Cyclamatic controller problem...

[evendine]

After many years of good service, my standard Cyclamatic controller cut out recently while going downhill, which was a surprise. Symptoms were clicking at the motor when the throttle was opened and no power.

Got it home, opened it up and on looking it over spotted a charred spot on a 5v wire leaning against the shunt. Checked the MOSFETs and one set was fried. Decided to check the electrolytics while there and found the one next to the fried MOSFETs tested bad. Replaced all three associated with MOSFETs along with the 5v regulator smoothing cap, which was bulging and the fried pair of MOSFETs.

Plugged it in to my Lyen controller tester (Halls + phase wires to tester, throttle and power to the controller) and get no motor output. The 5v indicator on the tester is flashing regularly, as is the controller LED (which from other threads i understand indicates the controller is basically functional). The 5v rail looks a bit low (4.7v) and have checked throttle voltage range (0.87 - 4.0v) to make sure the controller will initialise properly.

The tester throttle indicator doesn't light up, interestingly, no matter which throttle position. Thinking about replacing the 5v regulator, but guess the processor may be damaged?

The bike works fine with another controller. Any thoughts on what the problem with the original might be much appreciated...

 

[evendine]

Think the microcontroller must have been taken out when the MOSFETs blew...

Bought a KU65 and fitted it, but although lights on the power meter come on and the motor sounds as it usually does, no rotation of the wheel. Could this be a stuck clutch, perhaps?

 

[Deleted member 4366]

Is the motor running backwards? Disconnect the controller from the battery completely, short its battery wires to discharge the main cap, reconnect and try again.

Your PAS will probably work backwards until you flip the magnet disc.

 

[evendine]

Interesting - so you can 'hard reset' the controller as you might do with a laptop which is hanging on boot - makes sense

Getting maybe initial 10mm (slow) forward movement when power applied via new BMS supplied thumb throttle and standard phase connections only. Initially thought must be the phase wires / motor halls misaligned, but little difference when phase power wires tried in all possible configurations (some move backwards, some a bit forwards) .

The odd thing is, in standard config, if you put your hand on the motor, feels like it's running and it sounds right. Have checked the motor halls with a multimeter and they look ok volts wise, but not sequencing even if wheel rotated manually. That's why i was wondering if the clutch (if there is one?) or maybe gears are the problem...

Btw, running with no PAS connected initially - assume that's ok for test with this controller?

 

[evendine]

Have just realised this is a different bike to the one mentioned at the beginning of the thread (have 3 in the house the the mo!), so not sure hardware is ok. Motor phase coils measure around 0R3. Two motor halls are high (4.6v) and one low (0v)...

 

[Deleted member 4366]

The KU65 has automatic phase and hall sequence detection, You can put the wires in any sequence and it'll still work.

I had a motor running backwards. I tried every sequence, but each time, the controller adjusted itself to run the motor perfectly backwards. The reason was a slight hall sensor position error, which can't happen on your motor.

My guess is that you have a hall sensor issue. The controller can run sensorless, so try it with the hall connector disconnected, then to reverse it swap any pair of phase wires.

You need to test that all the hall sensors are switching properly. With the halls connected and the controller switched on, stick your probes up the back of the hall connector and check the voltage while you rotate the wheel slowly backwards. you should see it jumping from 0v to 5v a few times each rotation. check black to blue, green and yellow.

 

[evendine]

Thanks Dave - found the KU65 through one of your earlier posts - really impressive Cyclamatic compatible controller for not much money. Ordered 4 from BMS along with some thumb throttles and power meters and they delivered them within a week.

The weird thing is that in sensorless mode, the motor sounds as through it's running normally, but the wheel doesn't turn. If you put your hand on the motor casing it feels as though the rotor is spinning internally.

Checking the hall sensors (when plugged in) as you suggest, they don't change, but one is switched and the other two not, with normal looking voltages. If the motor isn't engaging with the drive, would this explain the lack of hall activity, do you think, or should the halls change if the rotor is turning?

 

[Deleted member 4366]

Each hall should switch on/off when you rotate the wheel backwards. That's definitive. Can we have an answer: How many of the three halls switch on/off when you rotate the wheel backwards? I can't understand what you're saying about them.

To run the motor sensorless you must disconnect the hall sensor connector otherwise the controller will attempt to use the sensors. Can you conform that you tried it with the connector open?

When you run sensorless, the controller can't tell the direction of the motor, so it will go one way or the other. Whatever way it goes, it can be reversed by swapping any two phase wires.

It seems that you can hear/feel the motor running. Can you confirm that you can hear the characteristic whine of the motor and not some other noise, like hum from a stationary motor?

If it's running backwards, you should see the wheel slowly turning backwards when in the air because there's a small amount of friction in the clutch. If it rotates slowly forward with no power so that you can hold it with one finger with light pressure, then it's the clutch or gears that are faulty. If it's the gears, you can often feel that by rotating the wheel backwards by hand while it's running, which will engage any good parts of the gears so you feel a bit of a kick.

