Probe Basic - intermittent erratic probing

More
15 Oct 2024 12:16 #312145 by spumco

Try using just fast probe and zero out slow probe fr in the parameters page of probe basic, sometimes to slow of a break causes issues

Will do.  Any idea why this is the case?

I forgot to test using the cable (bypassing potential wireless issues) last night, I'll try that as well.

If all else fails, I've got a difference wired probe (ITTP) I could scab in and test to eliminate the probe being the problem.

Please Log in or Create an account to join the conversation.

More
15 Oct 2024 12:55 #312150 by Lcvette
the nature of tripod probes, seems that a faster approach speed yields a more positive switch occurence and less chance of a bounce. its not constrained to linuxcnc, this is a known issue on many control systems. on my machine i use only a single touch with fast feedrate and a feed of 8-10ipm. as long as the calibration is done at the same speed it yields extremely consistent precision and repeatability, +/-0.0001"

the xy clearance is also important, you should have enough distance to allow the probe to be moving at speed and not contact during acceleration. post a screenshot of your probing parameters and i can take a look.

fwiw, wired probes have much less of a chance of having issues, i have found that the wireless probes can false trigger, depending on the type, an IR type isn't great around flourescent lights and radio type can pickup interference. the higher end probes typically use a combination of both to minimize false triggers and interference issues. cheaper ones are typically a single layer type.
The following user(s) said Thank You: spumco

Please Log in or Create an account to join the conversation.

More
15 Oct 2024 13:17 #312151 by spumco
Not at the mill now, but here are the parameter values I was fiddling with last night:

Traverse: 50-60ipm
Slow probe: 1-4ipm
Fast probe: 15-25ipm
XY back-off clearance: 0.025"-0.050"
Z back-off: 0.025"-0.050"

I can certainly test other values tonight - perhaps start with slow=0, and fast=10ipm?

Please Log in or Create an account to join the conversation.

More
15 Oct 2024 13:33 #312154 by tommylight
Just shooting from the hip here, but did anyone try to add 22-100nF capacitors to probe outputs or controller/Mesa inputs?
That is nano farads.
Seems like the probe is being triggered from vibrations while the machine is moving, but i do not have one to test.

Please Log in or Create an account to join the conversation.

More
15 Oct 2024 13:46 #312157 by Lcvette

Not at the mill now, but here are the parameter values I was fiddling with last night:

Traverse: 50-60ipm
Slow probe: 1-4ipm
Fast probe: 15-25ipm
XY back-off clearance: 0.025"-0.050"
Z back-off: 0.025"-0.050"

I can certainly test other values tonight - perhaps start with slow=0, and fast=10ipm?

yeah bump up that clearance, you are likely blowing through that small clearance amount on the fast feed rate if using a wireless probe, give it 0.100 - 0.1500" if using fast and slow probe, make sure you calibration is done at the same final feedrate.  to start i would use just a single touch with slowfr zeroed out. try the below parameters and report back:

 
Attachments:
The following user(s) said Thank You: spumco

Please Log in or Create an account to join the conversation.

More
15 Oct 2024 14:52 #312163 by spumco

Just shooting from the hip here, but did anyone try to add 22-100nF capacitors to probe outputs or controller/Mesa inputs?
That is nano farads.
Seems like the probe is being triggered from vibrations while the machine is moving, but i do not have one to test.

Tommy,

I've got a motion-type filter in place so that the probe signal is only connected to motion.probe-input during a probing move.  That has eliminated the 'probe-tripped-during-a-non-probing-move' errors I was constantly getting.

However, that wouldn't do anything for vibration-induced probe triggers during a probing move.  That issue is what I was hoping the debounce filter would eliminate.

Please Log in or Create an account to join the conversation.

More
16 Oct 2024 12:11 #312245 by Lcvette
get it sorted?

Please Log in or Create an account to join the conversation.

More
16 Oct 2024 18:34 #312285 by spumco

get it sorted?

Hijacked by a 'domestic' crisis last night, no time to test.

I'll have a go tonight and report back.
The following user(s) said Thank You: Lcvette

Please Log in or Create an account to join the conversation.

More
16 Oct 2024 22:22 #312307 by spumco
Well that did the trick.

Removed the second slow probe, increased the fast probe speed to 8, and cut debounce back down to 2.

10 out of 10 clean probes!

Upped the speed to 25, re-calibrated, and it's still repeating perfectly.  Deviation between pocket probes is maybe 0.0001" 2 out of 10 sequential probes.  The other 8 of 10 are perfect repeats.

Even found a probe rotational position where X and Y are a perfect match.

Thanks a TON for the assistance.  Now I just need to figure out how to incorporate the probe's error signal in to the probing routines so it stops if the RF error is present during the probing.
The following user(s) said Thank You: tommylight, Unlogic

Please Log in or Create an account to join the conversation.

More
16 Oct 2024 22:52 #312311 by Lcvette
Great news! Glad you got it sorted! 25 is a bit fast for small holes, I would recommend finding a balance to be able to achieve full acceleration in small holes to ensure the probe event is at speed for best consistency. But finding the sweet spot is easy now that you're not getting errors.
The following user(s) said Thank You: tommylight, spumco

Please Log in or Create an account to join the conversation.

Moderators: KCJLcvette
Time to create page: 0.103 seconds
Powered by Kunena Forum