Closed and Open loop current control for Brushless Servo Motors 3

[FeX32]

Hi guys,

I have been having an issue finding and specing a drive/amplifier for a motor (

http://www.baldor.com/catalog/BSM90N-275AA)

Closest I have is this:

http://www.kollmorgen.com/hi-in/products/drives/servo/s700/s700-servo-drive/

(to which I'm not sure will work)

I want to be able to run it in torque/current mode (with internal fast PI loop). But, I also want it to act like an amplifier and not have an internal current loop active.
That is, I would like to be able to have a mode where I can supply a voltage via my controller and have the Amp feedback the current, and then my controller reacts externally.

Are there any motor drives that are high performance in their built in current loops (via analog in) and also with the ability to run open loop current control?

Thanks for the help,

Fe

 

[BrianE22]

Machine tool drives come to mind. Fanuc for one.

 

[mikekilroy]

Sure. We supply the drive you want everyday. you are on right track.
Can give you full tech details if you want. It is called akd-b-01206-nban

http://www.KilroyWasHere<dot>com

 

[FeX32]

Hi Mike,

And this version of the drive has open loop current loop capabilities? That is, I apply an input and it supplies this voltage or a voltage*gain to the motor?

Thanks,

Fe

 

[mikekilroy]

Sure.

Why not download the various manuals and the software (workbench) and you can learn more.

But I am having a real hard time understanding any reason one would want a sophisticated brushless motor in total open loop voltage control. Seems to me if you want crude, you just buy an induction motor and cheapest vfd you can find.

Are you aware that such control is very hard to even make a synchronous motor like this turn? Change the load maybe .01% and it will drop out of synchronization immediately! If you run the motor with nothing hooked to its shaft, you probably can set a 60 sec or longer accel ramp to get it stay in synch to get to speed, but as soon as you attach that shaft to any real world device with friction, stiction, torque perturbations, it will be very hard for you to not slip out of synch and stall. Try it if you question this: I assume you already have that Baldor motor, stick it on a cheap vfd if you have one around and try to run it and vary the speed. I would bet with any load attached you will not make it go. I have had, about 60% of the time, luck making open loop go like this, but it never amounted to anything that could do any useful work of course.

Please share WHY you possibly would want a high performance servo motor and then propose to run it in such a sloppy mode?

Anyway, for technical answer: open loop volts_in = volts_out * K is another way to say v/hz (aka skalar) mode. Go to motor tab in workbench, pick motor type 2 (rotary, v/f open loop) and you will get simple power supply with integral motor commutation and freq setting to match with no control.

You want also torque mode with "fast PI loop;" fast is a very relative term - what is fast to you?

http://www.KilroyWasHere<dot>com

 

[FeX32]

Thanks Mike.

I will have at hand a Real-Time controller that can run at 10-20 kHz in closed loop. This is what I will be using as a control. (so about 50us)

I am investigating 2 options. The reason I have 2 is I am concerned with the 4kHz bandwidth of the Analog in and out of the AKD (or any other) drives.
If I send a setpoint that updates every 50us to the current PI controller of the drive it will only update every 250us because of the analog in. And the same goes for the analog out.
I presume running it in open loop mode does not get around this 4000Hz bandwidth.

This is why I am investigating an Amp. thereby possibly I could get better resolution of control using my RT box directly? Maybe something like this(

http://www.aerotech.com/product-catalog/drives-and-drive-racks/bl-series.aspx)

but one with more power. I'm not sure though.

 

[mikekilroy]

I see where you are coming from now. Certainly a non digital all analog drive would eliminate all update time issues you may have. I wonder if that one you linked to in last post doesn't also have update time limits too? It does have total isolation between the current loop and output IGBTs - MIGHT be digital; worth checking. Also it does not state open loop control so check that of course.

But at the end of the day, I still wonder if YOU taking the current loop control is going to really gain you anything... Sure your loop is 10-20khz, but as you say, BOTH the analog input AND the analog current monitor output are SLOW... Drive PWM also is something like 8khz... The drives internal update rate of 0.67usec (1.5mhz)is way faster than your external loop, making it effectively analog anyway. If you could control the Iloop, you can adjust the PI to 5khz also; but my experience has been you probably won't be able to without oscillating on the type motor you listed.

This drive shows an in and out @ about 8khz BW for 3db points; check out the 0-10khz LPF descriptions on both (AIN.CUTOFF, etc)...

Then comes real world movement of the motor shaft itself. Again, my experience suggests to me that even if you had a 20khz Iloop, it would be questionable if that could manifest itself in motion on the output shaft of that large inertia motor.

What exactly is it you want to move that quickly?

http://www.KilroyWasHere<dot>com

 

[FeX32]

It does have update limits. They state 2kHz update rate which is far too low.
I like your idea of a fully analog Amp. Can you recommend one? Will they come with commutation or will I have to output this from my controller?
I need to control the current loop because I am investigating adaptive control methods at very high speeds. I thought that maybe I could “cascade” the control where I can use the internal PI on the drive then use my adaptive controller to modulate the setpoint. But then this 4kHz concerned me.
I will not need to drive inertia. I am only requiring to control the motors torque at speeds (3000rpm). This speed will be accomplished via another motor.
Thanks,

 

[mikekilroy]

did you checkout deltatau.com? they might have something fast enough with their raw drives?

Not sure where you saw 2khz update rate? That cannot be right...

