Vibration & no torque on acceleration, sensorless controller

[slackard]

Howdy,

Crystalite 5000 series motor, 72v 40a Sensorless speed controller.

After hitting a bump today i started getting the jittery vibrations I've seen described in a couple other threads and in this youtube video: https://www.youtube.com/watch?v=FIWSeFec1SA.

The trouble in those threads was always a cut hall wire or damaged hall sensor.

My motor has hall sensors and a 5 wire plug, but it isnt connected to anything as my speed controller is "sensorless". Nonetheless, i inspected the wires and do not see any breaks or damage externally. I have not yet removed the side cover, since because i have a sensorless controller i dont think a hall sensor wire could come into play.

As an added complication. i also disassembled my speed controller for inspection. No obviously damaged or burned ICs, but i found that one of the large capacitors had wiggled loose and was floating free (assuming this is what broke free when i hit that bump today). I found it's home on the board and soldered it back in place, but still no change in the behavior of the system. still acts just like this : https://www.youtube.com/watch?v=FIWSeFec1SA.

So, the motor still turns and up to a decent speed, but i get intense vibration and no torque at low speed. That tapers off as speed increases, but the amp readout seems to jump around, sometimes 35a, sometimes 4a, and sometimes 0a, with half to 3/4 throttle.

Something is out of phase, or some other nonsense to cause this. Replacing that loose capacitor didnt help, assuming when that popped loose, something else cooked and im ¯\_(ツ)_/¯

Keeping in mind that i'm on a sensorless controller, can a short in one of the 3 mains cause this vibration, & should I be pulling the side cover for internal motor inspection? Or, is this likely a partially cooked SC?

Last question, what size gear puller do i need? Seeing a lot of 6" options at o'riley auto but assuming that's too small?

any advice or guidance is much appreciated!

Thanks!

 

[slackard]

Update:

Strong continuity on the multimeter between all three main leads. With my meter in the 200 Ohm setting, i read 0.4 between each of the main leads.

Doesnt look like a broken or weak connection on the main leads. Willing to pull the side cover if necessary, but still not sure if thats the right course of action. Leaning towards needing a 2nd speed controller to test/replace

Kinda flailing in the dark here, so again, any advice is much appreciated.


Thanks!

 

[d8veh]

I would suspect one or more mosfets in the controller is/are faulty.

If the capacitor that broke free was the one across the battery wires. The free leg would be carrying battery voltage, so it could do a lot of damage to anything on the pcb that it came in contact with. Maybe it wiped out one of the transistors that switch the mosfets.

You need to try a new/different controller. There's no reason to suspect the motor.

 

[Ykick]

Try the blown MOSFET test - http://www.ebikes.ca/learn/troubleshooting.html

 

[slackard]

Ok, thanks for the replies... just did the mosfet test...

With my multi-meter in continuity mode, here are my results;

Control test, touching multi-meter leads to each other: display reads 000, audible beep

G+B: display reads 497, no beep
G+R: display reads +/- 500, then rapidly climbs until reaching 1, no beep

Y+B: display reads 497, no beep
Y+R: display reads +/- 500, then rapidly climbs until reaching 1, no beep

B+B: display reads 497, no beep
B+R: display reads +/- 500, then rapidly climbs until reaching 1, no beep

Not sure how to intemperate the 497 values, but definitely not seeing a dead short on any connection.

Next steps? back to the motor? without a spare of either, its tough to rule either one out completely.. Still looking for as much guidance as possible.

Thanks

 

[slackard]

For good measure i put the multimeter in 200k ohm mode and did the same test..

Control touching multimeter leads to each other: 000

G+B: 9.4
G+R: starts at 9.4 & slowly climbs to 20+

Y+B: 9.4
Y+R: starts at 9.4 & slowly climbs to 20+

B+B: 9.4
B+R: starts at 9.4 & slowly climbs to 20+

All three leads are behaving the same way and again, no dead shorts.

Maybe the climbing values have to do with the collector charging or discharghing? If i short the Red and black wires, the climbing values reset to their baseline levels before climbing again.. Im thinking this is normal transistor operation but do lack that essential MSEE

Thanks again to anyone who can help me along here!

 

[amberwolf]

Climbing value is caps charging.

When that capacitor broke loose, then whatever voltage spikes it was there to suppress suddenly wouldn't be suppressed and would get to whatever parts they shouldn't. If those parts were sensitive enough and the spikes were high enough, they could be damaged by those, sometimes in a way that just makes them flaky, and not actually totally failed. That flakiness can often be such that it *only* exhibits a problem at some particular operating range, and only while powered on. So any unpowered testing that's done may well show no problem.


However, it is much more likely that there is an intermittent connection problem somewhere, either a short or an open (most likley) that may appear ok when you're handling the wires/connectors but then when they're being ridden/operated is not quite enough connection to work as desired.

Given the symptoms, it's probably a phase wire from the controller to the motor, either at the connector, or inside the motor, but possibly even a conductor broken inside the insulation, anywhere along the wire lenght from controller to motor.

