Re: [CAD_CAM_EDM_DRO] EMC, INI , MIN_FERROR and acheiving maximum accuracy

Asim Khan wrote:

>I had posted the same question before on the same forum, hoping that
>jon elson, paul, ray henry and les watts may find some time
>replying my post!!
>
>

I did reply, on 3/20, but you must have missed it. Here it is, again :

hi!

>I've recently runned the spindle of my old gold hurco milling machine
>which now has EMC with STG-II as machine controller.
>I just now have made my spindle running. and i have cut some
>circles cut on a wooden log. now I want to super tune my EMC system
>so that i can have maximum possible machine resolution available.
>
>Here are some of my machine settings:
>
>MACHINE resolution = 0.03 mm or 0.001 inch (as depicted in machine'
>smanual)
>Encoder resolution = 200 (lines) x 4 = 800 counts per revolution
>
>lead screws of my machine take 5 revolution for 1 inch
>
>therefore i have in my ini the setting of INPUT_SCALE parameter
>to 4000. (INPUT_SCALE)
>MIN_FERROR value in my ini is 0.01
>
>I want to know what minimum value i can set for MIN_FERROR ?
>so that my machine gives maximum accuracy
>

MIN_FERROR and FERROR do not affect accuracy at all. What the do is
control how much error is permitted before an e-stop happens. MIN_FERROR
is the minimum limit on following error. FERROR increases the limit
based on
current velocity. So, the limit at any particular moment is MIN_FERROR+
(FERROR * velocity in units/second). So, if FERROR is .005, and velocity is
1 unit/second, then it would add .005 units to MIN_FERROR.

>>i've observerd that decreasing the MIN_FERROR severely affects
>>the tuning of beds and causes all time famous and bad "following
>>errors".
>>
>>Please put some light over super tuning of machine for acheivement
>>of maximum possible accuracies. What are the things that limit it
>>and what things must be done to get an EMC based system running
>>at maximum possible accuracy
>>
>>
>
>

What you need to do is get the logging function working, and install the
plotting
software that converts the log files into X-Y plots. You plot commanded
position
against actual position, and adjust parameters until the lines lie on top of
each other. The P,I, D, and FF0, FF1 and FF2 parameters are the ones
that control
the servo loop performance. Generally, you want to set P to the maximum
value
that allows the system to be stable. Have your hand on the E-stop
button when
playing with this - it can go wild very quickly, and could possibly tip
over a
Bridgeport-style machine! When you find instability, turn P down a bit.
Then
put a small value, like 5.0 in FF1, and adjust until you get better
following of
velocity changes, like the starting and stopping of jog movement.

Jon