Kollmorgen 400W troubleshooting and probably upgrades

[Jon@LiionWholesale]

Hey guys!

Hope you can help with our new project. I found a Tsunami GT scooter (which I guess is made by Currie) that had dead SLA batteries and no charger for pretty much free. According to the owner everything worked great until the batteries died a few years ago and then it just sat in their garage. So the project would be to make up a lithium ion pack (NMC cells - probably Samsung INR18650-29E) for it and bring the Scooter back better than ever. Then we'd see how good the motor was and either run it with the motor as is for a while as phase 1 then upgrade it down the line, or if the motor isn't so great we'd do a different motor and controller right off the bat. Then of course we'd blog about the details and make some good content for people trying to do similar things. We just picked up the scooter.

So the good news: Well the motor has the part number CTI-187-2 on it and from googling it looks like that's the Kollmorgen 24V 400W brushless motor that seems to have a very good reputation around here, at least as of a few years ago.

The not so good news: Just to test everything out I hooked up two spare 12V lead acid car batteries I had in the shop together for 24V and plugged it in in place of the dead batteries. Immediately upon turning on the on switch the main fuse (30A) blew pretty fantastically. Ugh. Seems pretty clear there's a short to ground. I confirmed that the motor has the ESC onboard like it's supposed to (looks like it's potted into the motor housing). First test I unplugged the motor from the rest of the scooter and unplugged the batteries and checked resistance from power to ground on each separate circuit to narrow it down.

Battery harness with no batteries attached: good (open)
Scooter harness: good (open)
Motor power leads: Bad, I assume(0.2 ohms...basically shorted together)

I opened up the motor and it looks like i've got 5 hall sensor wires and the 3 motor phase wires going between the controller and the motor. Looks like all the ESC circuitry is potted and not accessible.

So my question to you is, where do I go from here? I'm reading that the controller on these motors isn't great so I'm guessing that's the problem, but how do I confirm that the motor is good but just needs a new controller? I don't have any other brushless controllers to swap (well other than a traxxas one for my slash 4x4 at home but I'm guessing that won't work). Having an engineering degree that was heavy on electrical stuff helps but motors are one thing I never studied so I'm pretty clueless in this area especially when we're talking brushless.

Since I'm pretty sure that is the cause, once that's confirmed if anyone also has any recommendations for what kind of controller to use instead and how to maximize power from this motor I'd appreciate it. I'd like this thing to be kind of silly powered, I was thinking ultimately we'd probably do a 48V pack although I'm open to input. It won't be driven that far so reliability isn't too much of a concern but I also want it to have a reasonable chance of lasting. Think like a race car, not expecting it to break but at the same time it's not the end of the world if it does.

Any input would be much appreciated, thanks!

 

[John in CR]

The easiest way to test the motor itself is with an ebike tester. Cut the phase wires and hall wire off of the built in controller to run them outside of the motor. Then it's fairly straight forward to test the phases for shorts, and that the halls function. Look up a thread for doing that.

The motor apparently does fine on 48V. You just don't want to go too high with the current and keep the gearing low so the motor runs less stressed. The magnets are a one piece ferrite ring on outside of the rotor, and guys who've tried 60V the magnet ring flew apart. There's some old threads in 2008 or 2009 with more info.

They're solid little motors other than the crap built-in controller. I've got a few of them in a box somewhere just waiting for an appropriate use. I got some for $32/ea when someone at the factory in India screwed up and wired the controllers to make them spin the wrong direction for the drives they were intended.

Once you clip the phase wires can you do me a favor and measure the phase to phase resistance, so I can get a true idea of their current handling? Simply pass a fixed low current through and measure the voltage drop. Resistance is equal to the voltage drop divided by the current. A 12W 12v bulb with a 12 volt battery makes a convenient fixed current source of about an amp.

 

[Jon@LiionWholesale]

The easiest way to test the motor itself is with an ebike tester. Cut the phase wires and hall wire off of the built in controller to run them outside of the motor. Then it's fairly straight forward to test the phases for shorts, and that the halls function. Look up a thread for doing that.

The motor apparently does fine on 48V. You just don't want to go too high with the current and keep the gearing low so the motor runs less stressed. The magnets are a one piece ferrite ring on outside of the rotor, and guys who've tried 60V the magnet ring flew apart. There's some old threads in 2008 or 2009 with more info.

They're solid little motors other than the crap built-in controller. I've got a few of them in a box somewhere just waiting for an appropriate use. I got some for $32/ea when someone at the factory in India screwed up and wired the controllers to make them spin the wrong direction for the drives they were intended.

