Kelly KLS 7218-M Identification not working

JHA

10 µW
Joined
Jan 12, 2019
Messages
5
Hi All,

I’m a long time stalker of endless sphere and the great info available here, but this is my first post. I quite enjoy the electric vehicle thing and I have built three e-bikes (really just plug and play), I have an electric skateboard which is also occasionally used to get me to and from work. Occasionally I don’t fall off too which is nice.

My most recent project is also a bit of a step into the unknown for me as it’s a bit bigger and less plug and play than previous tinkerings. I’ve taken an old quad bike with a broken gear box and have almost finished converting it. This has been a lot more involved with modification to the chassis, making frames for the motor, battery and also including a new shaft to provide gear reduction via chain drive to get a bit more torque to the ground. I’ll post some pictures and write up when complete, it will have:
- Golden Motor 3kW HPM3000b
- Kelly Controller KLS 7218-M controller
- 20S-10P 34Ah battery from EM3EV (I didn’t build it myself but did make the box for it)
- Cycle Analyst for info and throttle control / limit
- 72V to 13.8v DC-DC converter for the quad bike lights, horn and to power the controller (12V source not the B+/B- which is from the battery at 72V nom).

Right now, I’ve encountered a bit of a problem. I can’t get the Kelly Controller to perform the auto identify function. I’ve had a good read through the posts here and I seem to be having different issues to others. When I perform the process there is no movement of the motor at all and after a while I just get an “internal error” and “identification error” message on the interface screen.

I have gone through the user programmable parameters and have reviewed the items that can be seen under the “monitor” window. It seems that all is working. I can see the hall sensors from the motor change when I rotate the motor. I can see the RPM change when I move the motor by hand. One of the posts shows the order the hall signals should be arranged and I seem to match this, so pretty sure I’ve wired it up right:
https://endless-sphere.com/forums/viewtopic.php?f=28&t=95368#p1398898

However, I have had no success when I have tried to run the identification.

When I start up the controller, I get.a repeating beep which I take to be 1-1 on the error code (I.e. perform identification). I enter 170 into the program and write it to the controller and turn it off, wait for a minute or so and re-start. On the re-start the beeping ceases and nothing else happens. The motor does not turn (at all) and remains completely still & silent. Most posts indicate some (erratic) movement of the motor as the controller tries to identify all the phases and hall arrangements. Nothing from mine at all.

After a short or a long time I get two error messages on the monitor screen, “internal error” and “identity error”. Once I saw a “reset error” but it was accompanied by the other two error messages. In all cases the motor did not move (some of the forums indicate there are a few movements of the motor.

At this point I have tried to either restart the controller (leaving 170 In the identity tab) and I have also tried writing 85 in the tab and then restarting the process. In both cases I just end up with the same result, that is the controller not doing anything.
When I write 85 in and restart, the error beeps start up again in the same repeating “beep, pause, beep” fashion as when I first started which I take to be the “1-1” error code to perform indentification.

I have been through all the wiring and measured voltages and all appears to be hooked up correctly and verified by the monitor tab. It also seems as if the program has identified the phase order for each of the hall signals and yet no change to the error codes and I cannot drive the motor at all.

It is as if there is no voltage provided from the controller at all.

I am running a 20S 10P Lipo battery (nominally 72V, charged to about 81-82 volts at the time of testing.
I have a cycle analyst to monitor with the throttle controlled via the CA. I have tested the output here from about 1v to 3.8v input to the controller.

I have seen one post that suggest putting diodes between the motor hall signal and the controller which I have not tried (yet) but as I can read the hall signals and RPM in the monitor tab on the controller software I believe the controller can identify the hall signals OK.

Any help appreciated!

(PS - I will try to upload some photos of the program GUI and layout, please excuse the wiring in the photos, it’s all a big mess as I’ve been trying to trace everything back to make sure it’s connected correctly).940F2DE2-9EF3-4BD3-B735-755503043129.jpegView attachment 6
 

Attachments

  • A75D567E-FE5D-4D0E-9FD5-177AD89F5A63.jpeg
    A75D567E-FE5D-4D0E-9FD5-177AD89F5A63.jpeg
    159.1 KB · Views: 4,009
  • F2D78C76-A078-4593-9E2F-0E920D5C9A4C.jpeg
    F2D78C76-A078-4593-9E2F-0E920D5C9A4C.jpeg
    98.6 KB · Views: 4,009
  • 809332F4-64BD-4BA6-AE18-E731EE119909.jpeg
    809332F4-64BD-4BA6-AE18-E731EE119909.jpeg
    113.4 KB · Views: 4,009
I think your controller has an issue. If you look at the status tab it says Controller temp 246c!
It's also reporting ''internal error'' on the status screen.


