why does generic motor jam on crystalyte contoller?

[SniperGaulois]

I am trying to fit a generic motor that I got for spares on a crystalyte motor, I know it runs okay because I tested it on an identical bike.....all the colours in between the generic motor and the crystalyte controller match perfectly, but the motor is stalling, it goes backwards or forwards by 1 inch and then jams.

there is a major difference in the motors, which both have start immediate capability:
the crystalyte has half as many magnets around the outside... the one on the right in this photo http://crystalyte-europe.com.linux19.hostbasket.com/ROTORS.jpg whereas the generic motor, the magnets are a half as wide and there are twice as many...

is that the reason why they are incompatible? the only other thing that might have happened is when I put araldite glue over the hall sensors to keep them safe from moisture:|

 

[fechter]

Most likely the color codes are not the same between the motor and controller. You'd think they could make this "standard" but every manufacturer seems to have their own variation.

You can easily blow up your controller with the wrong combination, so you need to be careful when testing. With an invalid combination, the current can go sky high as the motor phases are fighting each other. Some kind of additional current limiting is a good idea when testing phase combinations. I use a bench power supply with a current limiter, but if you don't have access to one of these, a large light bulb, like a car headlight (over 5 amps) or resistor in series with the battery might work. A smaller than normal fuse for testing might be a good idea too.

Test with the wheel off the ground.
When testing, you really need some way to measure the current draw. A CycleAnalyst, Wattsup or similar meter is good. A mulitmeter with a 10A current scale will work too.

You may have to try every possible combination of phase and hall colors to get one that works. When you get the right combination, the motor will start by itself from any position, and the no-load, full throttle current will be "normal". For a Crystalyte motor, that will be somewhere around 2A.

Invalid combinations will result in stalling, failure to start from some positions, excessively high no load current, weird noises, reverse direction, etc.

There are 6 possible combinations of phase wires. For each of these, there are 6 combinations of hall wires. That's 36 total possibilities. It sometimes helps to make a chart to keep track of which ones have been tried.

If the motor is made with 120deg hall spacing, but the controller is made for 60deg, no combination may work properly. Crystalyte motors are 120 deg. This is not easily changed.

 

[SniperGaulois]

36 possibilities? I thought that the phase wires were positive, negative, and some kind of third one...ooops! That is outrageous!... the spacing of the sensors is very similar,I would have to open up the motor to look again. would there be an obvious difference and spacing?
is it faster to match the possibilities by opening the motor it is the
to the

 

[steveo]

i've had the same problem as you when i used a chinese knock off ...

black & red all match ..

as fetcher said .. wire the sensors in differen combinations till you get it and do the roll back test till it rolls backwards smothly ..

-steveo

 

[SniperGaulois]

I went off to have a scramble and it looks like the voice recognition has been playing poltergeist in that last post.
it turns out, the sensors inside the generic motor are twice as far apart than the 406. arg... but I have read that the difference in degrees is actually due to the middle one being switched around? now that I have glued really strong glue on them (that wasn't very clever) I don't know which side the writing was on! I will take your advice with the current limiting and I will try now I have six possibilities from checking the sensor colours on the inside-they are indeed different.
Thanks steveo...I guess all the more satisfying once it is rolling. so basically I want i to roll backwards and forwards smoothly? ...ok

 

[fechter]

Thanks for the reminder Steveo. The rollback test is a safe way to test with no throttle.

I just wasn't sure if it would always detect a bad combination. When you have a good combination, there will be virtually no cogging when rolling backward. I'll have to test that idea some more.

 

[SniperGaulois]

I've been through the 36 combinations and the best I could do was get it rolling around thudding. if the sensors are spaced twice as far apart but all facing upwards, probably it means it's a different angle. crystalyte is 60° 5V ? guess I will have to try and access a duplicate controller/change the hall sensors.

when I look on digikey, what kind of sensors should I buy?

 

[fechter]

The physical spacing of the sensors is not what counts, it's the electrical spacing.
This depends on the sensor spacing and the magnet spacing.

A while back we speculated that if one sensor (the middle one?) was faced in the opposite direction, it would make the motor a 60 degree if it started out as a 120 deg. I cannot verify that works.

