Dynomotion

Group: DynoMotion Message: 11503 From: lmp582002 Date: 5/9/2015
Subject: BDC Motor

Greetings,


I am considering using the SnapAMP to run a very high speed spindle motor (40,000 RPM Target) and have a few questions.


1. Will the SnapAmp have any trouble controlling a motor at such a high speed?


2. Can the SnapAmp run with a Y configuration? I believe this would allow me to use a higher voltage power supply - 730KV see below.


3. How is the SnapAmp configured to run a BDC motor without encoder feedback?



Specs:
RPM/v: 730kv @ "Y" config. / 1280kv @ "Δ" config.
Resistance: 0.0132ohm @ "Y" 730kv / 0.0046 @ "Δ" 1280kv
No Load Current: 3.2A @ "Y" 730kv / 4.2 @ "Δ" 1280kv
Can Diameter: 56mm (actual motor diameter) 
Can Diameter inc. Water Jacket: 63mm
Length: 102mm
Shaft Size: 8mm
Weight: 971g


Thanks!


Scott
Group: DynoMotion Message: 11505 From: Tom Kerekes Date: 5/9/2015
Subject: Re: BDC Motor
Hi Scott,

Regarding:

#1 - probably too high of a speed.  KFLOP+SnapAmp updates the commutation only every 90us.  At 40,000 RPM that would be:

(you didn't state the number of poles - I'll assume 4 poles)
40,000RPM / 60 sec/min = 667 rps
1/667rps = 1.5ms/rev
2 cycles for a 4 pole motor - 750us/cycle
Only 8 commutation updates / cycle (every 43 degrees of phase)

It might work somewhat with just the right commutation but certainly not optimally.


#2 - I believe it should work in either configuration


#3 - you will need to determine the rotor angle somehow and set the commutation Angle in a User Program.  See the HallCommutate.c  example.  Although in your case a UserCallback should be used to update the angle every 90us whereas a normal User Program only executes every 180us.

I don't see an inductance spec.

If you decide to try it please post your results.

Regards
TK


Group: DynoMotion Message: 11506 From: lmp582002 Date: 5/9/2015
Subject: Re: BDC Motor
Tom,

You are right it is a 4 pole motor. What would you consider to be a minimum for commutation updates per cycle to have a motor function well? I am also considering a 8 pole motor running at 12,000 RPM which would have 13 updates.

If I use 2 KSteps (running multiple stepper motors) and a snapamp to run the spindle will this have any effect on how often the CPU can update each cycle of the motor?

Regarding point 3, I am curious how the rotor angle is normally determined? Can back EMF from the motor be used? The snapAMP home page says the following. I understood this to mean back EMF, but maybe the standard options are hall sensors or encoders.

"We designed the SnapAmp to easily connect to the KFLOP with a single ribbon cable, enabling KFLOP to perform 3-phase brushless sinusoidal commutation for brushless motors without the need for hall sensors - though hall sensors can also be used."

Thanks,

Scott
Group: DynoMotion Message: 11507 From: Tom Kerekes Date: 5/9/2015
Subject: Re: BDC Motor
Hi Scott,

I'm not sure what the minimum number of updates would be.  To do nice sinusoidal commutation many would be required.  The Torque Constant should have a sine wave function.  Perfect commutation would always be at the peak of the sine wave and theoretically provide 100% Torque.  As the rotor moves it should drop off as the number of degrees from the peak.  Normal Hall sensors only transition every 60 degrees (+/-30 degree range).  At 30 degrees torque should still be at 87%.   There will always be other factors that reduce the efficiency.

SnapAmp was designed with basic CNC Servo motors in mind that are typically limited to 4000 RPM or less.

KFLOP services up to 8 Axes every 90us regardless of the number of axes in operation.  But for more of the Axis in operation less of the DSP time is leftover for a User Call back and the running Threads.  This description may help:

For SnapAmp to commutate without Hall sensors the assumption was that there is an encoder input.  SnapAmp doesn't have a means of sensor-less commutation.

Regards
TK

Group: DynoMotion Message: 11510 From: lmp582002 Date: 5/11/2015
Subject: Re: BDC Motor
Tom,

That makes sense. I am wondering if it still might work relatively well. In thinking about how a brushed DC motor works it is a complete square wave driven from the brushes? Do you think it would function in efficiency similar to a brushed motor?

Thanks,

Scott