Re: [CAD_CAM_EDM_DRO] Re: Suggested velcotity params for servos?

At 05:08 PM 12/20/2004, you wrote:

>--- In CAD_CAM_EDM_DRO@yahoogroups.com, "AbbyKatt" <cnc@a...> wrote:
> > Can anyone recommend velocity params for servos? I've got the big-
>big
> > servo motors from Dan Mauch's site (I can't remember which ones they
> > were, but they cost $199/each) and I'm running these with geckos at
>60v
> > in a 1/3 timing belt setup on my mill/drill. I'm running with
>TurboCNC
> >
> > My max reliable (without faulting) speed is 35Khz
> > That's pretty easy to guage.. But I was wondering about the starting
> > frequency and acceleration?
> > I know steppers can have a high start frequency, since they
>effectively
> > "hop" to the first position with a high torque, but what about
>servos?
> >
> > I'm really not sure here.. Any good way to derermine?
> > I'm worried that I'll be accelerating too fast, or starting too
>fast and
> > sacrificing accuracy as the servo lags behind just far enough not to
> > affect the signed 8-bit lock (128+- count) of the servo. But at the
>same
> > time, I don't want to enter 0 as the start rate if I can do better.
> >
> > Any suggestions? Good ways to check (without a DSO?)
> > Thanks!
> > ~Abby
>
>Setting up your parameters is a very important step in ensuring
>operation without step loss.
>It is however a lengthy process covering several days of retesting.
>
>Start out with your Max Pulse rates.
>Set your Acceleration to a very conservative setting when testing.
>A very noticable Acceleration setting that will ensure the motor
>wont be inhibited by this maybe 2-4 seconds to reach top speed should
>be fine.
>Program long enough moves for testing so top speed can be reached.
>Now increase the Max Pulse rate for the axses untill you get into a
>stall condition. Reduce the setting about 10 percent.
>Repeat with all axes and keep notes of the settings you come up with.
>
>After the Max Pulse rates for all Axses are established test with all
>axes moving together at max rates.
>May add some weight to the table to simulate a load.
>Note when 2 Axses of same resolution are moving a 45 degree angle
>may be programed to ensure both run at Max capacity.
>You may need to ajust your Max Pulse rate down again for this test
>because your Powersupply will likely not be able to provide the same
>amount of current when all motors are running.
>The last phase of this test has to happen first thing in the morning.
>You want to repeat this test after the machine has cooled of and the
>oil in the slides has thickened overnight because this will add to the
>resistance of the slides to move.
>If your machine passes the test your parameters should be fine else
>reduce and try again the next day.
>
>Now you can begin setting Acceleration deceleration.
>Basicly the same sequence as before just using the acceleration
>Parameters.
>Few people like to run there machines close to the maximum
>acceleration settings posible a stiff machine puts a lot of
>wear and tear on all components.
>The settings you will come up with are just pices of info what is
>possible.
>When a axes locks its Error register has overflowed +-128 pulses it
>would be good advise to Lower the setting significantly to get a
>lesser following error.
>There is a great deal of personal preference where you want your
>final settings to be.
>Most like to see a noticable Acceleration deceleration lasting for
>maybe a second or so, this makes for smuther transitions and causes
>less wear.
>Lower settings also make a system less sensitive to increased masses
>of tools or parts.
>A properly setup machine will be much less likely to lose steps even
>under the least favorable conditions, contrary bad setting can cause
>problems under some conditions that will be hard to pinpoint and
>toubleshoot.
>Good Luck

I am going to jump in here with an addition to what Torsten has said.
That deals with the START velocity of your servos. I originally set mine up
with the maximum start that I could get reliably.
Big servos can tolerate a big start velocity (ask me how I know:}); but
much to my horror, I had a project that required the use of 1/8" end mills
and every time I started a cut, it would break the mill off! I finally
figured out that my start velocity was to great and it was actually jumping
and then slowing down. Also, buy slowing down the start speed, I was able
to increase acceleration.

Just food for thought.




Bubba

OLDER THAN DIRT
Country Bubba
(Actually the inventor of Country and Bubba)
In God We Trust

LaGrange, GA

http://bellsouthpwp.net/A/r/Arts_home_page/cnc/

[Non-text portions of this message have been removed]