it's behaving very well.

Thanks to everyone who chimed in!

SplinterOz
05 April 2017, 0425
Glad you came back with the fix as it will help others.

Bernel
05 April 2017, 1426
Well done and useful information. I will be interested in how your Kelly performs.

Sent from my SM-T310 using Tapatalk

andy-meyer
26 December 2017, 0530
Hi,

I had the same Problem with an ME1302 motor and the same truble ti fix it. After sending the controller to Kelly, they had check them und the controller was OK.
I had order an new ME1302 motor and than it works. In the new ME1302 motors is not an RLS Sensor build in. They have an new new Xingli sensor.

How is youre Power at Startup like an weelie? I have now the follow Problem on my controller (700A):
On Startup (0rpm), the power is max 50A
on more than 60~100 rpm the power is max 200A
on more than 100 rpm the power is max 500A

The battery is not the problem. How are you hyppe with the power?
Regards, Andy