Diagnose Pendleton not working

[AndiB]

My partner's Pendelton Somerby stopped working recently.
She heard a low clunk noise just before it stopped providing power assist.

Initially I thought it was a broken hub going into the hub (which certainly didn't help), but after fixing that the bike still doesn't kick in the motor.

The battery is (supposedly) fully charged, with all green LEDs lit on the battery.
The controller behaves normally too: full battery indicator, no flashing lights of any kind.

What can I do to diagnose what's wrong with the bike?

Should I check the battery? Or the fuse? Though the controller is coming on, so I'd imagine there's power coming through...
How can I check there's actually power going to the motor?

Thanks!

 

[Sturmey]

On my daughters 2017 pendleton, there is a small led fitted to the pedal sensor on the inside of the chainwheel. This (If I can remember) normally comes on at startup and flashes when you pedal. (It may be covered in dirt) This failed once due to water penetration. I removed its sealant and dismantled it and dried it out and resealed it with clear builders silicon and it worked. The only other fault that occurred on this bike was with one of the handlebar brake switches but these can be disconnected if necessary.

PS Try pushing the 'walk assist' button on the front of the '790' unit with the bike turned on. Normally, the wheel will move slowly and weakly and if this is the case, you may have a pedal assist sensor problem or problem with the magnet wheel....

 

[saneagle]

The clunk indicates a fault in the way the power comes to the motor.

First thing to check is that the connector near the motor on the motor cable is pushed in all the way to the marked line and not just in tight.

Next thing to check is that the wires aren't melted where they come out of the controller. The controller sits in the compartment at the front of the battery. You have to open the compartment to check in there.

Lastly, if it happened when you were going slowly up a steep hill, you might have blown a MOSFET in the controller. The only way to confirm that is by doing a very simple test with a meter.

 

[AGS]

If it’s the same motor as in your other thread then one of the phases is cooked and the motor is toast.

 

[Nealh]

If a phase bullet has melted or overheated at the controller end then the hub should be ok , before coming to any terminal conclusions and worrying the op he needs to fault find first.
If the controller it toast and needs replacing then it is an opportunity to upgrade all to a KT system for £80 - £100.
Back in March I did the same to a neighbours similar step thru bike and it changed the ride for the better.


A simple manual test for mosfet failure is to rotate the drive/motor wheel by hand , if one feels resistance then usually a shorted mosfet has occurred or the wire/ connection has melted. By disconnecting the three controller phase wire bullets the motor resistance shoud free up and freewheel quite well , this will tell you the issue is the controller.

Using a meter to determine if the controller mosfets are at fault or carry out Hall sensor tests is a more definitive way to fault find and it is imperitve to do so if trying to identify a motor Hall issue.

 

[Nealh]

If it’s the same motor as in your other thread then one of the phases is cooked and the motor is toast.

Quite right AGS , I have just seen the pic you refer to.
It is toast.

 

[AndiB]

If it’s the same motor as in your other thread then one of the phases is cooked and the motor is toast.

Quite right AGS , I have just seen the pic you refer to.
It is toast.

Thanks @AGS and @Nealh , it is indeed the same motor as in my other thread.
For future visitors' reference, the "toast" picture is below.

Would it be better to swap the full wheel now, with a suitably rated motor inside, or can I just swap the motor insides?
Any reputable sellers of wheels / motor parts you know of?

Thanks!

 

[saneagle]

Thanks @AGS and @Nealh , it is indeed the same motor as in my other thread.
For future visitors' reference, the "toast" picture is below.

Would it be better to swap the full wheel now, with a suitably rated motor inside, or can I just swap the motor insides?
Any reputable sellers of wheels / motor parts you know of?

Thanks!

You're getting ahead of yourself. First, you test to see what's wrong, then you fix what needs fixing.

 

[AndiB]

If a phase bullet has melted or overheated at the controller end then the hub should be ok , before coming to any terminal conclusions and worrying the op he needs to fault find first.
If the controller it toast and needs replacing then it is an opportunity to upgrade all to a KT system for £80 - £100.
Back in March I did the same to a neighbours similar step thru bike and it changed the ride for the better.


A simple manual test for mosfet failure is to rotate the drive/motor wheel by hand , if one feels resistance then usually a shorted mosfet has occurred or the wire/ connection has melted. By disconnecting the three controller phase wire bullets the motor resistance shoud free up and freewheel quite well , this will tell you the issue is the controller.

