low speed control

[Yahoo1]

I am using AC brushless servo motor in my motion control system. From the motor's spec, the speed can goes from 0 to 4000RPM. I am wondering how low speed of the motor I can get in my system. If I want 0.01RPM for the motor, what pre-requisite is needed, resolution of encoder, mechanical accuracy, control tunning?
How high encoder resolution is needed at least for this low speed?

Any opinion is very appreciated.

Thanks

 

[Barry1961]

I would guess at that the big question would be how accurate you need to be at this low RPM. When a servo is at zero speed it usally oscillates around it's calculated "perfect" zero. To lower your following error while doing a low speed move you would probably want much higher PD gains. Usally the servo is much more accurate the the mechanicals.

Do you have a application or just wondering?

Barry1961

 

[Yahoo1]

I have an application but I am wondering how I can implement it.

 

[sreid]

Running at low RPM can be a real challenge. A high resolution encoder is essential (the more the better). This may limit the top RPM. A partial solution is to use sin-cos encoders and do flash interpolation of the encoder signal.

Other limits must be considered. If the control has a 16 bit DAC (15 bit resolution +/-), the speed resolution for a 4000 RPM top speed would be 4000/2^15 = 0.122 RPM.

Tell us more about what you want to do and we can make better recomendations.

 

[Yahoo1]

Thanks for all of your opinions.

I am using a AC brushless servo motor to control pump's injection speed. Sometime we need it run at very low speed.
The motor we are using has resolver with 1 speed and 500ppr. The driver converts it to quadrature incremental encoder signal, so the controller can see 4096ppr.

sreid, which top speed are you refer to, from spec or real max speed in the application?
The motor's speed at bus voltage is 4000rpm, but use it up to 2000rpm. And the resolution of encoder feedback is 12bits on driver, which is Baldor SD26M and controller, which is NI's PCI7342. From your post, it is absolutely not possible to get 0.1rpm for my system. The limits is from the resolution of the encoder feedback. Am I right?

Do I describe my situation clear enough?

Thanks.

 

[sreid]

You are correct, the feedback resolution is usually the first thing that limits low RPM performance. If there is not at least two encoder edges per servo update time, there is no stable velocity estimation. Also as the edges get farther appart, there is a latency in the velocity estimation that introduces a delay in the servo loop which limits the servo gains that can be used.

The above is not an absolute but points to the problems that will be encountered.

 

[Skogsgurra]

Hi Yahoo1,

Do you really need to go all the way up to 4000 RPM? I guess not.

First; define your speed range - say 0,01 - 10 RPM (which I think is adequate in an injection pump application, correct me if I have misunderstood your requirements).

Second; select drive technology - I think that a stepper would suit you better than an expensive servo motor with encoder or resolver.

Third; calculate what gear-box you need between motor and pump. You will probably need to gear down something like 100:1, which implies a worm gear or a two-stage gear.

Four; enjoy sucessful operation.

 

[Yahoo1]

Hi skogsgurra,

Thanks for your points.
First, the stepper motor is not suitable for my application. The pump has speed control and pressure control function.
Actually, we used a 1:200 gear box between motor and pump. But sometime we still need to run the motor at about 0.1rpm. Highest speed is about 2000rpm.

 

[sreid]

If the controller will let you set 0.1 RPM, then the motor will run at an average speed of 0.1 RPM. It might actually start and stop but you probably won't notice it at the output of a 1200:1 gearbox.

 

[TwnB]

How is the gearbox connected in the system? You say it is a 1:200 gear ratio. Does this mean for every 1 RPM of the motor the output shaft of the gear box rotates 200 times? SO at 0.1 motor RPM the output of the gearbox is 20 RPM. If so then maybe it is possible to put the encoder on the output of the gearbox. The only problem you would have then would be figuring for the backlash of the gearbox which could be compensated for by having another sensor on the motor that would give you an input to calculate backlash.

If the gearbox is connected the other way around then the output of the gearbox at 0.1 motor RPM would be 0.0005 RPM which makes no sense to me but then again I don't know what your application is.

Your other option would be a larger gear ratio.

 

[Yahoo1]

The gearbox is a speed reducer.

The PID loop time of the controller is 62.5us. If the lowest speed is 0.1rpm, what encoder resulotion should be used for this application.

I am trying to find a way from controller side if it is possible because the gearbox is there already and we have some space limit.

Now I am using 1 speed resolver and 1024ppr, 12bits resolution of encoder signal on amplifier. If I change the resolver to 4 speed and the amplifier with 4096ppr, 16bits resolution encoder, how about the cost?

Thanks.

Thanks.

 

[blacksea]

One more question.
Why you don't increase a resolver resolution by Baldor driver? Most of them have such opportunity - up to 4096 PPR maybe reached.

 

[EdDanzer]

You should choose the encoder count by speed limitations of the control for maximum resolution. Most controls have a maximum encoder frequency, divide this by your maximum RPM to find the maximum encoder count.

 

[jbartos]

Suggestion: The motion controls is often designed by software, since there are various parameters and variables involved and certain type of hardware available on the market and feasible for the project. To design the control system by an empirical approach does not necessarily lead to the best solution or to a solution at all.