The Crystalyte motors use a Honeywell SS40A
http://rocky.digikey.com/scripts/ProductInfo.dll?Site=US&V=480&M=SS40A

You can also get them from ebikes.ca

You might want to test each sensor to make sure it is actually toggling beween high and low as the rotor is turned. It is very easy to fry them if you accidently put voltage to the output pin.

 

[SniperGaulois]

Thanks a lot, its way more complicated than I thought just having a brushless motor! I have tried with a multimeter to measure the amount of peaks coming from the sensors on the generic motor...

it seems to be every 20-30° approximately, one time I counted 18 changes over the whole wheel turn, and the next two times I counted about 15, sometimes going over the same spot there is a peak that appears and disappears so I don't know really, but I do know that the generic motor has at least twice as many individual magnets as the 406 design.

 

[fechter]

I wouldn't worry about how many times it switches in a revolution, but as long as each hall output is switching as you rotate, you should be able to make it work.

 

[rkosiorek]

i built a simple tester to test the hall sensor output. it helps identify sensors hat don't work or ones that are out of sequence.



originally posted here:

http://www.endless-sphere.com/forums/viewtopic.php?f=2&t=278&st=0&sk=t&sd=a&hilit=tester&start=180#p37079

rick

 

[rkosiorek]

just for interest, i have used the Crystalyte controllers with various non crystalyte motors. including BMC, Kollmorgen and a Cyclone 500W with a blown controller. with the Cyclone i left the Controller board in place i just isolated the hall sensors and lead those out to an external connector.

when testing controllers after upgrade or repair i often use a small industrial motor as the load.

rick

 

[SniperGaulois]

I figure that is using LEDs to measure the output,- what does the extra precision allow for measuring compared to a multimeter? I am still mostly using electronics lab 2000 for young people to make radios with, I don't understand anything.

-with so many bare wires it is hard to tell when there might be a false contact, especially when changing around 36 possibilities, what kind of systems do you use to go fast and reliable with the contacts?

yesterday the sensors seemed stuck on 2 different controllers, the original controller for this motor reads 5 V and the crystalyte reads 13v, so I hope the sensors have enough resistance to handle it.

-do i have to be very careful with the phase wires so they don't come into contact-do they all have a square wave going in between zero and positive?

it's quite hard going actually-I don't even know if the wheel has the same angle let alone compatible voltage sensors and how to tell what it is. I have some spare SS40A on the way, same type as crystalyte, but I won't even know which way round to fit them!

Here are some photos in case anyone recognises it, it's from a 17cm/6.5"'wide fork, eco brand whirlwind bike, it weighs the same as crystalyte, and the axle bolt is 3 mm thicker and has more than 5 cm sticking out on either side of a regular bike fork...the original controller uses 5V on the sensors. I will fix it up to crystalyte even if it takes lots of work, it's good education for me.

 

[rkosiorek]

I figure that is using LEDs to measure the output,- what does the extra precision allow for measuring compared to a multimeter? I am still mostly using electronics lab 2000 for young people to make radios with, I don't understand anything.

you probably do not have 3 multimeters hooked up to the motor at the same time. so ou cannot see the relationship of the Sensors turning off and on and in what sequence. The 3 LEDs make it a quick easy to see and understand visual representation. with a multimeter you will only see that each sensor just alternately turns off and on. the LEDs not only show that they alternatetly turn off and on but that they do so in a distinct pattern. read the post that i linked too earlier.

-with so many bare wires it is hard to tell when there might be a false contact, especially when changing around 36 possibilities, what kind of systems do you use to go fast and reliable with the contacts?

I usually just use a screw terminal block and i only change the wires on one side of the block.usually l leave the controller side connected and only change the motor side. i use 2 blocks - one for the halls (5 terminal) and heavier a 3 terminal one for the phase wires

yesterday the sensors seemed stuck on 2 different controllers, the original controller for this motor reads 5 V and the crystalyte reads 13v, so I hope the sensors have enough resistance to handle it.

Hall sensors are most commonly rated for 24V max and 50mA current sink. at least the SS40A are. so the voltage is not an issue.

