Version 2 Crystalyte Controller information

[fechter]

Does Shenzhen have the same value of R and C as V2 ?

Yes, 3.3k and .01uf.

 

[The7]

http://uk.youtube.com/watch?v=Icslk6jMyJY
Let me know if this is what you need to see, if so I'll do the rest of the controllers!

Which controller did you use?

I took photos from your video and compare them with Fechter's and mine.
Fechter is using Shenzhen controller.
My photo is for Schwinn AL1020 motor using old version Crystalyte controller

Observations:

1) My motor phase volage is symmetrical at partial throttle.
2) The motor phase voltages of your and Fechter's look similar at partial throttle and are NOT symmetrical.
3) When free-running at throttle off, the motor voltage is symmetrical and is due its back emf.

Comments:
a) AL1020 motor is Y-connected and Puma is delta-connected.
b) These controllers are seemed to be designed for Y-connected motor.
c) When the phase relationship between the Hall signal and motor voltage is matched, the motor voltage will be symmetrical.
d) Suppose the phase relationship between the Hall signal and motor voltage is matched when the motor is Y-connected. Then there will be a mis-match of 30 deg in phase relationship and this would show up in an non-symmetrical voltage wavform. Seems this is case for Puma motor.

 

[Knuckles]

The PUMA is delta-connected? Heck, is a BMC motor delta-connected also?

From ... http://en.wikipedia.org/wiki/Brushless_DC_electric_motor (yes very noob-ish on my part) ...

"There are also two electrical configurations having to do with how the wires from the windings are connected to each other (not their physical shape or location). The delta configuration connects the three windings to each other (series circuits) in a triangle-like circuit, and power is applied at each of the connections. The wye ("Y"-shaped) configuration, sometimes called a star winding, connects all of the windings to a central point (parallel circuits) and power is applied to the remaining end of each winding.

A motor with windings in delta configuration gives low torque at low rpm, but can give higher top rpm. Wye configuration gives high torque at low rpm, but not as high top rpm.

Although efficiency is greatly affected by the motor's construction, the wye winding is normally more efficient. Delta-connected windings can allow high-frequency parasitic electrical currents to circulate entirely within the motor. A wye-connected winding does not contain a closed loop in which parasitic currents can flow, preventing such losses.

From a controller standpoint, the two styles of windings are treated exactly the same, although some less expensive controllers need to read voltage from the common center of the wye winding."

-And just for other "non-oscilloscope owning" noobs like me ... This is my controller on ACID ...

[youtube]s1eNjUgaB-g[/youtube]

 

[Jozzer]

Thats a prototype controller made on behalf of the puma factory for their motor. I think we sent one to Fech a while back too. It was running on 60v..
I also have a Kelly 100A controller, that seemed to work quite well. At least you can choose 60/120 degree halls on the kelly.
I'll post up a vid tomorrow if I get a chance, and an analogue x'lyte for good measure.

Steve

 

[The7]

From ... http://en.wikipedia.org/wiki/Brushless_DC_electric_motor (yes very noob-ish on my part) ...

A motor with windings in delta configuration gives low torque at low rpm, but can give higher top rpm. Wye configuration gives high torque at low rpm, but not as high top rpm.

This description is not quite correct for brushless motor (with 3-phase wnding) because it depends on the applied voltage.

If a certain motor is designed to run at an applied voltage of 415V in wye conection, the same motor will have identical rpm. torque and power characteristics in delta connection when the applied voltage is 415/1.732 = 240V

 

[Knuckles]

Thats a prototype controller made on behalf of the puma factory for their motor. I think we sent one to Fech a while back too. It was running on 60v..
I also have a Kelly 100A controller, that seemed to work quite well. At least you can choose 60/120 degree halls on the kelly.
I'll post up a vid tomorrow if I get a chance, and an analogue x'lyte for good measure.

Steve

The SHENZHEN controller has a jumper for choosing 60/120 degree halls as well. See fechter diagram ...

 

[solarbbq2003]

if the 250watt ( or 350watt as is called) bmc is running ok on v2, and only difference to the 500watt is the coil thickness they might lead in another direction for solution, I'm guessing they are not same though, maybe 250watt is wye connected, lower nunmber of mag poles???

 

[solarbbq2003]

thanks The7 for detailed description of calculating low pass frequency, I'll try some of the suggetions as soon as have next batch of v2 controller ( in the air now)

 

[The7]

if the 250watt ( or 350watt as is called) bmc is running ok on v2, and only difference to the 500watt is the coil thickness they might lead in another direction for solution, I'm guessing they are not same though, maybe 250watt is wye connected, lower nunmber of mag poles???

It is quite possible for a smaller power version to have different no of pole-pairs.

Crystal 5000 has 12 pole-pairs while 4000 has only 8 pole-pair.

