How to reverse motor direction - old Chinese car

[Klasik]

Hi all,
I bought this old car. It was working fine except some vibrations, so I decided to open motor and change bearings. After reassembly motor runs backward. Car has normal gearbox, so now I have 5 speed for reverse, and only one slow speed forward . I am desperate, because I dont know how to reprogram controller, and motor encoder is new for me.

It looks like BLDC motor, 3 phase, rotor is magnet. but dont have hall sensors. It has some kind of encoder (see photo), with 6 wires going in. I would be very happy if anybody recognize it and tell me how to reverse this motor safely ?? (maybe just switch any of two phases)

Something must have happened, maybe controller went to its defaults for some reason. I connected everything as it was before. Any idea guys ? Thank you!

motor:

controller:

 

[amberwolf]

Did you reconnect all the wires to the same places they came from?

If not, you might want to check your reference pics (or notes) from before taking it apart, and compare to after to find out what's different.

If you don't have any pics or notes, you can try just swapping two phase wires, since it's unlikely the controller would have any automatic detection routines.

But be gentle with throttle when retesting, and watch for excessive current, to make sure you don't have some sort of "false positive" forward combination. (depends on how it's encoder works)

 

[Klasik]

I am quite sure that all wires are reconnected to the same places. In the beginning I couldnt believe my eyes. Main phase wires are marked A B C, on both controller and motor, and 6 small wires are hard-wired(one connector).

I checked my photos to compare. During reassembly I also opened controller just to see whats inside. (I removed one board, to see what is under - another board , at this point I decided to close everything as it was).. This is one thing which could be reason, controller went to defaults or sth.

I also added my own watt meter to the system. This should have no influence.

I tried to disconnect encoder, then it doesnt run at all. I wanted to swap two phases, but noticed that there is still high voltage on them (battery disconnected). This fact, with combination of 2 degrees celsius in my garage, convinced me to go home and ask folks in forum.

Later I will try to swap phases, but I can imagine that it wouldnt be enough. I really would like to know what is this encoder. Could it be hall sensors ? Why 6 wires instead of 5... How to make it run reverse ?

Problem is that I have no documentation. Manufacturer of motor and controller seems to be non existent anymore. All my e-mail go to spam. I was looking forward to run this car in spring. Now I am stuck. Help appreciated !!

controller inside:

 

[Lebowski]

Could it be an encoder like this ?

https://endless-sphere.com/forums/viewtopic.php?f=30&t=31584#p458133

 

[amberwolf]

I am quite sure that all wires are reconnected to the same places.

Then it is strange that it is running backwards. :/

Unless the encoder was physically removed from the motor and reinstalled offset so it is reading magnets incorrectly (assuming that's what it does), I don't know a reason it would run backwards now.

If the encoder has a separate magnet or optical ring that it reads, if that ring was reinstalled offset or flipped over, perhaps taht could cause the problem.


If something got disconnected inside the controller when it was disassembled, perhaps that is causing it?

Or...maybe the reason the car was available in the first place is that it randomly swaps motor directions?

I wanted to swap two phases, but noticed that there is still high voltage on them (battery disconnected).

That's probably from the internal capacitors in the controller. Those should discharge over some time (minutes to hours).


If you post pics of things (best way is to attach them directly to the post in the attachments tab below the box you type in), it's possible someone might recognize something and be able to better help.

(if the links that are in your posts are supposed to be pics, they don't show up for me).

 

[Klasik]

Thank you for support! I was stuck sitting in freezing garage few times, now I have feeling that I am getting somewhere.

amberwolf: yes, I already posted couple of pictures, I will attach all of them now as you said.

Controller has big 23pin signals connector. One of wires is probably FWD/REV function. Unfortunatelly I have no idea which one. So as you said, maybe this wire was somehow grounded/ungrounded during reassembly. (unintentionally...weak connection). Connectors look good, so I dont think so, but maybe.

Encoder has no optical ring or magnet, only this what you can see on photo. Small metal rotor inside this resolver, has some kind of "groove and tongue" system, so it should be in correct position. Phases are labeled and the rest of connections are hardwired - connectors.

