Gem e2 immobile - error 45, error 81 in high mode

Field windings would not get energized until the pedal is pressed

Ah, more complicated than I hoped. But maybe worth a try.

I’m not sure what it will tell us

If pressing the pedal and applying small voltage/large current to A1, A2 on the motor generated even a little motion then might that eliminate the motor as the culprit?

If we were starting to narrow it down to a fault in either the motor or the controller then, what with the motor being by far the less complicated of the two, I was hoping to established its credentials before I considered opening up the controller.

might that eliminate the motor as the culprit?

I wouldn’t necessarily call it eliminated.

The controller is upset about something. I consider it a sophisticated device that looks at a handful of parameters set in place before it considers it safe to proceed. These params are there to protect itself and everything it is connected to.

Hooking a power source direct to the motor has no pre-check/protection and (in theory) will either produce a reaction or not. The problem my brain is tripping over is that how do you know if a positive running response is correct/normal? Assuming you are not testing at the full 72v volts so you don’t know if the speed it turns is normal. The motor could still be cooked and turn a little, but not like it. To make this a valid test you probably need to do this on several motors and more/less have some sort of control group.

Checking your controller
While I have not done this, I hear you could also stick a 100w light bulb across A1/A2. Pressing on the pedal should produce a variable light show. Then you need is a mirror ball and a pole in the garage and Bob’s your uncle!!

Yes I see your point.

I had thought we had narrowed it down to either

  1. something in motor is broken
  2. something in controller is broken

But I wasn’t considering another highly likely option
3. The controller is seeing something it doesn’t like and therefore not proceeding to apply power to the motor.

But if that were the case (i.e. the controller knows something is wrong), I would have expected to see an error on the display. I don’t (in Low mode).

how do you know if a positive running response is correct/normal

The motor can be turned by hand, but I see no motor motion at all and no volts across A1/A2 when I try to operate the vehicle “normally”. If the motor rotated at all when powered by the bench supply that would tell me that the complete lack of motion is not due to the motor being completely dead.
I do get your point that its possible that a barely working motor might be operating outside the parameters that the controller believes are acceptable.

That the -81 error is triggered (albeit in High mode only) a little while after motion might normally be expected has led me to believe that the controller thinks everything is okay and only complains when it doesn’t see the tacho pulses that it was expecting. But I shouldn’t make too many assumptions about the decision processes in the controller. I have regarded the -81 error as being caused by the controller not seeing pulses it was expecting and the -45 error as a issue only associated with the high power demanded by High mode. If I could get the controller to drive the motor in Low mode, I’m hoping both errors will go away if the vehicle is only ever driven in Low mode.

I didn’t get a chance to do the resistance measurements today. I will try to go to the site tomorrow.
A simple short to earth might explain everything (i.e. controller thinking that its driving current, but its not flowing through the motor; no volts across A1/A2; the -45 error due to no back e.m.f. from the motor). But I am clutching at straws.

Thanks again for your help with this.

Also- In review (so I don’t lose the clues in between all the chatter)

  • the -45 goes away when one of the A wires is unhooked. Still shows the -81.
  • rolling the car a bit will register 1~2 mph on the speedo.

Does the -81 appear instantly?

In review (so I don’t lose the clues in between all the chatter)

I completely agree that its important keep a focus on what we know and what we believe. In the past I have jumped to conclusions that are not yet proven and it has been a distraction. Also some initial observations have later been shown to be under unrepresentative conditions and their results have been misleading. It’s worth re-checking anything we are basing our decisions on.

the -45 goes away when one of the A wires is unhooked. Still shows the -81.

Yes I believe that is the case. I will specifically check again.
It’s a shame that the service manual doesn’t give more details about the specifics of exactly what causes -45.

rolling the car a bit will register 1~2 mph on the speedo.

Yes, though I don’t actually roll it. The vehicle is on axle stands so I use my foot to rotate the wheel whilst looking at the display.

Does the -81 appear instantly?

