Re: [CAD_CAM_EDM_DRO] lead error versus backlash compensation

flyspeck1@... wrote:

> First off...
> Thanks to all who responded to my question about way
> covers. I'll pipe in with a comment or two as well.
> Neoprene is strange stuff; it comes in all sorts of
> forms... Everything from a permeable "sponge" type that
> is used on wet-suits, or a totally water tight material
> for dry-suits, and finally to.... you guessed it.... way
> covers. In fact, MSC sells a set of for BPort type
> mills that are specifically listed as being neoprene.
> Anyhow, neoprene does work well and is more than
> suitable for light milling applications; Jon Elson
> touched on this. It might not be suitable for 15HP Haas
> power mills, but for the mills that 99% of us are using,
> neoprene works great. McMaster-Carr, by the way, seems
> to have a fairly nice selection. Again, thanks to all.
>
> Now for my next question, and one I'm almost embarrassed
> to ask (because I should know the answer). What is the
> exact difference between lead error and backlash?
> Backlash is pretty easy. It is the amount of rotation
> your screw turns (translated to distance, of course)
> without moving the table when reversing direction. My
> understanding of lead error is that it is the amount of
> error between the commanded position and the actual
> position that a control will tolerate before tripping a
> fault.

No, that is usually called "following error". Lead error is the
difference between commanded position as determined by the
position of the leadscrew, and the actual position of the table.
If the distance between every turn of the thread is exactly .20000",
or whatever lead the thread theoretically is made to, then there is
no lead error. If the distance between every turn was .19999", then the
lead error would be .00001" per turn. If it was a 5 TPI screw, then
the error would be .00005" per inch. But, in most cases, especially
with precision screws, the error is not totally consistent, and you
have to have a calibration table, called a machine error compensation
table, entered into the control.

> Is that accurate? My Hurco control that I'm slowly
> bringing alive (great project, by the way... I've had a
> ball with it and learned a ton about archaic controls)
> has the ability to set DIP switches for both. I can
> measure the backlash easily enough, but the lead error
> is where I'm questioning how to proceed. If my
> summation is correct, I assume that you just keep
> increasing the lead error until the machine no longer
> trips out with an axis error (assuming such a situation
> really does not exist). Is my thought process correct?

I don't know whether this is correct or not. They may use a different
term for it. If you set the limit very tight, you may be able to trip it
by manually applying torque to the motor. If so, you could then
figure out what the trip point is, and if it corresponds to your switch
settings.

Generally, you want to keep the following error (or whatever it is called
on a particular machine) small, and optimize the servo amps and
CNC control's response to minimize the following error developed
at any particular speed or acceleration. Old controls with slow
minicomputers and low servo update rates tend to have more following
error than newer controls, in general.

Jon