Once you clip the phase wires can you do me a favor and measure the phase to phase resistance, so I can get a true idea of their current handling? Simply pass a fixed low current through and measure the voltage drop. Resistance is equal to the voltage drop divided by the current. A 12W 12v bulb with a 12 volt battery makes a convenient fixed current source of about an amp.

Thanks so much for the help! That's really helpful.

I just purchased a Lyen ebike tester, hopefully that's still a good one. When it comes we'll see how the testing goes.

I can for sure measure the phase resistance for you. We do tons of testing here on battery internal resistance I can just hook up to one of the load testers. I'll do it when I get the ebike tester and start on this, probably in a week or two. Do I need to do all 3 phase wire combinations?

I did some more searching for old threads, seems like as you say they're pretty good at 48V but 60V might be too much. Seems like 35A is doable @48V with pretty decent reliability. That's more than I need so I'll plan for 30A 48V and see how it goes. I think I'm going to go with the Lyen 9 fet controller, seems good.

Do we have any idea how much power or motor speed this thing runs at with 48V? Really hard to find actual specs on this thing.

 

[John in CR]

If you can get rpms, then simple measurements give great info using Mile's spreadsheet. The info needed is Kv, rpm per volt (which also gives you Kt, the motor's torque per amp...phase-to-phase resistance in the windings (which also gives you an idea of the current limits you can run)...and finally the no load current at 2 different rpms.

When trying the wiring config, the best advice I can give you try them with only small pulses of throttle. I use a couple of thin alligator clip jumpers on one of the battery connections to limit current like a fuse. When trying different combos, keep either the phases or the halls static, and try all 6 possible of the ones you're swapping. That will give you a valid combo for either forward or reverse. Once you think it's the right combo, then measure the no-load current, which should be quite low with a valid combo. To spin the other direction, simply swap 2 any 2 you were keeping static and go back and find the right combo of the originals you were swapping. Just because the motor spins doesn't mean it's correct. A valid combo will start smoothly and quietly, and low no-load current at WOT confirms it's correct.

 

[Jon@LiionWholesale]

Well I got my Lyen tester.

Luckily I found a pinout for this motor here. So connecting it was easy with no trial and error required.

I checked the phase wires first. If I spin it slowly I get no lights. If I spin it pretty fast by hand then all three phases light up simultaneously. Is the correct behavior? I see the documentation says that they should light in sequence, but then it notes some geared motors they will light up simultaneously. Do you think this is the correct behavior for this motor? I emailed Lyen and he said it's good before I finished this post but I'll leave the question for input anyway.

I also saw some people recommended a quick check where you spin the motor by hand and connect the phase wires together. I did verify that spinning the motor was easy without anything connected, then connecting any of the two wires together made it much harder, then connecting all 3 made it really tough to turn.

I also checked continuity from the phase wires to the casing and none had continuity which is good.

I checked the hall sensors with the Lyen tester and they look good, lighting in sequence.

So it sounds like the motor is likely good. Any other possibilities for failure that I haven't tested yet?

While I was in here with the motor and controller completely disconnected I did some more testing and found the ground wire on the controller's hall sensor wires has continuity to both the ground and power on the power connector going into the controller from the battery. I'm gonna go ahead and say that is my problem.

Also, John in CR I did your phase resistance test. I ran a quick 7.7A pulse since I couldn't find anything better at 12V and got 0.7V drop across the phases so that should be 0.09 ohms. The voltage drop was consistent across all 3 phase combinations. Does that sound in the right ballpark?

 

[John in CR]

The motor sounds like it's all good, so just clip the wires going to the potted controller. Don't bother trying to get that stuff out, because it will just turn into air space.

.09 ohms phase-to-phase resistance is nice, and means the motor can handle a healthier current than I would have thought.

Other than running an exterior controller, the only mod I've thought would be a good improvement is to make a fairly wide and deep cut into the ridges of the aluminum shell for adding some large thick heat sink fins. Then a centrifugal blade on the drive shaft to draw air thru the fins would keep it cool.

Of course, now I'm more interested than ever into turning a pair of mine into outboard motors for a lightweight catamaran. If that resistance is correct, then I'm sure one could outdo the small size torqueedo, and being direct drive and an inrunner it could be much more quiet and efficient too. I'll definitely include one in the motors I need to get measurements for Miles motor spreadsheet, and then we'll have some hard info to base decisions.

John