RE: [CAD_CAM_EDM_DRO] gecko servo drive +5tpi screw?

The story with the Gecko drives servo units is a little different. Most
stepper programs (what you need with the G320) were not designed with the
step rates and feeds that the servos are capable of. Luckily if you keep the
encoders within reason all works great. To figure your step rate start with
the line on the encoder disk. Multiply that times 4 to get the effective
steps per revolution (4000 with 1000 line encoders). Now multiply by the
number of motor revolutions per inch (5 in your case if you have no belt
reduction, most commonly 10 as a 2:1 belt reduction is pretty normal). This
would give you 20000 or 40000 steps per inch or a mathematical resolution of
.00005" or .000025". If you want 100 ipm (not unrealistic as a rapid rate)
you need 2,000,000 or 4,000,000 steps per minute or 33,333/66,666 steps per
second. Multiply this X 3 axis's and you are looking for the poor PC to do
some pretty amazing things.

Now, Mariss is going to be releasing a new model in 10 days or so that will
have a pulse multiplier built in that can be set for anything from 1 X (no
multiplication) to 10X so that in the above example the PC would only have
to do 3,333/6,666 steps per second. That will make high line count encoders
usable. On the G320 I would stick with an encoder that gives you 10,000 to
15,000 steps per inch and verify that your software and computer are up to
the task. Even with a 250 line encoder and a 2:1 reduction on a 5 tpi screw
you get a calculated resolution of .0001. Probably better than the machine
is up to.

Tim
[Denver, CO]

> someone posted before,that for a servo system to work ok the
> resolution should be 10x the needed resolution.
> I've being thinking in upgrade my grizzly to servos using the gecko
> servo drive and 500 lines encoder + a 5tpi screw,that would give me a
> resolution of 0.0001" Would it work ok? or should I get a more
> expenssive encoder(1000 lines).
> 0.0001" is the target resolution.