I suggest you contact Kelly support for warranty repair.
 
246C temp reading in the controller would cause controller to cut the power and create overtemp error from the controller at the same time.

I think you can do this for the basic fault analysis:
-Please disconnect everything but halls, phases and throttle to limit error possibilities
- Can you see throttle percent go up in program when you turn throttle?
- You have a potentiometer throttle set in the program. Do you have this? Probably not an error but anyhow..
- Never seen auto id error.. it wasn't in the program before. If you see halls then these should be OK. Are all phase connections (to motor) solidly connected? if any of the connections would be broken then i guess this error might be shown.

Apart from this i have no ideas. maybe you damaged something when you connected CA?
 
larsb said:
246C in the controller would be cutting the power but you should get overtemp error from the controller at the same time..

I don't think so, because it's not reading the thermal sensor in the first place.
My guess would be that the thermal sensor got disconnected somehow, ie. broken wire.

Whatever the case it's not gonna solve itself by checking the phase connections.
I wouldn't even open the Controller if it's still covered by warranty.
 
Ah, i didn't mean to touch the internals. What i mean is the connections to motor from controller. If there is a mistake there which causes a broken connection then it would probably show on the controller through the auto id error code. Motor could also be having a short circuit internally causing error. It's easy to check continuity on the phase-phase connections.

These are unlikely errors, i agree on that 8)
But could be worth checking as a final resort as the warranty replacement time and shipping cost is a hassle.

And it's reading the temp since it's got a reading :D
The limit for overtemp in the controller is 100 deg so if sensor readout is that high then overtemp should be reported as an error - but it doesn't.
 
Thanks for the responses.

In monitor mode during the identify operation the temperature reads 246. When not performing the identify operation this seems to read correctly with temperatures around 22degC which is about right for my garage. Motor is also cool to touch and has not had any movement or sound so I don't think it has received any current. In the manual there is the following text so I don't think this is an issue:
Note: Thermistor is optional. Default to KTY84-130/150 or KTY83-122 When controller temp item shows
246 in monitor screen of user program,it is normal.This item is only useful for internal calculation for
MCU.246 is about 11 degrees in environment.
I have assumed a KTY84-130/150 thermistor based on what I have seen from Golden Motor but not sure if this is right. Seems to work ok.

- Can you see throttle percent go up in program when you turn throttle?
YES, although this is not touched or used for the auto identify option.

- You have a potentiometer throttle set in the program. Do you have this? Probably not an error but anyhow.
YES, both in the cycle analyst and in the Kelly program. It works as expected on the monitor page although there is not much travel in the thumb control so this could be a bit tricky if I ever get it going.

- Never seen auto id error.. it wasn't in the program before. If you see halls then these should be OK. Are all phase connections (to motor) solidly connected? if any of the connections would be broken then i guess this error might be shown.
YES, connections are tight with no play at the terminal. I can't see the connection inside the motor but assuming these are good as I've not touched anything inside the motor.
 
I have a document from golden motor stating 5 pole pairs for the 3kW motor and 4 for the larger 5kW motor. I understand this is only used to calculate the speed and not used for the identity. I will try with 4 pole pairs (8) to see if that works.
I would have expected some rotation of the motor regardless of this setting, but got nothing?
 
The pole setting doesn't affect this. I think your controller most likely has issues but would still recommend to disconnect everything not needed and try again and also measure the continuity from phase connections through motor for all the phase combinations and check the connection points on controller.

I've had a controller where sealant from production was smeared on the connections so that there was minimal conduction - how unlikely is that? :D
 
larsb said:
I've had a controller where sealant from production was smeared on the connections so that there was minimal conduction - how unlikely is that? :D
Fairly, but not unheard of. When I worked at Honeywell CFSG decades ago as a final test tech for flight instruments & computers, I ran across quite a number of PCBs with connectors that had conformal coat on main connector contact surfaces. Caused all sorts of wierd problems, depeending on what signals/etc the pins were for. After the first couple, I just started out checking every single contact on every board for such contamination, even if it didn't exhibit a problem in testing, and found plenty. After that, the assembly line implemented a full inspection before they reached me, and this cleared up that problem, because it found the sloppy assemblers causing the problem (usually by failing to maxk off the connector prior to auto-spray coating, but sometimes drips/smears during rework of failed components, requiring recoating of that PCB area. Once the problem was fixed, the only time we saw this was when stuff came back from the field where it had been worked on by airline mechanics/etc.
 
