Re: [CAD_CAM_EDM_DRO] Re: Bridgeport Ballscrew Backlash

Mark Vaughan wrote:

>2 Are the servo drives set up correct. When you set a servo up, it's
>response will look nice with P and D parameters, but it will stop before it
>gets to position or just after, there will always or near always be an
>error, it needs a little more oomph on the last bit.
>

This is an incredibly good point. Your system is open-loop back to the
computer (I am
assuming, with Mach 3), so you have no idea what the actual position
is. You only know
where you commanded it to go, and that it is within +/- 128 counts so
that the Gecko
didn't fault. (It really has to be much closer, but depending on a
range of parameters
such as motor resistance, Ilimit setting and supply voltage it can be
tens of counts.
Mariss at Gecko swears this is impossible due to the I term, but a
friend of mine
has a machine in town with G320 drives, and you can turn his handles
1/10th of
a turn and they will just stay there! This is about 50 counts on his
motors. He really
needs to increase his supply voltage, but it still shouldn't do this.)

There is a test point inside the Gecko drive where you can monitor
instantaneous
position error with a scope or DVM. I think the manual tells how to
convert voltage
into position error, in encoder count units.

It might be a good idea to read the position error at the ends of travel
when doing
this test. Just the overhanging weight of the table will increase
friction at the ends
of travel. It may be that all the error you are seeing is following
error, and maybe
electronic tuning on the servo drive would reduce it substantially.

Jon