Using a meter to determine if the controller mosfets are at fault or carry out Hall sensor tests is a more definitive way to fault find and it is imperitve to do so if trying to identify a motor Hall issue.

I've got a multimeter, how would I check the MOSFET in the controller, please?

 

[saneagle]

I've got a multimeter, how would I check the MOSFET in the controller, please?

Set the meter to read resistance 20k or 200k scale. With the controller disconnected from the motor and battery, stick one probe in the red batttery wire and the other in each of the three motor phase wires. All three readings should be the same and in the range roughly 4k to 24k. Repeat using the black battery wire instead of the red. Typically, a blow MOSFET shows as a dead short, but sometimes, you see other anomalous readings.

 

[Nealh]

BlownMosfets.cdr (ebikes.ca)

 

[AndiB]

Set the meter to read resistance 20k or 200k scale. With the controller disconnected from the motor and battery, stick one probe in the red batttery wire and the other in each of the three motor phase wires. All three readings should be the same and in the range roughly 4k to 24k. Repeat using the black battery wire instead of the red. Typically, a blow MOSFET shows as a dead short, but sometimes, you see other anomalous readings.

BlownMosfets.cdr (ebikes.ca)

Thanks guys.

Just measured the MOSFET:

• Each phase wire to ground / negative: approx. 19.5kOhm
• Each phase wire to positive: approx. 800kOhm

There's no discernible difference to me when rotating the wheel with controller connected / disconnected.

Can't find a wiring diagram, but very naively I would say high resistance to positive is fine, assuming MOSFET is off (not powered)?

Next: measuring motor phase resistances...

 

[saneagle]

Thanks guys.

Just measured the MOSFET:

• Each phase wire to ground / negative: approx. 19.5kOhm
• Each phase wire to positive: approx. 800kOhm

There's no discernible difference to me when rotating the wheel with controller connected / disconnected.

Can't find a wiring diagram, but very naively I would say high resistance to positive is fine, assuming MOSFET is off (not powered)?

Next: measuring motor phase resistances...

800k is much too high. Check again and keep your fingers away from the probes. 800K is typically your resistance.

Neal forgot to mention that you have to rotate the wheel backwards to test for resistance; however, that test only checks for a shorted MOSFET. According to your measurements, all three are open circuit.

 

[AndiB]

On my daughters 2017 pendleton, there is a small led fitted to the pedal sensor on the inside of the chainwheel. This (If I can remember) normally comes on at startup and flashes when you pedal. (It may be covered in dirt) This failed once due to water penetration. I removed its sealant and dismantled it and dried it out and resealed it with clear builders silicon and it worked. The only other fault that occurred on this bike was with one of the handlebar brake switches but these can be disconnected if necessary.

PS Try pushing the 'walk assist' button on the front of the '790' unit with the bike turned on. Normally, the wheel will move slowly and weakly and if this is the case, you may have a pedal assist sensor problem or problem with the magnet wheel....

After replacing cables to avoid any electrical shorts, I now have at last power to the wheel.
Although only in walk assist mode. Normal pedalling does not kick in the motor.

As per your advice @Sturmey , I've checked the PAS sensor: red led turns on with power-up from the handlebar controls. Then it stays solid red and doesn't flash as the magnets move along.
Is there a test to check if it's actually reading the magnets? Like should it flash when I wave a magnet in front of it?

Thanks!

 

[Nealh]

The Hall is just a switch is should pulse on/off as the Hall senses a magnet passing. A solid led suggests you have it wired up incorrectly or there is a wire short.

Show us your wire sequence /configuration for the PAS connection.

One can test the PAS with a meter , probe the back of the PAS connection.
One should see approx. 5v between Gnd & 5v wire and then probe the Gnd to third Signal wire whilst rotating the pedals, one should see the voltage fluctuate between approx. 0.6v - 4.4v or there abouts.

 

[AndiB]

The Hall is just a switch is should pulse on/off as the Hall senses a magnet passing. A solid led suggests a Hall failure/short so sounds like the PAS needs replacing or you have wire dit up incorrectly.

Thanks @Nealh .

It's still the original sensor that came with the bike, and it used to work until recently. So I'm going with "it's broken" and will put in a replacement...

 

[Sturmey]

After replacing cables to avoid any electrical shorts, I now have at last power to the wheel.
Although only in walk assist mode. Normal pedalling does not kick in the motor.

