How to bypass internal controller on brushless motors?

friedwires said:
forget repairing the original controller. its fully potted, making it impossible to work on. what you need to do is get that controller out of there. there are three pairs of large wires(phase wires) and a five wire ribbon(halls wires) you need to get to them without damaging them so you can solder to them and bring the wires outside the motor. i found the best way is to get the controller right out of the way. boiling will soften the potting enough for you to dig it out to get at the screws that fasten the controller to the inside of the face plate.once the screws are out,you can pry up under the circuit board and break it out of there piece by piece with pliers. as far as wiring diagrams go you're gonna have to wing it. i may have kept some notes somewhere but either way we can walk you through it. i would recommend a lyen 12 fet controller for this motor,and i'd also recommend running at 48-60v

awesome. again, where can I find ribbon?
which Lyen FET? http://lyen.com/ please link
 
I doesn't have to be a ribbon, just any cable with at least 5 conductors. A round cable will probably be less prone to damage. I used a piece of cord off a telephone (6 cond.), and used the modular connector to make the connection to the controller. I have tons of ribbon cables from old computer stuff. You can 'zip' off 5 strands from a wider cable.
 
fechter said:
I doesn't have to be a ribbon, just any cable with at least 5 conductors. A round cable will probably be less prone to damage. I used a piece of cord off a telephone (6 cond.), and used the modular connector to make the connection to the controller. I have tons of ribbon cables from old computer stuff. You can 'zip' off 5 strands from a wider cable.

good ideas!
which Lyen FET? http://lyen.com/ please link
 
Yes, but may require swapping hall / phase wires.

One quirky feature of these motors is the hall sensors are located between the teeth in such a way that they are NOT neutrally timed. This causes compatibility issues with some controllers. With an analog Crystalyte controller, it runs OK, but has quite a bit of advance on the timing. This causes a higher than normal no-load current and probably some efficiency loss. It also makes it blow chunks if you try to use the reverse switch on the controller. It can only be set up to run one direction, but can be set up in either direction.

Apparently there are some controllers specifically matched for BMC motors that compensate for the timing advance. A sensorless controller would do this too.

When looking for a controller, be sure it has been successfully tested on a BMC motor. BMC hub motors have the same winding / hall configuration.
 
i have two of these controllers running bmc motors without issues http://endless-sphere.com/forums/viewtopic.php?f=31&t=17683 but if you're on a budget, i have one i can sell you. its a bmc 36v 20a that i ran on my first bmc build . its a good controller but it has a soft start. it was also made to mount on top of the motor with spacers.
 

Attachments

  • 04141725a.jpg
    04141725a.jpg
    13.2 KB · Views: 2,859
friedwires said:
i have two of these controllers running bmc motors without issues http://endless-sphere.com/forums/viewtopic.php?f=31&t=17683 but if you're on a budget, i have one i can sell you. its a bmc 36v 20a that i ran on my first bmc build . its a good controller but it has a soft start. it was also made to mount on top of the motor with spacers.

I appreciate that... but I would want to get at least 40a/36v-48v controller. so far the cheapest I can find is about $200.

I bought a 600w 24v BMC motor from thesuperkids.com. hooked it up and turned on the switch, and blew my 40a fuse. double checked everything. even tried my 300w bmc successfully. totally bummed. it came with several CAPs installed on the power line into the motor (with shrink wrapping). I didn't expect to see those. I guess it makes sense, incase something burns out.... im guessing the caps are what go first.
 
here is the one I am thinking about getting. I dont know much about these controllers... yet. sounds like you do though! if you know a better product for less price... I am all ears (eyes).
http://www.electricrider.com/crystalyte/x-ct3640d.htm
and this looks about the same... im surprised there are so many styles.
http://www.ebay.com/itm/Crystalyte-36-48v-40a-Controller-e-bike-scooter-IRB4310-/180715899251?pt=LH_DefaultDomain_0&hash=item2a1381c973#ht_500wt_1185
 
friedwires said:
this is your best bet http://endless-sphere.com/forums/viewtopic.php?f=31&t=17683
I saw that but it says
Low voltage cut-off: 62.5v
 
The low voltage cutoff is fairly easy to change. I'm not sure if it can all be done in software, but Lyen can tell you how to do it.
 
ok I bought one of lynn's controllers. I look forward to using it.
then today I just found this... scroll down about half way down the page to where it says "Magic Brushless Dual-Mode Cruise Controller -- for Electric Bicycles, Scooters!" (BAC-281P). looks like it's $65 w/the programming cable.
http://www.goldenmotor.com/frame-controllers.htm

thoughts? opinions? seems too good to be true.

be sure to check out the home page if you wanna see some super bad ass motors http://www.goldenmotor.com
 
ok to follow up. I bought an older burned out 600w MAC/BMC motor (fried internal controller) originally for a synergy/lashout e-bike. I successfully bypassed the internal controller and hooked up a Lyen 12 fet 45amp controller set for 36v(http://endless-sphere.com/forums/viewtopic.php?f=31&t=17683), wired for CCW direction for my scooter (mongoose FS rocket). Thanx Lyen!
Prior to switching to this new setup, I was using the mongoose's original 450w 24v (model:XYG-6B) overvolted to 36v using a currie style 36v/30a controller. I haven't done any efficiency tests yet. so this reply is mostly about speed and power observations between these 2 motors.

Originally on my scooter:
Brushed motor:
24v/20a= 480w (about 15mph) (5mph uphill and sometimes quits)(http://www.monsterscooterparts.com/24v45moupkit.html)
and when over volted:
36v X 30a = 1080w (about 24mph) (7mph uphill and quits less often)
note: at speeds over 2600rpm any throttle action beside FULL confuses to motor and creates bad thumping/knocking noises. if I need to slow down, I have to slow down below 2600rpm to accelerate again. and as long as I do this, no bad noises occur.