-do i have to be very careful with the phase wires so they don't come into contact-do they all have a square wave going in between zero and positive?

Yes, be careful the phase wires will have a strong square wave. shorting them can destroy the controller by frying the output FETs. While testing the hall sensors DO NOT CONNECT THE PHASE WIRES. leave the ones from the controller isolated and insulated from each other. Once you have all 3 hall sensors working properly you can proceed to find out whether the sequence of halls and phases matches for the motor to run properly. Connect up the phase wires. Leave the hall sensor wires alone. Test all 3 combinations of the phase wires to see if any work. If they do not switch 2 of the hall wires and test all 3 combinations of phase again until either the motor runs or you have exhausted all possibilities.

it's quite hard going actually-I don't even know if the wheel has the same angle let alone compatible voltage sensors and how to tell what it is. I have some spare SS40A on the way, same type as crystalyte, but I won't even know which way round to fit them!

most hall sensors are set so that the face with the label is the most sensitive. that would be standard. the sensor is mounted so it is closest to the surface of the molded plastic housing. bipolar sensor like the SS40A and others used in motors are set so that the open collector output pulls down and latches to GND in the presence of a "S" magnetic field and turns off or opens in the presence of a "N" field.

with the motor apart, you can test the sensors by using a small hand held magnet. bring it close to the sensor and it will conduct. flip the magnet over and it will turn off.

most common mounting is for the numbered face to be visible when glued into place. but look at your motor before you rip it apart to make sure how they had it.

Here are some photos in case anyone recognises it, it's from a 17cm/6.5"'wide fork, eco brand whirlwind bike, it weighs the same as crystalyte, and the axle bolt is 3 mm thicker and has more than 5 cm sticking out on either side of a regular bike fork...the original controller uses 5V on the sensors. I will fix it up to crystalyte even if it takes lots of work, it's good education for me.

that would make it a 165mm spacing between the dropouts. that is about 20mm wider than standard for the usual bikes we are used to. most likely the bike was meant to be used as a domestic product inside China only. i have never seen one identical to it.

120deg. sensor placement is by far the most common. it is a very good chance your motor would be 120deg.

rick

 

[ratatouille]

There are 6 possible combinations of phase wires. For each of these, there are 6 combinations of hall wires. That's 36 total possibilities.

Hum, aren't there only 3 combinations of the phase wires ?

In the new motor, if the hall wires are not in the same position, you only have to change the phase wires sequence for matching the new hall wires.

Original combinations (in the rotation sequence)
Hall\Hall-Controller-Phase/Phase
B-\B-B/B
Y-\Y-Y/Y
G-\G-G/G

New combination : hall wires do not match so we have to find the new phase wires sequence

(wires to switch in italic)

Hall (new)\Hall-Controller-Phase/Phase (new)
Y\B-B/B <- No ------ Yes -> Y\B-B/Y
G\Y-Y/Y<- No ------ Yes -> G\Y-Y/G
B\G-G/G<- No ------ Yes -> B\G-G/B

Hall\Hall-Controller (new)-Phase/Phase
B\G-G/G<- No ------ Yes -> B\G-G/B
Y\B-B/B <- No ------ Yes -> Y\B-B/Y
G\Y-Y/Y<- No ------ Yes ->G\Y-Y/G

 

[fechter]

There are 6 possible combinations of phase wires. For each of these, there are 6 combinations of hall wires. That's 36 total possibilities.

Hum, aren't there only 3 combinations of the phase wires ?

Motor_______Controller
B__________ B B G G Y Y
G__________G Y B Y G B
Y__________ Y G Y B B G

For each phase combination, you have to go through the same premutations for the hall sensors.

 

[SniperGaulois]

Thanks, that is very helpful and it saves anyone reading blowing up their controller! I must have killed the sensors somehow because two of them are staying flat at zero and the other one is staying flat at 13... maybe it's something to do with the bonding glue I put on them. would it be safe to use wood epoxy when I replace them? it's all I have around.

the funny thing is, I managed to get a combination which was similar to having a single hall sensor defunct, it was turning around perfectly but thudding... I think that definitely means that I had the right order but that one of the sensors was dead which is very hopeful! because all the other 36 combinations that I tried the motor wasn't really turning at all...