The simple answer is “No”.
Firstly, I should say that -81 is only initiated in High mode. It can be generated by pressing the pedal very slowly. It seems that some time after the relay has clicked -81 is displayed as I continue to press the pedal gently. My personal assessment is that the controller is expecting pulses and generates -81 when it doesn’t see them, but that is just a theory and I could be wrong about that.
Once -81 has been triggered in High mode and I switch to Low mode, I then see -81 being displayed almost immediately after the relay clicks. [I need to repeat this test to be certain, but its my recollection.] This is the only way to get -81 in Low mode. It reverts back to the original operation after a full power cycle (both key and master switch off and on).

Stuff I believe to be true along with the relative confidence I have in the statements [20% is still true in my eyes, just difficult to prove conclusively]

  1. the vehicle does not move under its own power [100% confidence]. Until the point it failed it had been working - it had been reversed out and was about to be driven off. [90% confidence]
  2. the motor can be turned by hand [100% confidence]
  3. -45 error is displayed in High mode if the pedal is pressed quickly [95%]
  4. -45 is triggered by a certain level of power demand. [20%]
  5. If one of the A cables is disconnected from the controller, -45 is not seen [90%]
  6. If I press the pedal slowly in High mode with A cables connected, -81 is displayed . [95%]
  7. If I press the pedal slowly in High mode with A cables not connected, -81 is displayed . [50% - needs checking]
  8. -81 is not displayed immediately after the relay clicks. It happens shortly after as the pedal is gently pressed. [90%].
  9. -81 is caused by the controller not seeing the tacho pulses that it expects to see [20%]
  10. This would explain why -45 is seen with a quick press, i.e. it is encountered before the -81 is triggered.[20%]
  11. If I switch on the vehicle (which is up on axle stands), and watch the display whilst the front wheel is rotated (with my foot), the display shows up to 2 mph depending on how quickly I turn the wheel. [95%]
  12. The speed sensor is fully functional [75%]
  13. At the point when the motor should be moving, I see no voltage across the A1/A2 terminals on the motor. [40% - needs checking, along with field terminals]

My current possible theories:

  1. There is a short between the A1 and A2 terminals on the motor. All the current that the controller thinks it is putting through the motor is actually going through the short. [20% - I will hopefully know more after I do the resistance measurements on the motor]
  2. The output stage of the controller has died. The controller is driving the MOSFETs (?) but no current is being driven over the A1/A2 cables. [50%]
  3. Some fault condition has been recognised by the controller which causes it to make the decision not to drive the motor. That said, it still clicks the relay in all modes and doesn’t display errors in modes other than High. [5%]

Hoping to go to the vehicle today.

Thanks again.

This would seem to cast doubt on my second theory…

If the output stage was dead, what difference would the cable make?

A few things still to try

  1. Put meter on motor. (check ohms)
  2. Clean out motor (as best as you can).
  3. Check/free up brushes.
  4. Clean the comms if you can. Don’t try and knock the end off. It’s not easy.

I know… it’s odd and throwing me off. Having stuck brushes means the A1->A2 would be an open connection and it should be the same as removing one of the A cables. Possibly looking for a A → case connection? It’s all worth checking just to be sure and a quick service is not a waste of time.

It’s sure sniffing out like it is a controller issue.
You say you aren’t seeing any V coming out of the A cables but I don’t find when/how you are measuring this. Realize- as soon as an error is triggered there will be no output. If no errors appear on low or reverse then check for output in those conditions.

If this is truly a FET problem then it should work in Rev since it uses a different bank of FETs.

I have resistances measurements.
Test probe to Test probe = 0.2 ohms (this is the resistance of the test probe wires and so should be subtracted from any other measurement)
F1 to F2: 1.2 ohms
A1 to A2: 0.6 ohms
F to A: >100Kohms
A to chassis: >100Kohms
F to chassis: >100Kohms

I have voltages. All are with vehicle power switched on, in Low mode with pedal pressed and don’t change with actual pedal position. No errors displayed.
Across A1/A2 (motor connected): 0.2V
Across A1/A2 (motor disconnected): 40V
Across F1/F2 (motor connected): 4.4V
Across F1/F2 (motor disconnected): 69V

I didn’t do the bench supply test as I didn’t think that the field windings would be sufficiently powered.

