Ecrazyman controller with Kollmorgen?

Thanks for posting that Peter. Yes, it can make your head spin. I use a piece of scratch paper to keep track of the combinations.

In therory, if you know the wiring for one direction, it should be possible to predict the wiring for the reversed direction.
 
Ok, here is CW
[pre]Power
A Green
B Blue
C Yellow

Hall
A Blue
B Green
C Yellow[/pre]
 
I rode the bike to work. I made it about 6-8 miles before the LVC started kicking in (3.9k ohm resistor). My Fluke said the battery voltage was 31.8v. I forget the initial voltage, but I think it was a little bit above 40v. This was using 4 Makita li-ion batteries from Jonathan.

FYI for those who come after, modding the 48v controller for 36v operation is very easy. It took me about 10 minutes, most of it spent getting all the darn screws off. I have a really awful soldering iron and it still wasn't too bad.
 
Sure... I knew I should have gotten the 48v version! lol. It should be interesting to see the 36v version when I get it to find out how different it is. From the auction photos it appears the 36 and 48v ones use the same case.

Did you notice a big power difference between the stock Kollmorgen 24 volt controller and the e-crazyman at 36 volts?
 
Now I've got myself worried about my decision to get the 36v version since it's only is rated at 350 watts. I've read that the Kollmorgen is rated at 400 watts (even though electricscooterparts.com lists it at 300). Hopefully it will still provide more rpms than the stock controller.
 
fitek said:
Ok, here is CW
[pre]Power
A Green
B Blue
C Yellow

Hall
A Blue
B Green
C Yellow[/pre]

Peter, are the A, B and C wires like shown in my picture on the first page?
I want to make some updated pictures with properly color coded wires.
 
fechter said:
fitek said:
Ok, here is CW
[pre]Power
A Green
B Blue
C Yellow

Hall
A Blue
B Green
C Yellow[/pre]

Peter, are the A, B and C wires like shown in my picture on the first page?
I want to make some updated pictures with properly color coded wires.

I don't think so. The +/- for the halls appears to be correct, but to reverse direction to CW you don't touch the middle wire of the phase wires (you show as C) nor the middle wire of the remaining 3 halls (you show as B). Both halls and phase wires should be consecutive A-B-C. Since I haven't tried to connect the external controller yet (only reversed the rotation), I'm unsure whether for the phase wires A should be the black wire at the motor or the red one, only that B is the white one.
 
Fechter,

I have the 72V version of ECrazyman. Is the color coding consistent with what you posted, so once I know which phase wire is A, then yellow-blue-green is sequential as you colored the hall wires? With fingers crossed I'm down to a 50/50 shot at correct wiring on the first try. Let me know if you want original colors for the hall sensor wires of the stock motor, since those in your pic aren't the stock wires. I don't want to confuse the issue further by posting them now.

John
 
The color coding of the 'stock' wires does not seem to be consistent.

I'd like to make a diagram showing where they go right to the hall board. This will eliminate the variability in the wire colors.

If I get time, I have one burried in the garage somewhere and I could verify the correct wiring.
 
Fechter, I wrote down CW and CCW rotation for the Crystalyte controller. Here it is, using the lettering from your picture:

CCW
Power
A Grn
B Yel
C Blu

Hall
A Yel
B Grn
C Blu

CW
Power
A Grn
B Blu
C Yel

Hall
A Grn
B Yel
C Blu

As you can see I swapped B and C on the power wires, but I had to swap A and B on the hall wires.

So I drew a new Koll diagram in my notebook. The power wires are exactly the same, but the hall wires, instead of A B C, read B A C. Just swap B and A in the picture.
 
Now I'm beginning to understand why some end up going through almost random combinations to figure out correct wiring. What happened to putting A B & C consecutive? Now to top it off, why are you guys both putting ABC consecutive for one and not the other. Then to make matters more confusing, one of you does it on the phase wires and the other does it on the hall wires. I understand that Fetcher picked his up from a pic on VisforVoltage, which was an untested guess at a wiring change for increased speed (which didn't seem to make sense btw), but Fitek, what made you post BAC on your hall wire pic?

This Kollmorgen motor is as good a place as any to start, since with it opened up we can see the physical placement of the wires, so the ABC can begin with a logical start for a database. Then just plug in the color codes for a known controller proven to work, and just add motors and controllers to the simple database as they become known. That way guesswork wiring should only take place once by one person for any new unknown motor or controller. Plus, if they get a look inside the motor, guesswork could be pretty well educated. I'd prefer 1-2-3 , since it is a firing sequence, but someone already started the ABC thing.
 
Well the tabular results I posted are based on Fechter's image. Fechter's image is good because it names each wire-- so it gave us a common language to speak. But if you use the A B C coding from Fechter's image to attach a controller, it won't work because the hall wires aren't in order. So using Fechter's image, power wire X doesn't necessarily correspond to hall sensor wire X, get it?