Gubee geared hub motor has only 10 pole-pairs (and gear ratio of about 4 or 5 :1)

 

[fechter]

Observations:

1) My motor phase volage is symmetrical at partial throttle.
2) The motor phase voltages of your and Fechter's look similar at partial throttle and are NOT symmetrical.
3) When free-running at throttle off, the motor voltage is symmetrical and is due its back emf.

Comments:
a) AL1020 motor is Y-connected and Puma is delta-connected.
b) These controllers are seemed to be designed for Y-connected motor.
c) When the phase relationship between the Hall signal and motor voltage is matched, the motor voltage will be symmetrical.
d) Suppose the phase relationship between the Hall signal and motor voltage is matched when the motor is Y-connected. Then there will be a mis-match of 30 deg in phase relationship and this would show up in an non-symmetrical voltage wavform. Seems this is case for Puma motor.

Interesting observation about the symmetry. I hadn't noticed that before.

I think the asymmetry is due to the advanced position of the hall sensors, not really due to being delta connected. It would be possible to locate the hall sensors in a Puma so that the timing was neutral and the waveform would become symmetrical.

The amount of timing advance seems to be less than 30 degrees. I have not measured it, but I could calculate it based on the geometry of the magnet poles and the location of the hall sensors.

The timing advance is intentional. Tim Obrien has tested BMC motors that were both neutral and advanced timed, but otherwise identical. He much preferred the advanced version as it had better acceleration.

 

[Knuckles]

I just had a thought. Would a 'pedal first' controller be immune from the 'critical frequency' limitation when running a PUMA or BMC motor?
Or would it suffer the same 'jitter' at a critical rpm?

Heck if a 50-amp 72V controller was made without the need for Hall Sensors (pedal first) and it did not suffer from 'the jitters' then this could be a nice alternative.

 

[rkosiorek]

the only problem with a pedal first controller and the puma or other geared hub motor is that in order to start the pedal first combo you need to get the motor rotor turning. since the rotor is isolated from the outside hub and wheel/tire by either a freewheel or one way roller clutch ytou could never get it started.

what you would need is a combination of hall sensors used to start the motor and to take advantage at low rpm of the advanced timing for good acceleration and then at some frequency switch to a sensorless design for high speed.

rick

 

[docnjoj]

Yeah Rick! I didnt think of that! Way to go!
otherDoc

 

[Knuckles]

the only problem with a pedal first controller and the puma or other geared hub motor is that in order to start the pedal first combo you need to get the motor rotor turning. since the rotor is isolated from the outside hub and wheel/tire by either a freewheel or one way roller clutch ytou could never get it started.

what you would need is a combination of hall sensors used to start the motor and to take advantage at low rpm of the advanced timing for good acceleration and then at some frequency switch to a sensorless design for high speed.

rick

Yes of course. You are correct. This was discussed in other threads earlier. If this was constructed ... Do you think it would work around the 'critical frequency' jitter issue or do we still have to address this with a software code fix?

 

[fechter]

Yes of course. You are correct. This was discussed in other threads earlier. If this was constructed ... Do you think it would work around the 'critical frequency' jitter issue or do we still have to address this with a software code fix?

I don't think it will change the jitter problem.

 

[The7]

I think the asymmetry is due to the advanced position of the hall sensors, not really due to being delta connected. It would be possible to locate the hall sensors in a Puma so that the timing was neutral and the waveform would become symmetrical.

Would concur with you if the Hall sensors are placed for delta connection.

IThe amount of timing advance seems to be less than 30 degrees. I have not measured it, but I could calculate it based on the geometry of the magnet poles and the location of the hall sensors. .

From your photo, it is estimated to be about 22 deg.
Great if you could determine it from the geometry !

The timing advance is intentional. Tim Obrien has tested BMC motors that were both neutral and advanced timed, but otherwise identical. He much preferred the advanced version as it had better acceleration.

Would concur with Tim Obrien about his finding.

If viewing the BLDC motor as an 3-phase synchronous motor operating at unit power factor, the timing advance at light load is very small and will be large at heavy load.
Since the timing advance is constant for simple BLDC controller, some compromised value of fixed timing advance is used.

If It is compromised at heavy load, the motor will have a higher power output than the other compromised at light load. At the same heavy load, it will run a bit cooler. But it will run relatively hotter at light load when compared with the other.
Of course it is always hotter to run at heavy load than at light load.

 

[The7]

Some further comment on delta-connection:

From my engineering understanding:
a) If the voltage applied to the motor winding is purely sinusoidal, there is no difference in performance between the delta-connection and wye-connection when the correct value of voltage is applied.

b) If the voltage applied is not sinusoidal, there will be more harmonic currents flowing in the delta-connected winding. These harmonic currents will reduce the usful torque and increase the motor loss. The motor will run hotter.

c) Most designers prefer wye-connection for the reason as in (b).


