Re: [CAD_CAM_EDM_DRO] Re: Wierd Z axis in EMC
Posted by
Tim & Heather Smith
on 2002-07-12 06:47:11 UTC
Thanks for the info Ray,
I think I may have stumbled across the problem. First of all I tried another
PC and had exaclty the same problem.
The driver cards I have dont use step/direction signals, they use CW/CCW
signals.
I built an interface that will convert the step/dir outputs to CW/CCW. I
use optoisolators to protect the computer.
I Have turboCNC install on a dos partition, it has a setting where the width
of the control pulses can be set. Previously I had to have this set at
around 50us, when I tried it after failing with EMC I had to increase it up
to 60us to get the machine to work.
After talkng to some of the electrical gurus at work, we think this is
temperaure related, and I have been just borderline up til now. It is the
middle of winter here in Aust. and we have had some V. cold nights(for us)
here at around -5 to -10 deg C, (my shed is no warmer than this quite
frequently), we think that the cold temps may have slowed the opto's just
enough to cause a problem.
It turns out that the opto's I used (4n32) have a total switching time of
somewhere around 100us, which was probably why I had to set TurboCNC around
50us in the first place to get it to run. To cut a long story short I have
changed the opto's to faster ones( 4us switching time), now EMC seems OK
again (fingers crossed) and in TurboCNC I was able to reduce the switching
time to 10us and it still worked( set reset it to 20us just in case).
This weekend I hope to cast (I have a little foundry on the side) a couple
of gearbox housings which when assembled, I will fit onto my drive motors to
gear them down a bit, so I can get more torque out of them.
Hopefully this will mean less chance of missed steps.
Thanks again
Tim
I think I may have stumbled across the problem. First of all I tried another
PC and had exaclty the same problem.
The driver cards I have dont use step/direction signals, they use CW/CCW
signals.
I built an interface that will convert the step/dir outputs to CW/CCW. I
use optoisolators to protect the computer.
I Have turboCNC install on a dos partition, it has a setting where the width
of the control pulses can be set. Previously I had to have this set at
around 50us, when I tried it after failing with EMC I had to increase it up
to 60us to get the machine to work.
After talkng to some of the electrical gurus at work, we think this is
temperaure related, and I have been just borderline up til now. It is the
middle of winter here in Aust. and we have had some V. cold nights(for us)
here at around -5 to -10 deg C, (my shed is no warmer than this quite
frequently), we think that the cold temps may have slowed the opto's just
enough to cause a problem.
It turns out that the opto's I used (4n32) have a total switching time of
somewhere around 100us, which was probably why I had to set TurboCNC around
50us in the first place to get it to run. To cut a long story short I have
changed the opto's to faster ones( 4us switching time), now EMC seems OK
again (fingers crossed) and in TurboCNC I was able to reduce the switching
time to 10us and it still worked( set reset it to 20us just in case).
This weekend I hope to cast (I have a little foundry on the side) a couple
of gearbox housings which when assembled, I will fit onto my drive motors to
gear them down a bit, so I can get more torque out of them.
Hopefully this will mean less chance of missed steps.
Thanks again
Tim
>The retro sounds great. Good luck with its completion.in
>
>Which version of the BDI? Is this a download? If yes, did you run the
>checksum? How many axis defs do you have in the ini?
>
>I've not heard of a problem like this and I can't imagine where in the
>freqmod module it might be. Normally freqmod starts up with a maximum
>permitted frequency that is a calculation based on MAX_VELOCITY, UNITS, and
>INPUT_SCALE.
>
>I don't run with an mm definition here and I've heard of some setup
>nightmares and been party to a few myself. First be certain that you have
>commented out or removed or changed the line (about number 22) in the ini
>that says.
>
>; Startup codes for RS-274-NGC interpreter
>RS274NGC_STARTUP_CODE = G20
>
>You want the interpreter to default to G21. It will do that with nothing
>the ini or change this line to read G21.each
>
>Other things to look for include
>
>[TRAJ] LINEAR_UNITS should be set to 1.000
>
>[AXIS ?] UNITS should be set to 1.000
>
>You need to multiply the default
>
>FERROR = 1.000
>MIN_FERROR = 0.010
>
>by 25.4.
>
>I run [TRAJ] MAX_VELOCITY a bit higher than I set the same variable for
>axis. In the case of 188 mm/min I'd start out with the traj one to 3.25 or
>more and each axis to 3.13.
>
>Keep talking.
>
>Ray
>
Discussion Thread
smith67au
2002-07-10 20:53:50 UTC
Wierd Z axis in EMC
Ray Henry
2002-07-11 07:35:13 UTC
Re: Wierd Z axis in EMC
Tim & Heather Smith
2002-07-12 06:47:11 UTC
Re: [CAD_CAM_EDM_DRO] Re: Wierd Z axis in EMC