On Sun, 21 Oct 2007, Jon Elson wrote:
> Date: Sun, 21 Oct 2007 13:12:57 -0500
> From: Jon Elson <[EMAIL PROTECTED]>
> Reply-To: "Enhanced Machine Controller (EMC)"
> <[email protected]>
> To: "Enhanced Machine Controller (EMC)" <[email protected]>
> Subject: Re: [Emc-users] DC Motor and PID
>
> John Kasunich wrote:
>> Jon Elson wrote:
>>
>>> Kirk Wallace wrote:
>>>
>>>> I connected the motor to my lab supply and got the motor to turn very
>>>> slowly at about 4.5 Volts. At the slowest speed, it would only stop if I
>>>> reduced the voltage. So I guess I have a "good" motor and I just need to
>>>> work more on the tuning. Maybe, add more P to get the initial voltage up
>>>> to 4.5 but then add more D (?) to fight the oscillating. Well, I wanted
>>>> a real world test bed, I guess I got it.
>>>>
>>>
>>> Right, you want P as high as possible, and to the point that you
>>> need some D to prevent oscillation. Unfortunately on the Pico
>>> Systems PWM system, there is no current loop or tach feedback,
>>> so the tuning is a little bit more touchy than with velocity
>>> servo amps. But, usually you can get the following error down
>>> to really negligable levels and still have a stable servo
>>> response. Too much D and the whole system gets quite unstable
>>> due to lags in the loop and the effects of quantization of the
>>> encoder. Once you have P and D up about as high as you can get
>>> them, you reduce the rest of the errors with FF1 and FF2. It
>>> doesn't take much of these factors to make quite a difference,
>>> and it is easy to go too far and make things worse. And, you
>>> need something like 1 - 2 encoder counts worth of deadband to
>>> stop the buzzing.
>>>
>>
>>
>> Don't you use _any_ Integral gain?
>>
>> I guess there are as many ways to tune as there are people, but I can't
>> imagine not using I gain. In fact the drives that I work on in my day
>> job tend to use P and I only. No D, no FF. Of course these aren't
>> servo drives, just simple speed loops for the most part.
> Yes, I do, but it doesn't seem to do a whole lot. I usually
> have numbers between 1 and 5 there. But, I find the FF is a LOT
> more sensitive, and I can get the error down to practically
> nothing, 100 uInch or .001-.002 mm. I really doesn't seem to
> help much with the PWM system, maybe because it is not a true
> velocity servo. My usualy tuning protocol is to make a 3/4
> second move at various jog speeds, observe the error, tweak a
> parameter and repeat. My big problem is the unavoidable
> quantization noise of the encoder count being differentiated and
> turned into a huge signal with a lot of energy at
> 1/servo_period. I haven't had the time to work with it much
> more, but it looks like it needs some kind of fix to work well
> with my PWM servo system. I guess there is a lag in the system,
> most likely a combination of filter inductance and motor
> inertia, that makes it wildly unstable if you add too much D.
>
> Jon
>
Velocity feed forward is especially important with straight PWM ampilifiers as
it allows the system to approximate current control, that is, make the drive
torque independent of motor back EMF. It also has one advantage over current
control systems in that the inherent damping factor is higher (the amplifier
has a low output impedance)
One way to allow higher damping (and hence higher P) is to raise the servo
sample rate. For small or high performance motors, 1KHz is _way_ too slow, the
zero-order hold effect (because the velocity calculation from the previous
cycle is applied over the current cycle) introduces a very unwanted delay in
the D part of the feedback loop. This delay is inversely proportional to
sample rate.
Peter Wallace
Mesa Electronics
(\__/)
(='.'=) This is Bunny. Copy and paste bunny into your
(")_(") signature to help him gain world domination.
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users