As per your advice @Sturmey , I've checked the PAS sensor: red led turns on with power-up from the handlebar controls. Then it stays solid red and doesn't flash as the magnets move along.
Is there a test to check if it's actually reading the magnets? Like should it flash when I wave a magnet in front of it?

Thanks!

I tested the pendleton pedal sensor by substitution.i.e. I plugged in a spare KT mini type sensor into the controller, stuck my finger into its center and turned the magnet, and the bike sprung into life. So I was fairly sure that the original sensor was faulty. I then removed the sensor from the bike by opening the screw that holds it to the backing plate. On its examination, I could see the the sealing was cracked and water had got in. I completely removed the sealing and dried it out and resealed it and it worked.
This spared me the problem of getting a replacement.
In my situation, there appeared to be four options. The first was to get the exact kingmeter replacement and just transfer the sensor head over by removing the screw etc. The second option was to get a complete right hand replacement and change over the whole unit but that involves completely stripping out the bottom bracket. The third option was to fit the compact kt V12L sensor on the left crank as this appeared to work . The fourth option was to attempt a repair to the original sensor and as this worked out, it turned out to be the easiest option in the end.
But anyhow, back to your bike. What actually was the original fault to this bike?

 

[AndiB]

I tested the pendleton pedal sensor by substitution.i.e. I plugged in a spare KT mini type sensor into the controller, stuck my finger into its center and turned the magnet, and the bike sprung into life. So I was fairly sure that the original sensor was faulty. I then removed the sensor from the bike by opening the screw that holds it to the backing plate. On its examination, I could see the the sealing was cracked and water had got in. I completely removed the sealing and dried it out and resealed it and it worked.
This spared me the problem of getting a replacement.
In my situation, there appeared to be four options. The first was to get the exact kingmeter replacement and just transfer the sensor head over by removing the screw etc. The second option was to get a complete right hand replacement and change over the whole unit but that involves completely stripping out the bottom bracket. The third option was to fit the compact kt V12L sensor on the left crank as this appeared to work . The fourth option was to attempt a repair to the original sensor and as this worked out, it turned out to be the easiest option in the end.
But anyhow, back to your bike. What actually was the original fault to this bike?

Good question... I think it's several faults, really:

• The hub cable had been mounted badly, either from purchase or during an in-warranty repair by Halfords. As a result it had worn through the insulation and snapped a phase wire. Replaced the cable...
• Opening the motor (a bafang rear hub, 250w) showed a few charred windings, though the motor still works (with new cable) in walk mode.
• Now it looks like the PAS isn't working either. Ordered a replacement today.At £10 it felt like an easier option to trying to prod, breadboard, etc the original. Especially since the seller accepts returns. So

Just hoping that I don't have to also replace the controller. At that point I would have repaired/replaced all electrical parts of the bike...

 

[saneagle]

Good question... I think it's several faults, really:

• The hub cable had been mounted badly, either from purchase or during an in-warranty repair by Halfords. As a result it had worn through the insulation and snapped a phase wire. Replaced the cable...
• Opening the motor (a bafang rear hub, 250w) showed a few charred windings, though the motor still works (with new cable) in walk mode.
• Now it looks like the PAS isn't working either. Ordered a replacement today.At £10 it felt like an easier option to trying to prod, breadboard, etc the original. Especially since the seller accepts returns. So

Just hoping that I don't have to also replace the controller. At that point I would have repaired/replaced all electrical parts of the bike...

That's a good learning experience. Hopefully, you'll stay and show others how to do it.

 

[Sturmey]

Good question... I think it's several faults, really:

• The hub cable had been mounted badly, either from purchase or during an in-warranty repair by Halfords. As a result it had worn through the insulation and snapped a phase wire. Replaced the cable...
• Opening the motor (a bafang rear hub, 250w) showed a few charred windings, though the motor still works (with new cable) in walk mode.
• Now it looks like the PAS isn't working either. Ordered a replacement today.At £10 it felt like an easier option to trying to prod, breadboard, etc the original. Especially since the seller accepts returns. So

Just hoping that I don't have to also replace the controller. At that point I would have repaired/replaced all electrical parts of the bike...

By the way, I forget to mention to check the gap between the magnetic disk and the sensor, which I think should be about 2 or 3mm. My daughters pendleton had a really nasty habit (occasionally) of the chain falling off the inside of the front chainring and sticking to the magnets until I fitted a chainguide and shortened the chain.