 

[fechter]

You need to be very careful to avoid shorts when testing. If the output of a hall makes contact with a voltage source (like the hall supply), it usually blows instantly.

I use a barrier block similar to what Rick showed. The "Euro style" blocks from Radio Shack, to be exact. I find them very handy.

 

[rkosiorek]

great minds.......

 

[rkosiorek]

First the sequence of the Hall Sensors must be correct. they are the "BRAINS" of the whole thing. It's like setting the ignition timing and plug order in a Car. to show the physical arrangements for 120degree spacing the following wiring is universal as long as the sensors are spaced 2 slots apart.



the windings are next. the first slot on the right that is straddled by a hall sensor is filled with the wires for phase "a" conveniently it is covered by the sensor for phase "A" the next slot over has the back end of the phase "b" winding. the third slot has the phase "c" wires and is covered by the sensor for phase "C". the next slot has the back end of the "a" phase winding. the slot after that has the wire for phase "b" and it is straddled by the sensor for "b". the next slot after that has the back end of the winding for phase "c".

But anyways. if the hall sensors are wired correctly from right to left the colors that are used by crystalyte are Green Yellow Blue. start there and try all 6 phase wire combinations. one will work properly. if the motor runs in reverse flip both of the phase and hall wires for the yellow - green pair.

rick

 

[ratatouille]

For each phase combination, you have to go through the same premutations for the hall sensors.

oups, sorry, i forgot my math, you're right, 6 possibility

But why do you want to switch hall too ?

I changed the controller of my scooter. i plugged the hall wires as it was originally, the colors matched but the motor didn't rotate well. I switched the phase wires and found the right order for the motor to run smoothly. And that's all.

If you change the hall sensor in the motor, whatever the order and color of the wire. Once plugged to the controller, you have to find the right combination for the phase wires, ie when a hall is excited, the controller has to excite the correct phase wire. So for one combination of hall wires, there are 6 combinations for phase wires but once the good one found (phase wires), it's done. So 6 permutations maximum if you are not lucky.
As there are 6 possible combinations of hall wires, you'll have 6 right combinations for phase wires (but we need just one).

Or there must be something i don't understand and i apologize for disturbance.

 

[fechter]

If you can physically locate the order of the halls by taking the motor apart, it will eliminate some guess work.

It's very difficult to see where the windings start and stop with some motors.

Some cominations of hall sensors will not work no matter what you do with the phase wires. If the motor uses 'alien' color coding and you have to start from scratch, you may need to try quite a few combinations to find a good one.

Out of the 36 possible hall / phase combinations, I suspect there is more than one that will work. Maybe 3.

 

[rkosiorek]

only ONE of those 36 combinations will work properly with the motor spinning in the correct direction. one other will work right with the motor spinning in the opposite direction.

I don't think that i have tried all 36 combinations because i stop when i get to the one that runs properly. but several times i have gone through all 6 with the halls in one sequence, changed the halls tried the next six found the one that worked but in the wrong direction of rotation. then used that as a start point to test the other 2 combinations of simultaneously swapping 2 wires on the hall and the same 2 on the phase to find the one that goes in the right direction of rotation.

rick
l

 

[jakun]

There are 6 possible combinations of phase wires. For each of these, there are 6 combinations of hall wires. That's 36 total possibilities.

Hum, aren't there only 3 combinations of the phase wires ?

Motor_______Controller
B__________ B B G G Y Y
G__________G Y B Y G B
Y__________ Y G Y B B G

For each phase combination, you have to go through the same premutations for the hall sensors.

Hi I bought a new controller for higher voltage and cause the motor didn't run well I triet to do as you explained. Now it works well. But another thing that I detected is, that with an other hallsensor configuration than that which I am running now, the motor was verrry verrrry fast and did run in an other direction. Why is that?

 

[fechter]

I've heard of that before too. I think some combination gives you very advanced timing that speeds up the motor. I think under load, that combination would get into a shorting situation and cause the current to skyrocket.