AXIS CALIBERATION AND THE SAME OLD DIRTY POSITION COUNTER STUCK PROBLEM!! !

[Babar <riet91-at-yahoo.com>]

Hi !
we have got our board repaired (pld replaced) by STG
but the problem of counter stuck when dc-servo is run
from testing dialog box (by setting voltages) is still
there.
We had tried installing EMC with STG-II on a pentium-1
machine
with 64-MB ram. Now we see a bit improvement in
position counter stucking. now we see this problem
occuring for over 5V for each axis.
in past this used to happen for voltages over 3. This
seems to be good improvement by use of a slow machine.
now we are running the x and y axis by limiting output
voltage to +/-4V. (we have z-axis encoder problem
therefore z -axis is not being touched). Have
tuned the coefficients as follows 
P = 100
I = 1
D = 0

VELOCITY FEED FORWARD GAIN = 1
and all other set to 0
now we can jog both the axis smoothely. we even have
run them in mdi and auto mode.


PROBLEMS RELATED WITH CALIBERATION
we want to caliberate our milling so that a 1 inch
travel in emc becomes 1 inch in real!! we dont know
what screw picth we have for our all the three axis. I
tried to find them in machine manual but didnt found
them. instead i've found these:

	positioning accuracy: +/-0.001 in (+/- .03mm)
	system resolution: 0.0005 in (0.01 mm)
	repeatability: +/- 0.0005 in (0.01 mm)
	encoder lines: 200 

some parameters found in the ini are not
understandable even after reading the help files.
these are:

	INPUT_SCALE = ?
in trajectory section there is a parameter called:

	LINEAR_UNITS=0.03937007874016
what is it and do we need to change it?

in [AXIS_X] section there are these parameters called:
	UNITS=0.03937007874016
	MAX_VELOCITY

The following soft limits?
	MIN_LIMIT = -1000
	MAX_LIMIT = 1000

we are using 200 lines encoder so i have set 
	COUNTS_PER_REV = 800
is this right?


PROBLEMS RELATED WITH COUNTER STUCKING:
- i had heard that multiple read pulses to IC7166
during a single read   operation result in wrong
position counts reading. Can anyone tell me   if the
various code modules of emc software might be reading
the       position counter latches simultaniously in
different threads? or is    there any protection of
position counter latches being accessed only   by a
particular code module and a particular thread?

- did anyone check? if the code that is reading the
position counter 
  latches and validating the position counts is cpu
clock dependent?

- what may be the impact of fast frequency at
quadrature inputs?
  because we see counter stuck problem only when there
is fast motion.


HISTORY:
we are using bdiemc209-01 also had tried bdiemc212-01
but found same position counter stucking problem.

used generic.ini modified and renamed it to our
milling machine hurco.ini to suit our STG-II board 
mounted at address = 0x200. we are using tkEMC
display.

for auxiliary output we are using LPT port at address 
 = 0x378

have Quadrature encoder with with differential
outputs.

we had implemented the star ground. and infact played
alot with various grounding options but no use. now i
suspect that the problem lies with some code module in
emc. which troubles.

all the cycle times found in the ini were played with
but no use
in our counter stucking problem

various board base addresses of STG-II were tried but
no improvement
in our counter stucking problem.

regards
babar


__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com