How to tell if it is the controller, battery or motor that gave up?

[Chris Maluszynski]

I have an old Wisper 905 that I upgraded with a 48V battery and a 17A controller some time ago.
It was working great, until earlier this year, as my wife was pulling a heavy trailer with the bike it suddenly started cutting out power if PAS was set above 3 (out of 5). On throttle it would cut out immediately.

I posted a thread about it here https://www.pedelecs.co.uk/forum/threads/power-cuts-out-as-soon-as-load-increases.37619/ and got some good suggestions, but I now wanted to start over with a new thread as there was never a conclusion. I kept using the bike on PAS 3 and it did the job for me.

The other day though, it gave up entirely. Everything looks good, battery is at 54V, but nothing happens. Well maybe initially there is some kind of little "ticking" sound from the motor, but then nothing. On any level of PAS or throttle.

So what died? The battery? Even though it's showing full voltage? The controller (again!) Or the motor? Could I have killed the 250W BFUn motor with supplying it with too much power? Or could it be anything else? Someone mentioned a wire could have gone resistive, but how do I check for that?

Any other ideas?

I'm considering giving up on this bike, but it's an old dear friend, and I was hoping to keep it alive for a while longer.

 

[vfr400]

As mentioned before, the battery can be eliminated as a cause by a simple observation. If the LCD switches off when the motor cuts, the fault is between the controller and the battery. If it stays on, the fault is between the controller and the motor.

Electronics can't repair itself, so if it doesn’t work then does, it's not the electronics, though there are a couple of exceptions.

First, work on the basis that the electronics are OK, which means that you have a connection fault. The most likely would be the motor bullet connectors at the controller not tight enough, or the connector at the motor end not in all the way to the marked line.

The two exceptions are: 1. The temperature sensor is cutting the controller. In that case, the controller would fel hot when you touched it. 2. Dodgy mosfet. If you over-loaded them, maybe one nearly blew and has become compromised so it breaks down under load. That's just a theory. I'm not sure that it's actually possible.

Normally, problems are caused by simple things. You just have to use logic to find them.

 

[Chris Maluszynski]

Thanks vfr400.

I will look at the connections. But I wonder if I ever saw any connection at the motor end? The cable goes into the wheel as far as I know, are there any connections inside the hub? Which marked line are you talking about?

Re temp - as this happens immediately at startup, everything is cold, so it doesn't sound very probable.

Your theory about mosfets sounds very plausible to me, so if I don't find a faulty connection, I guess I'll be ordering yet another controller.

I wish I could get a 9fet controller into the box on the bike, but unless these have shrunk in size last time I checked, that won't be the case, so I suppose it will again have to be the 6fet one.

 

[Nealh]

The battery or controller is imv where the culprit lies.

If the controller is set for current control then as well as working in PAS3, both PAS 1 & 2 will work as they draw less current. If the bike cuts out using PAS 4, PAS 5 & throttle at all times then it looks as if the battery simply can't handle the current asked of it.
PAS4 delivers 50% / 8.5a whilst PAS5/throttle deliver 100% / 17a.
One will need to wire up temporary a voltage gauge/meter so one can see what occurs to the voltage just before it collapses. Using a KT lcd is no good as it will cut out when lvc is detected so one needs a totally independent reading that doesn't cut out. If voltage sag is high then cell aging means they are passed their best but can be mollycoddled for many years of use at lower power or as a get me home battery. Eventually the cells will be no more then scrap or one could reutilise them for torch use or other light use.

The 6 mosfet controllers suffer thermal cut out during certain uses like high PAS and slow riding on inclines (towing a heavy load may produce a similar effect) , one can tell if this is the case by touching the controller ( being careful as they will burn skin) after such a cut out event. The mosfets can reach the max temp quite easily under some conditions and temporary short out thermally, but once cooled will work again. They will though remain compromised and will readily short out under similar circumstances, I had one and temporarily solved it by adding an extra heat sink to the case and a pair of 5v mini fans. Eventually I got bored with extra rig moral and opted for a 9 mosfet model for more reliability.

 

[Nealh]

In the old thread the cells were stated as LG and battery as 14.5ah, do you know the actual cell it self ? LG M29 (2900mah) is one of the new type and is a replacement for the quite poor MG1 (2850mah).