One other difference I have is the controller pin 7 is powered by 12v (13.8v) from dc to dc converter, not the 72v directly from the battery. The B+ is 72v from the battery. Could this cause the issue above?
Markup of controller diagram below:
 

Attachments

  • 43E6E808-F444-4728-8806-2A49210EDECD.jpeg
    43E6E808-F444-4728-8806-2A49210EDECD.jpeg
    85.8 KB · Views: 3,963
So, it turns out I’m a numpty, the controller needs full voltage to pin 7 not 12v (13.8v) as I had supplied. I read the 12v signal supplies from bike switches and assumed all pins on this plug would be 12v.

Fany from Kelly Controls put sorted it for me and I must compliment him on the rapid and useful help. Unexpected but appreciated. Be great if all suppliers responded like Fany.

Thanks all for your help. Will post some picks of the settings later. This is a great resource and some valuable learnings throughout.
 
HELLO SIR
THIS IS SRUJAN CAN YOU PLEASE HELP ME HOW TO USE THAT KELLY KLS M SOFTWARE
 
I dunno that someone so incredibly impatient (five posts in 39 minutes screaming in all caps, starting minutes after joining) can be helped, even if he did actually have a question that someone could answer.

But I will leave what advice I can anyway, regarding the only information he's given us to work with: To find out how to use Kelly's software, the best people to contact are the people that made it: The Kelly controller company itself. Other than that perhaps look at the Kelly configuration program documentation thread by Glyn, somewhere in this list:
https://endless-sphere.com/forums/search.php?keywords=kelly&terms=all&author=&sc=1&sf=titleonly&sr=topics&sk=t&sd=d&st=0&ch=300&t=0&submit=Search

SRUJAN said:
HELLO SIR
THIS IS SRUJAN CAN YOU PLEASE HELP ME HOW TO USE THAT KELLY KLS M SOFTWARE

SRUJAN said:
REPLY ME SIR

SRUJAN said:
SEEKING FOR YOUR REPLY SIR

SRUJAN said:
SRUJAN said:
SEEKING FOR YOUR REPLY SIR

SRUJAN said:
IS THERE ANY ONE TO REPLY ME SIR
 
Hello to the group.

I have the same problem with my Kelly KLS-H controller.

He does NOT go back from the 170 to the 85 and also is giving me a "identity Err".

I uploaded a video.
Can anyone help to point the right path??

Video link on youtube:
https://www.youtube.com/watch?v=pAZBuSllmQY

Thanks a lot !!!
Cheers,
Marcel
 
Hi Marcel, fellow Dutch man? :eek:

I have had simular problems with getting a Kelly KLS through identification proces..
It needs to jumb back to 85 by itselve, and with me, it also stayed at 170.
It took me a few nights, to find this solution, which worked for and many others.. But not everybody.
But you can try:

To 'feed' the hallwires, you are supposed to use pin5 purple, 5volts.
The gamechanger for me and many others with a slightly different motor was, to use pin11 red, 12volts, instead. Met dezelfde zwarte draad, de hele printplaat is onderling geaard.

So I use pin11 12volt, with same black wire.

I heard that in newer Kelly instruction books it is also advised by Kelly..

It went through identification angle within 1 minute, the green lights have to turn red, then it must have jumbed back to 85, then ignition off-on, kelly light is green and you are ready to go.

Best you can do to rule out everything else, is just connecting the wires that are absolutely nessesary..

-Connect the wire you made from B+ with pin7 Pink Pwr
-Connect the thin hallwires and feed them with pin11(12v)
-Connect thick fase wires
-Connect Throttle wires
-Connect power faser wires red and black

That's all needed to get through identification..
Heel veel succes, lukt je wel!

Groet Bas
 
Hi I am using Kelly KLS14301-8080I.

I was also getting an auto-identification error.
My solution was to replace the number of poles with the factory default one.

The electric motor I use is 30 poles. But when I typed 8 poles the automatic identification was successful.
 
Back
Top