 

[d8veh]

I think that the capacitor (assuming the one across the battery) is to give extra instant power rather than for smoothing. The mosfets appear to be OK. I still reckon that the free capacitor leg touched something. The PCB works at 5v with a bit of 12v. If that leg touched anything, damage will occur.

 

[slackard]

This all makes sense.

My motor/SC connector is not the stock plastic nonsense, rather a set of fat 200A, gold plated bullet connectors, so I should be good there.

Based on your comments, and the symptoms (& based on my limited research), I agree it may well be a phase wire. Willing to pull the side cover , etc., but still not sure what size gear puller to get. Also still thinking I'll need a new/spare SC, given my situation of 100v capacitors breaking free & bouncing around the inside of the SC case...

Appreciate the comments & guidance. Happy to consider any additional comments before i get too crazy w $300+ purchases, etc.

Thanks!

 

[slackard]

I think that the capacitor (assuming the one across the battery) is to give extra instant power rather than for smoothing. The mosfets appear to be OK. I still reckon that the free capacitor leg touched something. The PCB works at 5v with a bit of 12v. If that leg touched anything, damage will occur.

The cap that broke free was located near the center of the board, near the mosfets and was bent on its side. I can snap a photo of the offending cap if it helps..

The only other caps near the size of this one were near the corner of the board and those bridge the main output busses. the cap that broke free was nont on the main power bus, rather away and separate..

I'll snap a photo.. stay tuned

 

[slackard]

 

[slackard]

Got the new speed controller in and all previous symptoms have vanished. The issue really was the old speed controller, which failed when that cap broke free.

The old SC does still drive the motor, but not without the vibrations & low torque i outlined earlier.. If possible I'd like to get the old one working again and have been scouring the forums for advice, etc in diagnosing the issue. maybe the controller is shot and replacing this or that transistor or mosfet is a feudal effort.. nonetheless the PWM features do still seem to work, rather it seems like an issue with improper phase or something.. I've got my multi meter handy if anyone has any guidance on Speed controller diagnosis & repair.

Thanks!

 

[d8veh]

The transistors that drive the mosfets are all around where that broken cap leg was. I still think that it touched one of them, which would have instantly blown it. This will give the same symptoms as a blown mosfet because the gate will be permanently closed or open.

 

[slackard]

The transistors that drive the mosfets are all around where that broken cap leg was. I still think that it touched one of them, which would have instantly blown it. This will give the same symptoms as a blown mosfet because the gate will be permanently closed or open.

The FETs are all FB4310Z... The transistors youre referring must be surface mount? not seeing any other dip style ICs, so it must be the small surface mount HY2D & HG1B transistors that are right near the fets... correct?

Will look at testing these with the multimeter later today... Not thrilled about trying to replace one of these surface mount chips by hand :/

 

[amberwolf]

It's not exactly easy but it can be done (I've taken some off a dead contrller and moved them to an almost-working one to fix a phase, once).

 

[slackard]

Ok..

admittedly i do lack that essential MSEE, but fwiw, i do own a multi-meter and have done my share of at-home soldering..

Not knowing a ton about electronics repair, though, i watched a bunch of videos on testing transistors, etc., & Found this video, which for me made a lot of sense... https://www.youtube.com/watch?v=_3G8t-cV1d8

Using the methods outlined in the video linked above, i used my multi meter to test all the transistors and diodes on the board with the hopes of isolating a faulty component.

There are 3 phases, and 3 sets of mosfets and 3 sets commensurate transistors, on the board. So, I expected to find the faulty component by testing each, and comparing to their counterparts from the other phases. Should work, right?

Alas, no luck. Seems like all the components behave similarly to their counterparts on each of the 3 phases & as such I am unable to isolate any single component as faulty.

So.. im totally willing to admit im beyond my depth and that i should simply scrap this controller.... but as a last ditch effort.. does anyone have any further guidance for me? Anything that moves me towards a working SC would be much appreciated..

Also, Does anyone do or recommend someone for SC repair? I spoke to the guys at Electricrider.com and theyre not interested in servicing this unit at all.

Also, fwiw, my New controller is not "sensorless" like this toasted one. I had my first ride on the new (hall sensor enabled) controller and the acceleration performance is awesome compared to the (now cooked) sensorless controller.. A lot more power and a lot less noise. A nice upgrade to say the least.

Thanks again!

 

[d8veh]

It's not the mosfets that have blown, but more likely one of the smd transistors that opens the 12v to the mosfet. They're all lying on the PCB next to the capacitor leg.

 

[slackard]

It's not the mosfets that have blown, but more likely one of the smd transistors that opens the 12v to the mosfet. They're all lying on the PCB next to the capacitor leg.

Right, I was already operating under this assumption and the SMD transistors are what I tested, and the diodes. It seemed like each component behaves just like the its counterparts on the other two phases, so I am unable to isolate a single faulty SMD component on the board. Possible my testing methods are faulty.. or multiple components could be faulty and behaving the same way, preventing me from isolating any 1. :/

So.. back to my previous message.. can any further insight be gleaned here? Thinking not at this point, and not terribly worried, but still happy to entertain any other ideas.

Anyone interested in a partially blown SC?