I would simplify your cutting out as suspect one, battery unable to supply the current demand. Evident as bike cuts out during high PAS and throttle use.

Suspect two , controller mosfets if the controller is hot to the touch after cutting out.

 

[Chris Maluszynski]

Hi again Nealh!
Thanks for your valuable input as always.

Not sure if I made it clear, but since the previous thread, the development now is that it no longer runs _at all_ not on any PAS level, even with the battery fully charged. There is no cut out, as it never starts to run in the first place. (the only evidence of life was some faint "clicking sounds" from the hub motor at some point. The display shows no signs of any problem, does not lose power, and keeps showing battery voltage 54V

The 6 mosfet controllers suffer thermal cut out during certain uses like high PAS and slow riding on inclines (towing a heavy load may produce a similar effect) , one can tell if this is the case by touching the controller ( being careful as they will burn skin) after such a cut out event. The mosfets can reach the max temp quite easily under some conditions and temporary short out thermally, but once cooled will work again. They will though remain compromised and will readily short out under similar circumstances, I had one and temporarily solved it by adding an extra heat sink to the case and a pair of 5v mini fans. Eventually I got bored with extra rig moral and opted for a 9 mosfet model for more reliability.

This would confirm vfr400's theory about compromised mosfets. Could this be what happened at some point, and then the mosfets have gradually degraded since and now are completely dead, thereby entirely cutting out any power to the motor?

I wish I could test the battery on a different bike in order to rule out or confirm that it is the culprit. Maybe I should ask around locally if someone has a bike that has this kind of battery and could test it. The other path is to again replace the controller and see if that solves the issue.

I guess this will be my path forward:

1. Check for bad connections at the controller end (motor seems not to have connections externally)
2. Try to find a different bike to test the battery
3. Order new controller (should I get the exact same one?)

Thanks again.

chris

 

[Nealh]

Use the simple how too mosfet test I posted in your other thread, if one gets a failed fet result a new controller will be needed.

 

[Chris Maluszynski]

Use the simple how too mosfet test I posted in your other thread, if one gets a failed fet result a new controller will be needed.

Thanks!! Didn't realize you posted it there. I'll attach the link if anyone else wants to see:

"One can measure the resistance on the phase wires to deduce whether one or more have failed.
Easy to follow guide."
https://ebikes.ca/documents/BlownMosfets.pdf

 

[vfr400]

The battery or controller is imv where the culprit lies.

If the controller is set for current control then as well as working in PAS3, both PAS 1 & 2 will work as they draw less current. If the bike cuts out using PAS 4, PAS 5 & throttle at all times then it looks as if the battery simply can't handle the current asked of it.
PAS4 delivers 50% / 8.5a whilst PAS5/throttle deliver 100% / 17a.
One will need to wire up temporary a voltage gauge/meter so one can see what occurs to the voltage just before it collapses. Using a KT lcd is no good as it will cut out when lvc is detected so one needs a totally independent reading that doesn't cut out. If voltage sag is high then cell aging means they are passed their best but can be mollycoddled for many years of use at lower power or as a get me home battery. Eventually the cells will be no more then scrap or one could reutilise them for torch use or other light use.

The 6 mosfet controllers suffer thermal cut out during certain uses like high PAS and slow riding on inclines (towing a heavy load may produce a similar effect) , one can tell if this is the case by touching the controller ( being careful as they will burn skin) after such a cut out event. The mosfets can reach the max temp quite easily under some conditions and temporary short out thermally, but once cooled will work again. They will though remain compromised and will readily short out under similar circumstances, I had one and temporarily solved it by adding an extra heat sink to the case and a pair of 5v mini fans. Eventually I got bored with extra rig moral and opted for a 9 mosfet model for more reliability.

The display would switch off if the battery cut, then it would have to be manually switched back on again to resume.

The motor connector will be about a foot from the axle if there is one.

 

[Chris Maluszynski]

Update:

I tested the mosfets with a multimeter. No continuity at all. Just "1"s all the way. Not sure if this is an indication of some kind of fault? According to the instructions it seems like maybe one could expect a lot of resistance but maybe not necessarily that?