I did some google research, and I think that it is SIN/COS type of resolver. Going to garage to think and maybe take some more pics.

 

[amberwolf]

Controller has big 23pin signals connector. One of wires is probably FWD/REV function. Unfortunatelly I have no idea which one. So as you said, maybe this wire was somehow grounded/ungrounded during reassembly. (unintentionally...weak connection). Connectors look good, so I dont think so, but maybe.

Can be wire broken anywhere along it's length, or at the back of a contact pin. Hard to diagnose without actually continuity testing everything from one end to the other.

Can also be contact pin that is not fully seated in connector housing, so ti back out as the connector is plugged in. Push on the metal fo the contacts from the back of the connector (if you can reach this) while it's plugged in but powered off, and see if any move inward. If they do, they might make a better connection now.

Encoder has no optical ring or magnet, only this what you can see on photo.

If it is a SIN/COS unit, then it *does* have a separate magnet ring on the rotor just for it, otherwise it can't do position sensing; this ring is magnetized in a gradual pattern around it from north to south in some number of poles, and the paired hall sensors are then offset at a distance that puts them 90 degrees out of phase from each other along that pattern.

The way yours looks, it appears to have the magnets on the rotor core inside the encoder's coil ring. Probably the coils get a voltage induced in them by the passing magnets, which are shaped the way they are to cause a sine-shaped pattern. The coils are probably wired in two sets (probably in series) and the controller compares them to get the SIN vs COS signals.

It could be some other type of resolver with more sensor inputs; I'm no expert at resolvers.



This is a SIN/COS sensor on a powerchair motor I have. It's not directly on the rotor, but rather on a separate bit of the shaft. I don't have a SIN/COS controller so I don't know what happens if the ring is offset from the rotor position.

Small metal rotor inside this resolver, has some kind of "groove and tongue" system, so it should be in correct position.

If that groove/tongue system is symmetrical, same all the way around, it won't stop it from being offset during reinstall, which may change how the controller senses rotor position, possibly even direction. Depends on how the controller works.

But it looks like there's four bolt holes around the sensor ring, so those ought to align with four holes on the housing it attaches to (even if they never put any bolts in there).


Most likely that small rotor is the magnet ring required by SIN/COS.

 

[major]

Hi,

Looks like a resolver to me. Try rotating the stator 90°, the aluminum ring with the 4 holes. Just a hunch.

major

 

[Klasik]

I tried to take car for a spin as is (backward.) And noticed, that regen doesnt work too. This makes me think that perhaps controller is in "reverse mode".
I am still uder impression that my work(reassembly) was done good. It is really strange.
I am trying now to understnand where all wires (23 pin connector) goes, and hopefully be able to select\exclude, which ones could be FWD REV switch. If this leads nowhere, I will probably try switching phases or rotating this resolver. Thanks for suport!

 

[amberwolf]

Before you spend a bunch of time with wiring that probably isn't the issue:

As brought up in a previous post, do those four bolt holes on the resolver ring have matching holes (or pins) in the main housing?

If so, try lining up the resolver ring with those.

 

[Klasik]

Hello amberwolf, my wire search was not very successful. Some wires leads to such places, with very bad access. Then I wanted to rotate resolver. I removed this aluminium ring, wanted to rotate resolver, but it seems to be pressed in. ALuminium ring is separate part. I cant move it, but also I can see that is aligned as should be (matching pin/holes). Also I am sure I did not move it before. So I decided to leave this fragile part as is.
But I have some relieve finally.
Since there was nothing else left, I started trying all kind of combinations of phases and resolver signal wires. Most of them, motor didnt work at all, was stuck. In the end of the day (almost gave up), I found one which works, and pushes car forward.
Switching phases wasnt enough, I had to phisically cut resolver wires and find combination by luck.
I dont know what happened. I still think that my wiring after assembly was good. Now is different than in the beginning for sure.

But OK, this desperate repair worked.. simple things are the best. This one wasnt for me. If I had service manual for this controller, at least, or so. That would be different story.