I tried sliding the brushes (pull back spring with pliers and pull on copper braid). I was able to wiggle (rather than slide) one of them. The other seven all appeared to be stuck fast. How would I go about freeing them? WD40?

I confirmed the earlier observations
With an A cable disconnected removed I do not see -45. I do see -81 (in High mode only) about two seconds after pressing the pedal. Thereafter -81 is also shown instantly after the pedal is pressed and also in Low mode, but not in reverse. To clear this state, I need to switch off both key and Master switch. Thereafter it returns to the state where Low mode does not initiate any errors.

Once triggered, -45 continues to be displayed even if the pedal is released.
Releasing the pedal after -81 will remove 81 from the display but it returns instantly after the pedal is pressed.

The only difference I can find between Low and Reverse modes is that Reverse does not seem to display -81. [75% confidence]

And here is a doozy. The A1 A2 cables are swapped (MA1<->CA2) but the F1 F2 cables are not (MF1<->CF1). [90% confidence]

After an hour or so of 3-in-1oil and WD40, four of the eight now slide. I’m hoping that that I’ll be able to free the remainder in a day or so.

Does this point the finger fairly and squarely at the controller output stage?

What are my remaining options now, other than taking apart the controller and looking for signs of an explosion?

Thanks,
Richard

Which four? One each in 12, 3, 6, and 9 o’clock positions?

Does this point the finger fairly and squarely at the controller output stage?

Not enough info. When were these readings taken? At idle? Someone mashing the pedal? Was 81 on screen?

Did you try a light bulb across the A1/A2 yet?

At the moment, its 2, 1, 1, 0, but I’m hoping that I can free them all up when I next go back to the site.

I found it easier to remove the motor so that I could rotate it to use gravity to help the lubricants penetrate.

??? How is this oriented?

All are with vehicle power switched on, in Low mode with pedal pressed and they don’t change with actual pedal position. No errors displayed.

I will try the lamp test (if I can find one).

I would like to get hold of a clamp ammeter to confirm that no current is passing through the motor. Maybe I can make a measuring device with a loop of wire.

Are there any other experiments I can do? What would confirm that it is a controller issue?

??? How is this oriented?

From recollection…

image

…but I will check. I am hoping that all 8 will be sliding when I next report back.

Ok, At least you have a couple opposing pairs. That should have run half of the armature and gotten some sort of movement. Hard to tell. If the comms are still real dirty it may not have good contact.

For orientation: Yeah, I know this is not a AMD brush pack. it’s just the first pic I found.

I’d say your controller needs to be sent off and looked at next.

What probably happened was that the brushes stuck but still had a bit of contact. If run like this it would move but probably threw some nasty arcs when brushes wore out. These arcs throw all kinds of voltage spikes back up to the controller and blows up the FETs.

Continue working to get those brushes free. Gentle/easy tapping on the carrier vibrates/rattles it a bit to help. On the CARRIER and not the brush. Brushes crack and chip when abused.

We are based in the UK. Do you know who we would send it off to?

Is there any merit in me taking it apart? I’m expecting to see a big black hole where a MOSFET (or MOSFETs) should be. Perhaps I could source replacements and fit them?

Thanks for your help with this.

We are based in the UK

Ugh. That makes it tough.
Is this a T4 controller?
For sure. Pop the lid off. Let’s look for carnage!

I think you identified it as a T6 controller

The best guidance I can find is https://www.youtube.com/watch?v=M26-BlexPKo (though ours is a T6). It seems to suggest that there is a relatively simple test that I can do for MOSFET integrity.

A circuit schematic would be great but I suspect I’m going to be out of luck there. But I’ll need to know what MOSFETs to use as replacements. I hope any dead devices are still recognisable and not vapourised.

Now that I have the motor off, I can stick it on a bench. What tests can I do of it?
I was imagining that I could apply a voltage to the field windings (12V? perhaps using a car battery or a car battery charger) and then ramp up the voltage on the A1/A2 connections using my 10A bench supply. Is that a realistic test?
What voltage should be applied to the field windings?
How much current should I expect the field windings to draw?
How much current should I expect the A1/A2 contacts to draw?

Thanks,