I think that's where the confusion stems from.

My image just shows which power wire goes with which hall wire. If you know the firing order of your controller, you can hook it up to the Kollmorgen using my image. I've got it working with two different controllers so far (Crystalyte and Ecrazyman).

I've edited my image to use the more conventional U V W phase coding so future users don't confuse our A B C tables used for testing with which phase each wire is associated with.

I also plugged my website on top of my thumb :)
 

Attachments

  • koll-innards.jpg
    koll-innards.jpg
    84.3 KB · Views: 4,282
UVW???????????? If that's some kind of standard, whoever came up with it should be publicly flogged! Your Kollmorgen hall sensors are out of order for their physical arrangement on the board, so the factory person who did it should also be taken out and flogged. My Kollmorgen's follow the logical sequence, and the physical sequence corresponds to the arrangement on the board. Not considering the power supply and ground wires for the halls, to reverse direction just swap the 1st and 3rd wires for both the hall and the phase wires...nice, easy and logical, though I don't know yet if the phases and halls are in the same logical sequence of both being in the same order in terms of rotation. According to the pics and sequence here http://visforvoltage.org/forum-topi...kollmorgen-24-volt-bike-motor-needs-reversing, they came pretty close to perfect logic except that someone messed up and got the ABC clockwise order confused and did a big batch as CBA. We should thank them, because we wouldn't have these quality motors so cheap without this screwup. I just wish they had done it with the 600W model with the inch long heat sink fins instead of the little bumps.

The 72v Ecrazyman controller I have seems to follow logic too, since it has 3 hall wire colors that are the same as the 3 phase wires. Plus the hall power wires are, you guessed it, red and black. Now I assume these wires are logically connected inside the controller, otherwise someone at that factory needs a flogging. Crystallite also seems to have their ducks in a row and follow some logic with yellow, green, and blue hall and phase wires.

A bit of an off topic question for Fechter, if I have a brushless hub motor whose hall sensors are placed evenly around the armature (18 sets of windings, 6 windings separate each sensor), is that a 120° setup? Logic seems to indicate that it is, unless the 20 magnets around the shell means otherwise.

John
 
I just received my ecrazyman 36v controller in the mail today. I'm going to post up some photos later today of the inside just in case anyone is interested in what they look like (much smaller than the pictures I've seen of the other ones, fewer fets). Now I'm really worried that it won't give me more power than the Kollmorgen's internal controller.

One strange thing though... He sent me what looks like a hall sensor with it. It's a plastic wheel with magnets and a wire with a pickup sensor. What are these used for?
 
there is no wrong way to wire up the phase wires or the hall sensors, either the motor will rotate clockwise or counterclockwise depending on how they are hooked up, there are no combinations that just wont work, if it spins the wrong way just either swap two phase wires, or swap two hall wires, whichever is easier.
 
Ok... Here's the inside of the 36 volt ecrazyman controller:

2621146073_ccdeae39f4.jpg


Is there anything that jumps out that I should change/upgrade while I have it apart (other than buying a new controller :)).

I may try to yank the controller off the 'ol Kollmorgen today and give it a shot.

I noticed that ecrazyman sells a 48 volt version that's the same size as this one. Can a controller with only 6 fets handle 48 volts?
 
dirty_d said:
there is no wrong way to wire up the phase wires or the hall sensors, either the motor will rotate clockwise or counterclockwise depending on how they are hooked up, there are no combinations that just wont work, if it spins the wrong way just either swap two phase wires, or swap two hall wires, whichever is easier.

I'd have to disagree with that. This would be true for a sensorless controller, but for one that uses the hall sensors, there are many combinations that will result in the motor just sitting there and buzzing or the current going sky high.

The A, B, and C designations I have on my drawing are right off the back side of the hall circuit board.
Kollmorgen hall connections Backside.jpg

I suspect the phase wires are the ones that are switched around. The halls sensors are in sequential order.

The new crazydude controller is totally different. 6 FETs could handle whatever voltage they're rated for, but not as much current at 12. I have no clue what the encoder wheel is for.

Both 60 and 120 deg. motors will have the hall sensors sequential, but the spacing relative to the magnet arc will be different. With a 60 deg. motor, there will be some positions where all three hall signals will be high and some where all three will be low. With a 120 deg motor, there will be a maximum of two hall signals high or low at any given time.
 
fechter said:
dirty_d said:
there is no wrong way to wire up the phase wires or the hall sensors, either the motor will rotate clockwise or counterclockwise depending on how they are hooked up, there are no combinations that just wont work, if it spins the wrong way just either swap two phase wires, or swap two hall wires, whichever is easier.

I would have to disagree with that too. I just spent a couple of hours messing around with mine to get it to do something.