New to my scooter:
Brushless 600w MAC/BMC (originally rated 600w w/internal controller... which I bypassed):
24v x 25a = 600w (I haven't rode one of these so dont know how fast it went/goes)
NOW: 36v x 45a = 1620w (presumably) (about 15mph) (about 8mph uphill)

My initial observations:
Im thinking something might be wrong. I haven't done any metering yet. I expected the MAC/BMC motor to perform as well in speed, if not better, than my over volted brushed motor. but sadly it does not. the MAC makes a pretty loud winding sound, especially with more load... and aside from the winding, seems to perform properly (I think?)... just not what I expected at all. infact now that Im writing this, I'm thinking maybe I got the hall sensors wrong? anyone have any experiences like this?

Is there a test I can do to figure out whats going on? or if you have any info that can help me understand why it's not to my expectations, please comment. like old magnets? or inferior version of this motor (not a super MAC)?
I should also mention, when I was testing it, and had it in it's original turning direction (CW) it sounded like it spun about 20% faster....!?!?
 

Attachments

  • 600w mac-bmc.jpg
    600w mac-bmc.jpg
    35.8 KB · Views: 2,795
The BMC motors have an odd hall timing that does not play nicely with some controllers.
It would be important to get some current measurments while running the motor but this would require an ammeter with sufficient range or a large precision shunt. If the timing is the issue, the efficiency will suck and the motor will take much more current than it should for a given power output.

There are some threads burried around that describe this timing issue. One person was able to re-time the hall sensors, which is done my mechanically moving them. Since they are in the slots, the slots had to be ground away slightly on one side to make room. It doesn't take much, maybe half the width of the hall sensor. The inherent timing is quite advanced, so retarding the halls could bring it back to neutral. This means move the hall sensors in the same direction as the rotor will be going.
 
ok I had the wrong phase and hall wire combination. now it works like a champ!!! 27mph on flat and 16mph uphill!!!! totally awesome and worth every penny!
note: I road around with the wrong combination for about 4 blocks. and today when I was testing it I noticed the phase wires on the controller had melted the housing a little (not enough to expose the wires). so I think its safe to say that Lyen controller (12 fet) can handle 10+amps for short periods of time (2-5 minutes) without frying it completely. phew!
 
OK, good.

I'm using an old analog Crystalyte controller. It behaves OK, but I know my no load current is higher than it really should be. It seems at higher loads the efficiency isn't too bad. Adjusting the timing is very difficult on mine since the halls are burried in the bottom.

Ideally, you'd want to vary the hall spacing and take current measurments, and set it at the point where the no load current is at a minimum, or slightly advanced from there.

In theory, it should be possible to use the shunt inside the controller to measure the current. You just need to access both ends of it with wires and send them to a voltmeter. If your controller has a CA plug, these wires exist in the plug. Even if you're happy with the timing, it's nice to have some kind of ammeter.
 
with no load I measured 3.02amps at WOT and 3.5 with the wheel attached but off the ground. good to know about the wires in the CA jack. I have a CA coming in the mail anyway.
thanx again
 
3A @ 36v is probably in line with what I see on mine. Essentially most of this power is wasted. With proper hall timing, you might drop it to around 2A.
 
wow! thats a 33% waist. is that true at higher speeds with load? the hall sensors in these motors are fixed to a circuit board and glued into place with epoxy. and if I were to move a sensor, it would have to move over to the next space between the 18 stator arms... which doesn't really allow for any subtle variations in placement. or is that what you mean? I have 5 more motors I plan to bypass, maybe i'll consider trying to make one more efficient.
 
mrbill did a lot of testing on this and has some threads somewhere.
Do a search on author "mrbill" and read his stuff. Great testing.

Here's one: http://endless-sphere.com/forums/viewtopic.php?f=28&t=22460

I think this is the one I'm thinking of: http://endless-sphere.com/forums/viewtopic.php?f=28&t=20910

He probably ground off more stator than necessary, but you get the idea. The halls only need to move over a little bit, but the only way to to grind the slot wider.

I think the efficiency loss is greatest at no load, so maybe not as bad as it looks.
 
oh yeah I have that page bookmarked. and I too realized that the hall sensors are probably placed by design w/loads in mind. unless someone can definitively tell me how, and what it'll do, and when it'll do it (optimals), then I think I'll just put that on the back burner. :)
 
It may not be worth the effort, but I'm sure you can improve on the stock configuraton.
In general, I think if you moved the halls about 1/2 their width in the direction of rotation by grinding out a notch in the teeth it would be in the ballpark.

When BMC designed the motor, they located the sensors midway between the teeth for ease of construction. I suspect the controllers specifically made for these motors have some kind of timing delay built in to compensate for the sensor alignment.

One possible approach that is easier to try would be to slant the sensors slightly so they face 'downwind' of the rotation direction. Stock, they are flush with the teeth, facing straight out. By canting them, the axis of their magnetic sensors will be tilted, causing a small shift in timing. Perhaps as much as 45 deg might have the right effect. This would still involve breaking them out of the epoxy and lifting the hall sensor board, but with heat it's not impossible. I don't know how much timing change you could get by doing this, but every little bit helps.
 
If the stock configuration isn't optimal, it's probably not just for ease of construction, but also keeping it dual purpose (reversible). how noticeable is the deficiency to single battery charge? maybe add a mile or two to the trip?
 
If the halls were located in the middle of the teeth, it could be neutrally timed and reversible (but harder to make). I don't know how much could be gained by fixing the timing, but I don't think its that much. Probably less than 10%.

Something like Burtie's timing adjuster circuit would be another way to test, and not need any motor mod. Likewise, a decent sensorless controller would sidestep the issue.
 
Back
Top