GearHeads Forum

Mark:

Probably one of the most annoying things is noise trouble. I once had a Cat Scanner that broke down every 4 or 5 days for about
3 months. It was a new type of Scanner and the experts in Germany were a bit mystified as well. They even sent a guy from Germany
to help out in the search. We found eventually it was the X-Ray Tube arcing occasionally and the spike of energy went through the entire
system and kept blowing a chip in a circuit board in a computer far away. This should have been impossible but we found a rubber hose that had wire
supports in it when it shouldn't have. That spike should have simply gone to ground, but that wire allowed it to run on a signal
cable and take out that chip repeatedly.
Long story I know, but its one of noise causing me over a month of anguish, work and about 20 circuit board at around 5K each.
Noise is hard. In your case since your not using a CO2 laser ( which are known for making arc's and noise spikes) its hard
to imagine where its coming from. I usually test like this..

1) Run machine with the Laser or spindle turned off. Any Z motion?
2) Run program with X axis powered off, any Spikes?
3) Same for Y axis, does the Z still tick with it turned off..

Id imagine one of the three tests to show no motion on Z... which helps to narrow the field.

Art

Thanks again, Art, for the advice. I've read through these bits several times and am trying to wrap my head around them.

1) you mentioned 3 axis movement making more noise. The spiral only moves in 2 directions. Not 3. Z axis "ticks"
2) the x-y movement (augs option) doesn't result in "ticks"
3) I tried circular movement (augs) no ticks there either

I'm guessing random noise also because...

I should say at this point there is one more data point. My laser jobs will stop randomly and intermittently via estop. Perhaps noise related too. The continue processor is a life saver but this is still a pain to have to restart 5 or 6 times during a 30 min job
I played a bit today just repositioning cables. No help.

I might be crazy but I originally built my cnc router around smooth steeper, c10 board, gecko stepper drives, 540 nema23 motors and mach3. I've run this for years and am very happy with it.

I built another control box with the pokeys57 in it powered with completely separate power supply. I use a parallel printer db25 switch box to toggle between feeding the drives via the c10 board and gecko drives for router. Switch A = router. Switch B = Laser. So I drove you down this road to agree (with no measurements yet) that the problem is likely noise related. It's not an intermittent wire. But it's not a real standard set up electrically.

So to chase the noise a little, I eliminated the switch box. Removed FEET of cable ( shielded cable) and found no difference in symptoms.

I'm going to try sticking some small caps on the control wires just to do some easy shotgun troubleshooting.

But what I will likely do is build a whole different machine and do the electronics cleanly.

That will take a bit and I've got to think about do I really want to do this with a little 10W diode? Maybe I'll go higher power. Galvo? CO2? Other? Research required.

Any thoughts or advise around any of that BS would be appreciated.

Mark

What I have seen happen is a small wire breakage or noise when a system is in a particular position,
so for example it may get a noise step at Y = .213 exactly..
In a spiral you pass all Y and X points on each pass, so you'd never know of the problem from doing x,y pictures,
Only a spiral would show it.. for example, imagine that when Y is at 40.12mms , it causes the Z step pin to short
to ground momentarily.. you'd never know unless you moved Y across that point repeatedly.. like in a spiral. If you lose
only a few pulses you'd never notice, its only when you repeatedly lose that pulse that you notice.

Spirals are actually a great way to test a systems performance for noise or wire losses. I remember way way back
when a guy named Fred at ( I think CNC systems or something..) made a cnc file of a parabolic curve that moved all
axis on every move. It was great at surprising people with how much loss or false steps they had.. Most 3d cnc systems
run at 2.5D most of the time, its when you do real 3d it shows its true nature..It was a popular file on my Mach3 forum
for a while till it was removed for claimed copywrite as very few files actually move all 3 axis at once.(or in your case..2)

Much of the time it is noise, and when all 3 axis start pulsing together the noise reaches its maxima, and your system
if it is susceptible to noise can lose or gain at that point. So I'd bet on either noise totals getting too high ( probable )
or a small wire break when the system moves to a particular x,y location... ( less probable).. or I have a bug I haven't seen,
( much much less likely :) ). I can tell you the DRO counts the positional pulses very very carefully, I have checkpoints
that recheck themselves with multiple redundancies.

Art

Thanks for the quick response, Art.

The DRO doesn't show any motion. It remains at 0. I've measured about -0.05" movement in the Z while the spiral diameter progresses about 0.5" diameter. I've also put my fingertips on the Z pulley and it "ticks" very consistently.

I'm on a day trip just now so I can't look into your suggestions until tomorrow.

I've given it some thought to just run the spiral gcode ( think I saw it in gearotics folder) and watch the Z.

Also might investigate using your motor tuning tool to look for motion there.

My other uses (now that I think about it) were using the xy motion option to do photo engravings instead of the spiral. I'll look at that too.

It's funny that the unexpected Z motion is so consistent (measured with my fingers) and I've never seen this before in the years that I've used this cnc platform (router spindle). I hope I don't need to break out the o'scope.

Mark

Mark:

By the way, the Aug file contains only laser power levels, it doesn't have any other control. Id check to make sure nothign is loose.
Art

Hi:

Did the Z DRO show any motion? If not it must be noise, Auggie will only move Z when either commanded, or if the Depth per Pass on the Laser3d tab is seto change depth each run of the program. It shouldn't move during the program however, that's usually noise.

Art

It has been a few years since I have used my cnc platform with my diode laser. After much head scratching and shaking ( what old guys do) I was able to get things functioning.

I did some vector engraving for my daughter in law with no problems. I generated that gcode with an independent cad/cam program.

I just tried to use AUGS to do raster engraving and that is where the funny came in.

I have only used AUGS once or twice so I'm just a beginner. I loaded a simple line art jpeg (not a photo) and selected the spiral motion option. Generated the files, loaded the TAP file and ran the program. Everything looked good for a bit until I noticed my Z axis was moving incrementally down. I put my hand on the pulley to make sure I wasn't hallucinating. After the spiral diameter reached about 0.5" I stopped and roughly measured my Z height. The DRO still read 0.0000 but it was actually about 0.05" lower. Hmmm.

I examined the TAP file and saw no Z movement called so I figure it must be the AUG file.

Any ideas what I may be doing wrong? I'm not anywhere near the machine now but I am wondering if the STEPOVER or the OTHER ( can't remember the other variable) should both be set to 0. I did set the OTHER to 0 but not the stepover since I figured that might be used to create the spiral.

Thanks for the assist.