I wonder how kol can state 8khz BW with 2khz update rate?? And so why offer 0-10khz LP filters on these sigs? I will ask because something doesn't make sense. will report back what I find.

http://www.KilroyWasHere<dot>com

 

[mikekilroy]

checked and as I thought:

"The latest AKD Installation manual on the website ( Revision U ) has added the specifications you are looking for.
I don’t believe they were published in previous revisions:"

Analog input: Firmware update rate: 16khz
Analog output: Firmware update rate: 4khz

So no issue on IN but 4khz out may be slow for you. but then you could just use realtime Hall sensors and gather that yourself anyway for true analog response.

http://www.KilroyWasHere<dot>com

 

[FeX32]

Thanks Mike.
I am not familiar with this Firmware update rate? How does it compare to bandwidth of the Ain or Aout?

 

[mikekilroy]

You are familiar with aliasing.... that is what this is about. There will be two BW or update times of significance to your project in this digital world: update rate and actual a/d BW.

I believe we will find the BW is >8khz on this product for both in and out. So an update rate of 16khz on the input is great in that it means you get two updates per possible significant BW change. It would also mean the output current feedback signal, at over 8khz BW, will only be seen every 2nd update! Nasty. I do not believe it, and I called the engineer on that. He has not responded yet - I suspect they will find an error in their math. But that should not be a show stopper for you, since you can buy a couple $1.00 (your in the USA?) hall donut sensors and gather the actual current real time yourself.

The only question that I see then remaining for this product is then if the 8khz update analog input rate is sufficient for your needs. If not, back to the drawing boards!

Might I ask WHY you are trying to do this, in more detail? Assuming it is not a patentable project, please share why you are trying to add some adaptive control algorithm to a servo motor current loop. You say for "high speeds." Current loop response is a function of motor L and supply V.... nothing to do with velocity.... I am overly excited to learn why you are doing this.

I cannot comprehend why adaptive current control is required at high velocities: velocities are controlled, in generic terms, around 100hz, current in 1000 hz; apples and oranges. I am very interested in what you are doing. Can you share?

http://www.KilroyWasHere<dot>com

 

[FeX32]

Thanks for the info Mike.

hmm.
I understand aliasing, but I am still a little fuzzy on how the BW can be >8kHz when the A/D converter runs at 4? Can you explain please?

I would be okish with 8kHz since it is double what I originally thought with these drives. What about the Aout? Can we confirm it is 8kHz too?
What about an Analog Drive? Are there any on the market? Would this solve my issue?

Regarding what I am doing. This is tricky, I cannot really reveal. But I will say that it has nothing to do with velocity. The current/torque needs to be controlled only. Any speed the motor sees will be due to the system it is attached to.
What is important to me is the high frequency torque profile that the motor shaft sees. It will be adapting (not extremely fast mind you) to some external sensors.
The reason I need fast update rates is to better approximate my torque profile I need.
For example, a 250Hz sinus wave torque... much better approximated at 10kHz than 4kHz. Also, 250Hz with 4kHz BW leaves 3.5 deg. of inherent phase lag.
The other thing is the max voltage. If the sample loop is too slow, the error in current will be too high and the required voltage to compensate for it will go above the voltage of the bus (320VDC)

Thoughts?

 

[mikekilroy]

Good. then WE are on the same page! I have not replied yet as I have not gotten the design engineers of that product to understand the ludicrousnes of giving an update rate lower than the BW! All I can say for certainty at this time is the ana input update rate is 16khz, and I ASSUME the a/d BW is >8khz. It appears, unless there is a mistake, the ana output is an issue for you since, although no limit in a/d BW of >8khz, it is only updated at 4khz rate. It would be useful for you I think if I could email you some pictures. Is that possible?

I know of no analog drive or I would have shared that, sorry.

Thanks for sharing more what you are doing. Would a drive with higher bus voltage eliminate some of the concern? Would a lower inertia motor help? Would an ironless rotor motor help? Would lower motor L help?

http://www.KilroyWasHere<dot>com

 

[FeX32]

Thanks Mike for looking into this.
I'd for sure like clarification on the BWs.

I'm not sure about increasing the bus voltage. I'll do some more simulations to try and see if it might help. Inertia would not really make a difference for this. Inductance would make it easier on the controller since the open loop response would already be faster.
The other thing is the 8kHz PWM. I wonder if we can increase this?

 

[FeX32]

Also I have some thoughts on motor drive combos:

Is there any disadvantage of using the AKD with the Goldline B 404 ? as opposed to using the S700 with the Goldline motor?

I want to compare AKD drive with AKM motor vs. AKD drive and Goldline motor
and these to S700 drive and Goldline motor.

I guess there must be some sort of differences, but I am unsure.
Thanks!

 

[sreid]

Just a thought,

How about using a linear servo amplifier? Varedan ( varedan.com ). They manufacture Analog Linear Servo Amplifies that accept two current command inputs (for two motor phases). The current loops can be set for 10 kHz bandwidths. The motor is commutated from the controller. Give them a call.

 

[mikekilroy]

Sreid, thank you for the source! Fex needs to exlore analog options. I will star you for the source suggestion!

Fex, you are asking questions about specific products that appears to me to go beyond SELLING which is not allowed on this forum, since it is for engineering ideas. I am not comfortable discussing your last post here. I do not want to sell you anything, but I can answer your specific questions. I feel to do so requires going off site. I am willing to post my email address if you wish to continue thoat specific info: mike at KilroyWasHere dot com.

http://www.KilroyWasHere<dot>com

 

[cswilson]

sreid:

I was thinking of mentioning Varedan (and also Trust Automation), both of which build good linear amplifiers for brushless motors that we have used successfully. The problem for the OP with these or comparable amps is that they have analog current loops (RC networks around op-amps) built in, and there is no way that I know of to disable them for the voltage mode that the OP wants.

At least with fully digital amps, there is the possibility of effectively disabling the digital current loop by setting certain gains to zero.

Curt Wilson
Omron Delta Tau

 

[mikekilroy]

Curt I mentioned Delta Tau PWM input drives... do you think they have anything to offer fex?

http://www.KilroyWasHere<dot>com