Is there any good reason why Puma chooses delta-connection?

 

[rkosiorek]

i think that it is an RPM thing. for a given voltage an otherwise identical motor will run at a higher RPM with the windings connected as a DELTA and lower RPM when connected as a WYE (star). the RPM is at the expense of torque though.

the power produced by either winding configuration will be the same though. power is analogous to RPM X Torque. so the Delta with higher RPM and lower torque would produce the same power as the WYE would using lower RPM but higher torque.

i would guess that in the PUMA/BMC they needed the Higher RPM to turn the wheel at an acceptable speed after the gear reduction and the gear reduction would more than compensate for lack of torque.

rick

 

[The7]

i think that it is an RPM thing. for a given voltage an otherwise identical motor will run at a higher RPM with the windings connected as a DELTA and lower RPM when connected as a WYE (star). the RPM is at the expense of torque though.

the power produced by either winding configuration will be the same though. power is analogous to RPM X Torque. so the Delta with higher RPM and lower torque would produce the same power as the WYE would using lower RPM but higher torque.

i would guess that in the PUMA/BMC they needed the Higher RPM to turn the wheel at an acceptable speed after the gear reduction and the gear reduction would more than compensate for lack of torque.

rick

Please note:

If a certain voltage is applied to a motor in delta-conection, it will have a certian rpm, torque and power charecterestics.

When such motor is connected in WYE, it will have the same rpm, torque and power charecteristics if that applied voltage is increased to 1.732 times.

 

[rkosiorek]

So for any given voltage the Delta will run faster than the WYE. so for an ebikes the design is limited by the Voltage of the battery pack. a DELTA wired motor will run faster on 24V or 36V than one wirred as a WYE. With the geared hub design the designers were not worried about the torque. they were worried about RPM, so I think that for the PUMA/BMC the designers chose the Delta configuration looking for that higher RPM.

having used a BMC geared hub at 24V and 36V i can say that it is definitely not a speed demon.

Rick

 

[The7]

So for any given voltage the Delta will run faster than the WYE. so for an ebikes the design is limited by the Voltage of the battery pack. a DELTA wired motor will run faster on 24V or 36V than one wirred as a WYE.
Rick

Agreed if the motor uses the same set of winding for delta and wye connection.

But, the motor could be re-wound with thicker wire and fewer turns per phase for running in WYE so that the same battery voltage would give the same rpm. torque and power.

 

[rkosiorek]

the question would be "IS IT WORTH THE EFFORT" to rewind one motor for yourself. think of it. cut off the old wiring, strip all traces of the shellac and paper insulators from the stator teeth, cut new insulators from paper, rewind all of those coils, re-dip the whole mess in shellac to fix the coils in place, re-mount the hall sensors.

that is a lot of work. especially the stripping/cleaning and rewinding parts

i have rewound Crystalyte 40XX and X5 motors from desperation or to test new ideas but it is not something that it is economical to do. at currewnt prices for copper rewinding a 40XX is about $45CDN and the X5 about $80CDN. add that to the time required and you do it either as a labor of love or insatiable curiosity.

i think that the most i would be willing to do on the PUMA/BMC is to rewire the stock winding as a WYE and modify the hall sensor pc board so that the timing could be adjusted. and then see if it will still turn a useful RPM. I know that the torque and acceleration would be improved. the only question is would it still propel the bike at a satisfactory speed using reasonable battery voltages.

rick

 

[The7]

the question would be "IS IT WORTH THE EFFORT" to rewind one motor for yourself. think of it. cut off the old wiring, strip all traces of the shellac and paper insulators from the stator teeth, cut new insulators from paper, rewind all of those coils, re-dip the whole mess in shellac to fix the coils in place, re-mount the hall sensors.
rick

If the motor is in burnt out, I think it is worth the effect.

But if an battery with higher voltage is at hand, it may be worthwhile only to reconnect the delta into wye and give it a test.
The original delta-connection would give a 20 deg advance in applied voltage wrt motor BEMF.
The re-connected wye would give a 10 deg advance in applied voltage wrt motor BEMF.
Would the 10 deg advance because it is closer to the ideal case.

The ideal case is when the applied volage is in phase with the motor BEMF.
Most motors with wye-connection could nearly achieve the ideal condition.
( From construction, Crystalyle 4000 has the applied voltge in phase with the motor BEMF)
( From viewing the waveform, Schwinn AL1020 has the applied voltage nearly in phase with the motor BEMF)

 

[monster]

i have a shorted FET on my 12 FET v2 controller. do i just need to replace the 4 FET's on the phase that has shorted or the whole lot? is there any thing else that will have to be replaced like the gate driver etc?

 

[solarbbq2003]

regen braking v2 controller:
abs pin to ground
select 60 or 120 degree:
12/6 pin to ground ( pin located just left of r26) gives 60degree