Can't dial in my new 48V regen system

[Electrobent]

Wiring double and triple checked and then tried swapping yellow and green for both power and hall wires and that got this far.

I have 10AH of NiMH and I road  about 5 miles (peddling for 4) and it dropped from 56V to 47V and seemed pretty useless there. I was expecting more.

I got pulses from all three hall effect sensors but I did not check for when they happened.  Could a sensor be loose in the hub and in the wrong position?

It probably goes about 20mph by itself on the flat with a 20 inch wheel and my 250 pound fat ass.  This seemed a bit slow.

What's next?  Keeping track of the angular distance between hall effect pulses for each phase?  What do I need? Three oscilloscopes?

[Hardcore]

I have 10AH of NiMH and I road  about 5 miles (peddling for 4) and it dropped from 56V to 47V and seemed pretty useless there. I was expecting more.

did you try a different battery as I think a drop of 9v is a bit much for 1 mile.

[Electrobent]

Pretty sure the batts are good.

That was 4 miles of pedelling while the motor was going full tilt and one mile of just the motor.

At the end, down around 48V it would surge:  go fast for a while and then slow down and then go fast again.

How hard is it to open these up to see if the Hall Effect Sensors are where they are supposed to be?

Its got to be easier than getting a response from Golden Motor!

[Hardcore]

you got:

GM 48v battery pack
GM 1000W wheel
GM regen controller

what about taking the magic controller and ride it sensorless

[Electrobent]

My battery pack is not GM
and my controller is not magic

All three Hall Effect sensors blink the 4.9 volts.

What I don't know is if they are blinking when they are supposed to.

I suppose one may be in the wrong place.

Need to find a 4 channel oscilliscope, build something with LEDs, or just take the cover off an look.

[Electrobent]

OK--my electronics tech built the LED flasher jig to test the hall effect sensors

Turning the wheel in the correct direction we got the following color patterns:

Blue
Blue and Green
Green
Green and Yellow
Yellow
Yellow and Blue

this repeated 24 times in one wheel rotation

Does anyone know if this is the correct order?

[Bikemad]

How hard is it to open these up to see if the Hall Effect Sensors are where they are supposed to be?

Check out this hall sensor replacment article on Instructables by Jeremy Nash

Link taken from this post

OK--my electronics tech built the LED flasher jig to test the hall effect sensors

Turning the wheel in the correct direction we got the following color patterns:

Blue
Blue and Green
Green
Green and Yellow
Yellow
Yellow and Blue

this repeated 24 times in one wheel rotation

Does anyone know if this is the correct order?

With 46 magnets, you should get 23 sets of pulses per revolution, but 24 is near enough.
(The first and last are probably the same one, but have been counted twice)

I don't know which order they should be in, but at least they do appear to be working correctly.

If the hall sensor signals are getting all the way to the circuit board inside the controller (no broken wires, poor connections or soldered joints etc.) Then my next check would be to make sure that all three of the phase outputs are:

(A) being sent from the controller
(B) actually reaching the windings inside the motor
(C) are in the correct sequence

Suggestion No 1
You should be able to check this by disconnecting the three phase wires and connecting a voltmeter between battery -ve and each of the phase wires in turn on the disconnected phase connector on the controller side.

Activate the throttle and then slowly turn the wheel by hand, the meter should indicate pulses of battery voltage on each of the phase supply leads.

If you don't get any pulses at all, go straight to the next suggestion.

If you only get pulses on two of the three cables, the fault will be in the controller itself, or the wiring attached to it which gave no reading.

If you get similar pulses on all three cables, continue to the next suggestion.


Suggestion No 2
Try connecting two of the three phase wires to the motor and see if it runs.
(You might need to spin it by hand to help it start.)
Repeat this two more times leaving a different phase lead disconnected each time and see whether the results differ.

If you find one of these combinations runs the same with one lead disconnected as it currently does with all leads plugged in, then the problem is either that particular lead or one of the two of motor windings connected to it.
 
I've run out of suggestions now, so let's hope something obvious shows up during these tests.

Alan
 

[Electrobent]

There was nothing between any of the phase wires and ground with the hall wires installed, power on, and turning the wheel by hand.

I opened it up and the phase wires are oriented Blue Yellow Green and if you flip it over the hall wires are in the same order--but that is really the opposite order relative to the rotation of the wheel.

My LED test jig shows my hall firing order to be Blue Green Yellow

So I tried swapping the Green and Yellow phase wires to get things in the right order

It was close but it is still not right.

I put it back the way it was supposed to be and got 24Volts AC between each pair of phase wires if I leave them connected and run the motor.