Clutches normally stick on rather than off when rusty, but I wouldn't discount rust as being a cause. I've heard of a couple of Q100s that lost drive because a roller's spring got caught under the roller, though that happens from over-powered ones. Whether it's the clutch or the gears, the motor would have to be opened to fix it.

 

[evendine]

Would be the first time I've had a problem with Cyclamatic motor hardware, despite much abuse!

Some answers to your questions:

How many of the three halls switch on/off when you rotate the wheel backwards? None - Lyen tester shows yellow and blue phase lights both on and green off. When measured with a meter, yellow and blue are 4.6v, green 0V. No change when wheel rotated in either direction.

To run the motor sensorless you must disconnect the hall sensor connector otherwise the controller will attempt to use the sensors. Can you conform that you tried it with the connector open? Yes

Can you confirm that you can hear the characteristic whine of the motor and not some other noise, like hum from a stationary motor? Yes - the sound changes with throttle setting and seems fairly normal if a bit rattley...

In sensorless mode, with a little throttle and moving the wheel forward manually, for perhaps 40% of each rotation there's a little forward motion of the wheel under power, but it's very weak.

 

[evendine]

Managed to get the motor apart - not hard to do, just removed the set of screws on the cassette side, the large spindle nut on the other side and tapped it through. A fair amount of water + rust in there - will post pics shortly...

 

[evendine]

 

[evendine]

Gears look fine. Not sure how to test the clutch. Rotating the rotor turns the planetary gears as expected, so not obvious what to check next...

Thinking about regreasing - the dark grey gunk is moly grease, which is lithium based apparently, so ok to mix with white lithium i guess?

 

[evendine]

Regreased gears with lithium, cleaned rust from surfaces as far as possible, tapped the clutch cover repeatedly with a hammer to try to free it, then reassembled and refitted the motor. One useful trick is to locate the axle in the dropout on the non cable side first and tighten the nut - makes getting the other side into the dropout much easier, more so if you flex the chainstay a bit. Not sure why, but it now runs with or without sensors, so a good result.

Checked the halls with the Lyen tester to make sure it really is running in sensored mode and all seems well. Suggests perhaps rotor and stator were seized as well as the clutch, so no relative motion between them.

One niggling issue is that feels like there's drag when free-wheeling. Could it be that the clutch has now stuck on? Not sure what to do other than run the bike and see if it frees up over time...

 

[Deleted member 4366]

Rotate the wheel by hand. You can feel the motor turning when you turn the wheel backwards. If it feels the same forward, the clutch is now stuck on. You can sometimes free it by whacking the wheel in the forward direction, but I guess that it has rust in it, which stop the rollers moving properly.

This looks like the correct replacement clutch, but you have to take yours right off to check it's the same and has the same number of teeth on the gears.:
https://bmsbattery.com/ebike-kit/635-gear-set-for-bafang-k5-swxu-swxh-motor-ebike-kit.html

 

[evendine]

Yep - you've described the symptoms exactly - drag in both directions and the free-wheel seems to be working (so can't be that), as can still hear the 'tick tick' in forward motion.

Funny - was reading one of your earlier posts about whacking the wheel with a hammer earlier on - the direct approach! Will give it a go...

Do you think it might be possible to drill lubrication holes into the clutch cover and free it that way? Thought about trying to remove the gear assembly, but not sure how to go about it tbh.

Thanks for the link, good to know it's possible to replace the gears / clutch for so little. Really impressed by the BMS kit which arrived earlier in the week...

 

[evendine]

Whacked the wheel with the side of a big wrench and it freed up third time! Genius! Made my day!

Noticed that rotating the wheel backwards caused it to stick again, but a few more whacks and it freed. Had to repeat about 5 times, but seems rotating the wheel backwards then whacking exercises the clutch, as it now appears to work fine even after rotating backwards.

Thanks again Dave!

 

[Deleted member 4366]

The ticking is the freewheel in the gears, which frees the chain from the hub. The motor's freewheel is silent. It that rivetted disc that holds the three gears. Yours is very rusty.

 

[evendine]

Yep - was thinking about how the water got in there but guess unless the motor were airtight, you're going to get condensation in cold weather. Gave everything a thin coat of lithium grease, which should help prevent surface rust in future hopefully...

 

[Deleted member 4366]

The most common way for water to get in is when people take the wheel off to fix a puncture, then put it back on upside down. The cable needs to go downwards, which is illogical if you don't know why.

Other than that, parking the bike on the prop-stand in the rain doesn't help because water can run down the axle and around the bearings.

It can get through the side-plateplate join if the screws are not tight, so a little sealant is always a good idea when you reassemble.

On some disk brake motors, it can get through the disc mounting holes if you don't have a disc fitted, and even with a disc, it can get down the screw threads, so a bit of Loctite there is always a good idea.