Funny thing is, it didn't work with the wiring Fitek came up with. I'll post up the combo that worked for me tomorrow some time.

Is there any way of knowing for sure if you have the wiring right? I just played around until the thing worked. Now it seems like the motor has a TON more rpms than it did at 24 volts. Of course that was the point of going to 36 volts but I'm just a bit paranoid I'm going to burn something up if it's wired wrong.

I just ran it for a few seconds at a time on my stand and it really flies. Not sure what it will be like with me on the bike but I have high hopes.

fechter said:
The new crazydude controller is totally different. 6 FETs could handle whatever voltage they're rated for, but not as much current at 12. I have no clue what the encoder wheel is for.

The controller in the picture is new, it's just the 36 volt version.
 
EVTodd said:
Is there any way of knowing for sure if you have the wiring right? I just played around until the thing worked. Now it seems like the motor has a TON more rpms than it did at 24 volts. Of course that was the point of going to 36 volts but I'm just a bit paranoid I'm going to burn something up if it's wired wrong.

I just ran it for a few seconds at a time on my stand and it really flies. Not sure what it will be like with me on the bike but I have high hopes.
I'm glad I'm not the only one going ???!!! about Dirty_D's comment regarding wiring combo's making no difference.

Maybe there's something to the high speed guesswork setting in that old V forum pic. Assuming correct connections from the factory the blue-green-yellow sequence from the controller should be the same in clockwise order for both the phase & hall wires. To change direction you swap A & C for both hall and phase using the a-b-c sequence pictured in the previous pic from the motor itself (ie green doesn't change). Going from 24v to 36v will result in a 50% rpm increase for factory wiring.

If your motor runs really well now, it was wired correctly at the factory, and you really have a W-V-U connection for the phase wires with a V-U-W for the halls, then the speculation in the V forum that there is a high speed wire combination must be true. If you find that your low end torque isn't up to par, then I believe the blue-green-yellow sequence clockwise for both phase and halls is the correct sequence to get your power back.

This success with what seems (to me at least) to be flipped A-B wires on the phase wires is confusing to say the least. I look forward to hearing how it works on the bike, since I have 3 Kollmorgens that I'd like to put to good use.

John
 
hmm thats strange, i cant see how changing the hall signals around could do anything other than change the direction of the motor, i did it on paper and switching anything just made the motor reverse direction, is it possible to get the grounds mixed up? maybe its only the 60 degree one that needs to be in a specific order, ill check it out.

ok i think i figured it out, if they are 60 or 120 degrees you can mix them up any way and it will work, with the rotation being in a certain direction, but if one of the hall sensors is flipped or is just set up a way to give the opposite output as the other ones for the same magnetic polarity, moving that hall that gives the opposite output screws up the order of signals that go to the controller, but you can still swap the other two halls and all it does is reverse the motor.

does that motor have one of the halls set up like that? it would be easy to test.
 
If the hall and phase wires are not in the proper sequence, one or more windings can come on at the wrong time and fight the forward motion of the rotor. Like having two windings going forward and one winding in reverse. This results in excessively high currents.

Other "wrong" combinations could result in a radical timing advance, which might make the motor go faster for a given voltage, but draw more current and be generally less efficient.

Here's a timing diagram for a 120 deg motor spacing:
 
yup, but for a normal 120 or 60 degree spacing without one of the hall sensors reversed, flipping them only results in the reverse sequence of hall states. another possibility of why it would only work with the hall signals in one order is that the controller logic isn't built to handle the reverse sequence because its not supposed to be biderectional.

heres what i got flipping different halls:

120 degree normal:
001
011
010
110
100
101

120 degree wire swap:
101
100
110
010
011
001

60 degree normal
111
011
001
000
100
110

60 degree wire swap
110
100
000
001
011
111

60 degree with center hall flipped
001
011
010
110
100
101

60 degree with hall flipped, swapping the two normal hall wires
101
100
110
010
011
001

60 degree with hall flipped, swapping flipped hall wire
111
101
100
000
010
011

as you can see 60 degree spacing has its own commutation sequence different from 120 degree spacing, 60 degree spacing can be converted to 120 degree by flipping the output of the center hall. if you swap that center halls wire, it results in a completely different hall sequence that will result in the wrong commutation sequence and the motor wont turn in any direction.

anyway, its still easy to wire if you don't know what wires go where and you don't know if it has a flipped hall sensor. it obviously wont work if the controller expects a 120 degree sequence and the motor halls are 60 degree without a flipped hall to convert to 120 degree.

1. wire up phase wires in any sequence
2. wire up halls in any sequence
3. swap hall wires until the motor turns smoothly in any direction
4. if motor is turning in the right direction go to step 6
5. if motor is turning in the wrong direction, swap any two phase wires
6. you're done :D
 
Back
Top