If this seems to pass the test, can fault at the controller be ruled out? Or could it be another issue?

I should maybe add that the light and horn works as normal. The display shows all the appropriate indicators when brakes are pressed, throttle engaged or pedals turned (pas). But the motor is dead silent.

 

[vfr400]

Update:

I tested the mosfets with a multimeter. No continuity at all. Just "1"s all the way. Not sure if this is an indication of some kind of fault? According to the instructions it seems like maybe one could expect a lot of resistance but maybe not necessarily that?

If this seems to pass the test, can fault at the controller be ruled out? Or could it be another issue?

I should maybe add that the light and horn works as normal. The display shows all the appropriate indicators when brakes are pressed, throttle engaged or pedals turned (pas). But the motor is dead silent.

You never told us whether the LCD went off when the motor cut.

Show us how you set your meter to do the measurement. It would be very unusual for all mosfets to blow, so I suspect measurement error.

 

[Chris Maluszynski]

The LCD never cuts out. Battery voltage shown on the LCD does not drop.

If I understood the mosfet-test instructions correctly, the measurements should _not_ show continuity, as that would indicate blown mosfets. The mosfets should have high resistance, but the question is if the multimeter should not at least for some of them show a little current passing through (a resistance reading other than max=1).

I followed the instructions: probes in the com and voltage ports and multimeter set to continuity measurment. I also tested it at the lowest ohm setting with the same result. Measuring between ground and phase wires, and between positive and phase wires.

 

[vfr400]

Meter should be set to 200k or whatever you have for resistance.

 

[Chris Maluszynski]

I set it to what the instructions said, "the beep mode" that measures continuity, and the next setting over, don't remember which one that was, probably 200k. If they were blown, I believe it would have shown close to zero and beeped, instead it did not beep (no current passing through) and kept showing "1".

 

[Nealh]

The phase tests are best done using the 200k resistance setting, normally measuring Black wire to phases should show a healthy 9 -10k range but as long as all three are near the same then that side of the fets is fine.
The Red wire test to phases can give a different result depending on the meter, some may show an ascending ohm measurement or simply show infinity or one. Again the results would show fets are good. if one see's zero/nought reading then there is a failure.

If using the diode/beep then no beeps mean all ok, if 1 was shown then all should be ok. A beep would indicate a fet circuit short or failure.

 

[Chris Maluszynski]

The phase tests are best done using the 200k resistance setting, normally measuring Black wire to phases should show a healthy 9 -10k range but as long as all three are near the same then that side of the fets is fine.
The Red wire test to phases can give a different result depending on the meter, some may show an ascending ohm measurement or simply show infinity or one. Again the results would show fets are good. if one see's zero/nought reading then there is a failure.

But does infinity (1) on the black wires indicate a fault or a bad reading or is this evidence of healthy mosfets?

 

[Nealh]

For your battery, try an independent voltage source to see if any changes of voltage occur. Rig up a temporary source to monitor on a ride or on a static test, for a static test have the motor wheel off the ground and apply brake pressure quite firmly (but still able to rotate) to simulate current loading.

 

[Nealh]

But does infinity (1) on the black wires indicate a fault or a bad reading or is this evidence of healthy mosfets?

Any reading of 0 is a fault, 1 indicates no short so all should be well.
To confirm the Black wire test use the 200k value and all three should show about 9.9k or there about, a weak fet may possibly may give a lower reading then the other two.
When I have tested I either get three good results or one bad one on a fet.

A mosfet that gets very hot and thermally cuts out would give an immediate beep/continuity under test, showing a short. Eventually cooled down it will (if not 100% compromised) return a reading of 1 but will remain prone to over heating and future cut outs.

 

[vfr400]

You need the actual resistance reading. No short doesn't mean good! Why do you have to waste all this time writing yourself and making us write arguing about it? It would have been much quicker just to measure the damn thing!

 

[Chris Maluszynski]

Sorry. I really do not want to waste anybody's time. I did the measurements according to the instructions, but only got infinity for results. Maybe there is something wrong with the multimeter. I will stop posting now, and just order a new controller or buy a new bike eventually. Thanks for all your help.