Crystalyte Phoenix motor problem

I've been working with a Crystalyte Phoenix e-bike setup for the past few weeks, but I'm having trouble with my motor.

A few weeks ago the motor (rear mounted) hopped out of the mounts and ended up damaging the power leads, causing them to short out against the metal inside body of the motor. I had to replace my controller, so I now have a new, presumably working unit.

When I turn the controller on the motor has quite a lot more resistance than when its getting no power. When unpowered I can easily turn the motor. When the power is on and I manually turn the motor, it also clicks along. Also, the resistance isn't constant - it feels like as I turn the motor the magnets inside are attracted to each other and the closer they get the harder it is to turn.

When I turn the throttle the motor will turn, but it makes the same clicking sound that it makes when I manually turn it under power.

I've check the hall effect sensor leads following the instructions from the Crystalyte UK website, and it seems to be working fine. I've also checked my battery voltage, and that's not a problem.

Any help here would be appreciated. Is it possible that the motor was damaged from the short out? Thanks!

On your problem I'll defer to the motor experts here, who know much more than I about stuff like this. Having a 5304, I'm curious though: under what conditions did your X5 jump out of the dropouts?

Were you jumping the bike? Under full acceleration? At what speed? Did this cause an accident?

Was the axle seated as deeply as possible?

Was the axle seated snug against the dropouts, or was there room for twisting movement?

Were the bolts reasonably tight?

Do you think the bolts may have loosened?

Were your dropouts damaged either before or after this happened?

If you fix the motor, what's to prevent the same thing from happening again?

It sounds like one of the hall sensor signals is whacky. It probably draws much more current than it's supposed to also.

Possibly the hall signal wires were damaged at the same time. Most likely failure modes would be an open wire (pulled apart), or shorted to another wire. If one signal wire shorted to another signal wire, it might "look" OK when you test it with a meter, but the pulse width would be messed up.

Try measuring resistance between the three hall signal wires (controller disconnected). I have no idea what it should be, but it should not be zero.

Another possibility might be the wiring is different with the new controller.

If the resistance goes away when you turn off the power, then the phase wires are probably OK.

The motor jumped out of the dropouts when I was starting from a dead stop. I think the problem was that I didn't torque down the nuts enough. Fortunately this was just down the block from my apartment and aside from feeling stupid about it no one was hurt. I plan on building a new frame next year, and for that I'll have much beefier dropouts to avoid similar problems.

I looked at the hall effect troubleshooting guide here: http://crystalyte-europe.com.linux19.hostbasket.com/checkmotor1.htm but since those instructions are for a different controller and motor I'm not sure that I'm testing the correct wires. Testing the leads with a multimeter shows that the sensor is at least outputting a signal as the motor turns, so I'm pretty sure that its at least somewhat functional. If anyone has any advice on how to properly check it and make sure that it works with the controller, please let me know.

Also, the controller is a 4048.

I did open the motor and rewired everything, since the initial accident tore up the shielding. I've gone over the wiring several times and I'm pretty sure that everything is set up correctly. Inside the motor there are no obvious signs of damage to the windings or hall sensor, but it is possible that I'm missing something.

How would the motor behave with no hall effect sensor? Am I correct in thinking that at high speeds the sensor is less important? I'm curious about this because the problem I'm having does not go away at higher speeds when I test the motor out on the bike (stationary, ofcourse).

fechter: I measure the resistance between the leads of the hall sensor, and when the controller is off the values are in the kilo-ohm range. When the power is on the multimeter reads max resistance (10 mega-ohms). When the power is on there is still continuity between all the leads.

Those readings seems kinda weird to me, does this suggest that I have the hall effect sensor wired up incorrectly?

Thanks for the help guys, I really do appreciate it!

This "clicking" noise you talk about. I have a 408 and it clicks when you are "backing up" the bike with the controller on but no throttle. I assumed this was normal but I don't like doing it. I figured it was the sensors. Don't know if this is the same type of clicking that you are having.

The hall sensor cable has 5 wires. One is ground (usually black; battery neg), one is ~10v steady to power the hall sensors (red), and the other 3 (usually yellow, green, blue) are the signals from the 3 hall sensors.

If you power the controller, no throttle, and measure the voltage of each hall signal wire as you slowly rotate the wheel, you should see the voltage toggle between high and low. Make sure the signal is toggling on all 3 signal wires.

If one hall sensor has a broken wire, the motor will not start from a dead stop in some positions. It will start from other postitons, depending on the alignment of the magnets with the poles, so to test, try starting the motor with the throttle many times and make sure it always starts from a stop.

With one hall out, one phase won't fire, so the motor will have less torque and will be noisier and have a jerky output, but it will still reach about the same top speed unloaded.

Your problem sounds more like two signal wires are reversed or two phase wires are reversed. Color codes on the controller vary depending on model, so matching everthing up correctly is not always obvious.

If wires are reversed, the motor will draw much more current than normal.

Do you have a way to measure the battery current?

I've got a watts up meter, so I'll measure current through the phase wires later on today.

The problem I'm having sounds consistent with that your describing for one or more faulty wires or sensors. The motor will NOT start from a dead stop at all positions. From those positions I have to manually rotate it a little bit and then it will go. Once it's spinning it will keep going as long as I give it throttle, but it does click and clearly doesn't operate normally.

This afternoon I'll test each of the signal wires and I'll post what I get. If one or more of the hall sensors is bad, does anyone know where I can get a good quality replacement via the internet?

Thanks again!

Okay, I tested the hall sensor leads, here's what I've got:

There are five wires: red, black, yellow, green, blue

I measured the following voltages between them:
red-black: ~6v, constant (doesn't change when the wheel is turned)
black-yellow: alternates between 0v and 14v when the wheel is turned
black-blue: alternates between 0v and 14v when the wheel is turned
black-green: 14v constant - doesn't change when the wheel is turned.

Does this indicate that the sensor with the green signal wire is bad? Am I correct in thinking that if the wire was detached from the sensor it would read 0v constant?

Looks like you found the problem. Yes, if the wire was disconnected, it would read zero. It might be shorted to the red wire.

Try disconnecting the controller and measure resistance between the red and green wire. If it's near zero, you have a short. It might be hard to distinguish between shorted wiring and a bad hall sensor.

I'm not sure where to get a replacement hall sensor if that's the problem. I think someone else has done it before. I would try DigiKey first.

Can you read a number on the hall sensor? I have a motor apart at home I can check if I remember to.

I've disassembled the motor and confirmed that the wiring is intact. On a test bench the one hall sensor won't change from 13v, even when directly exposed to a magnetic field, so I'm pretty sure that it's faulty.

I was able to order a replacement, but in case anyone else is wondering, on the actual sensor it says:

40A
606-LF

Anyone have any tips or suggestions in replacing the sensor? I'm planning on taking it out with a file and putting in the new one with an epoxy and silicone sealant.

This will do the trick...

http://uk.farnell.com/jsp/Semiconductors/Integrated+Circuits/OPTEK/OHS3020U/displayProduct.jsp?sku=1202503

Farnell sell to anyone (£20 miniumum order, they do FETs too)

Ah... you have the sensor, a bit of heat will degrade the epoxy making it easy to remove, heat it with a soldering iron ans pull.