Charging the bats now and then will systematically switch everything out

[Electrobent]

OK--to summarize so far:

48V GM 20" front hub
48V GM Regen Controller--the big black one with the resistors and diodes on the back of the board
40 NIMH 10AH 5C D cells in series--about 57 volts freshly charged--nominally 48V

Put it all together and am unable to find a phase combination that gives torque on the ground.

The instructions on the GM site for phase setting don't apply to this controller: to step between options you "bump" the throttle for s second--I guess it is an improvement--you don't have to flip the switch.  Finally sort out the procedure with Tom:

• Short purple-black pair
  turn on system
  bump throttle
  observe
  bump throttle
  observe
  . . . . .
  until you figure out the pattern and can stop on the best one--which is still nothing like how the motor really runs.
  unplug the purple-black jumper
  bump throttle again
  cycle power
  test

Now it should smoothly spin up fast.

I could get this far but when I put the wheel to the ground performance was pathetic and it sucked a lot of juice.

Some one Endless Sphere suggested swapping green and yellow on both the power and hall effect wires.  I did this and I think it got a little better but I don't know--I can make it sort of work like this under every combination of wires I have tried.  I just can't make it run right.

We built an LED Hall Effect test jig and it showed that all the sensors were working and firing in the order  Blue, Green, Yellow, and back to Blue again. 

I opened up the hub and found the power wires attached in an arc: Blue Yellow Green

I opened the other side and found the hall effect wires attached in the same order but on the other side of the wheel so in opposite order of rotation.

So I am thinking this is a two step switcheroo:  first we need to swap the yellow and the green to map the observed firing order of the hall effect sensors with the order in which the wires are soldered on the board.  Second, we need to swap the end points--the blue and the green--to reverse the order so the sides of the wheel (power and hall) are going the same way as the wheel is spinning.

I dug out my oscilloscope and put it across Blue and Green; Green and Yellow; and Blue and Yellow power wire pairs.  The patterns looked pretty much the same at full tilt.

My batts are charging and I'll try in on the ground tomorrow.

I must say that the workmanship on this hub is far better than it was on my GM 36 V from 2006 or so.  I had been planning to seal the flange with some sort of gasket goo for waterproofing but decided that the fit was tight enough without it and that I would just make a mess if I put it on.

I have photos of inside the hub and the scope screens for each pair if I could figure out how to include them.

[Bikemad]

I had been planning to seal the flange with some sort of gasket goo for waterproofing but decided that the fit was tight enough without it and that I would just make a mess if I put it on.

I have photos of inside the hub and the scope screens for each pair if I could figure out how to include them.

If you put a thin smear of Vaseline or grease around the flange joint it will help to keep water out and prevent corrosion at the joint.

To attach pictures to your post, clik the Additional Options... link and then the "Choose File" button and browse to the picture file on your computer.
Click (more attachments) to add up to 4 photos max.

Alan
 

[Electrobent]

OK Here are my pics:

Phase wires and hall wires

[Electrobent]

And here are the 'scope shots:

Do they look like they are supposed to?

[Electrobent]

OK I tried this:

Suggestion No 2
Try connecting two of the three phase wires to the motor and see if it runs.
(You might need to spin it by hand to help it start.)
Repeat this two more times leaving a different phase lead disconnected each time and see whether the results differ.

If you find one of these combinations runs the same with one lead disconnected as it currently does with all leads plugged in, then the problem is either that particular lead or one of the two of motor windings connected to it.
 
I've run out of suggestions now, so let's hope something obvious shows up during these tests.
 

When I unplug the yellow power lead and run the motor it runs slowly,
If I unplug the green or the blue, it runs at the same speed as before.

I can swap hall wires, go through the phase detection procedure again, and try it again and it is still the yellow wire that slows things down.

I put an ohm meter across the motor leads and all pairs read 0.6 ohms.

I don't like the conclusion I draw.

Anyone have a rosier one?

Why do all my scope traces look good AND all my coils measure .6 ohms.  Both of these should not be true.

[Leslie]

The motor seems to check out as long as you have done the testing and wiring properly.  The batteries seem good, as you are sure of this, the throttle works fine as the motor would eother tun constantly or not at all.  The last thing left is the controller.  Check if the wiring to the controller hasnt got any lose strands of copper touching things and the hall input wires are connected properly.  Ive had a single fet go out on a single channel and the performance sucked as it limped along.  Its a hard thing youre doing.  I do hope you fix things right.

The halls maybe right on the motor but maybe wired wrong in the controller.  Check your pins in your connectors.  Have you tried resetting the phase setting after you changed the hall order?

[Electrobent]

I was hoping it was the controller, but I just swapped out the power wires (G:B, Y:G B:Y, Controller: Wheel) and it is still the yellow wheel wire that slows things down when disconnected.  I think that the coil between the blue and green wires is bad.  But why do I still get 0.6 Ohms across it?