Probe Basic - intermittent erratic probing

[Deb12, LCNC 2.9, PB APT stable]

Setting up and calibrating my spindle probe and am running in to an odd problem. I'm not sure where to start troubleshooting.

When I run the calibration routine (ring gauge), about one in three times the routine 'misfires'.

• Behavior #1
  • After the second (slow) touch, the probe does not return to center before starting the next move.  The next move (opposite axis) is started at the back-off distance.
• Behavior #2
  • After a fast probe touch, it backs off but then fails to do the second touch and continues on to the next move in the routine.
• Bonus
  • The probe calibration offset field doesn't update after a successful calibration routine.
  • Example: probe a 0.7874" (20mm) ring
  • Results window: X-dia = 0.7886, Y-dia = 0.7884... but the offset field stays 0.000000 as if there is zero trigger distance.

The 'mis-probing' (#1 & #2 above ) has happened in both X and Y axis moves, and not at the same point in the routine.  I don't think it's an error in the subroutines as the behavior isn't consistent.

Settings:
Fast probe: 20ipm
Slow probe: 1ipm
Back-off: 0.025"
Non-probe traverse: 50ipm

I do have a false-trigger logic sequence in HAL to stop 'probe tripped during non-probe move' errors.  This logic only permits the probe signal to pass to motion.probe-input when a probe move is in progress. However, bypassing that logic (probe pin direct to motion.probe-input) didn't change or reduce the frequency of probing routine hiccups.

I've not done extensive testing, but similar behavior has happened twice during the four or five times I just tried probing the ring gauge with a 'normal' routine (not calibration screen).  Probe/spindle orientation doesn't make a difference.

I can fudge the probe tip diameter and wind up with the right dimensions, but I'm not sure where to start on the intermittent 'mis-probing' thing.  I can slow it down and try to see if the distance-to-go is appropriate before each move, but I'm not sure I'm going to catch it.

Any suggestions?

I use Probe Basic exclusively but I'm about 2-3 months behind on updates at the moment.

Haven't come across the errors you describe but I did have to add a debounce filter with a delay of 10.

I've been following the QtPyVCP and Probe Basic changelogs on GitHub and if my memory serves me right there has been some probing related changes in the past few months. Maybe Lcvette or TurBoss can figure out why you are seeing this behavior.

Thanks for chiming in.

I don't have a debounce (or dbounce) component in the probe loop, but I'm not sure if I need it.  I'm using a wireless probe (Vers. WLR v2) and I think the actual probe signal is debounced before the receiver sends the signal on to LCNC.

Could be wrong though.  I'll try direct-connecting the probe to the receiver (via cable) tonight and see if eliminating the 'wireless' part changes anything.

The frustrating thing is that the errors are intermittent, and I don't want to raise an issue on GH until I know for certain there's a subroutine error (and I suspect there isn't).

I had nothing but problems with my Vers WLR probes. I had 4 different probes ranging from v2 to v7 and all had intermittent receiver and trigger issues. I gave up on them even though I really wish they would work. There isn't another wireless probe in that price range that I know of. Not sure if you are experiencing any of the same issues but the odd behavior (especially #2) sounded like the problems I had.

I'll add a debounce filter later tonight and see how it goes.  I was tempted to get the latest v9 WLR because it has the spin-to-wake function - nice for keeping it in the ATC.

Temptation is dwindling, however.

I found there was an error in the ring calibration subroutine, it was corrected in the most recent update, so you will want to run apt update for that fix.

@Lcvette:

I just checked in to the forum to report that I'd found a parameter variation between my probe_cal_round_pocket version and the one on GH...

Too funny.  I'll get the latest version in a sec.

Regarding the erratic behavior, I added some dbounce and it was worse at delay=10.  Turning it down to delay=2 seems to have sorted things out for now.  Multiple probing attempts were successful - no fails - but the jury is still out.

Relayted question...  I found an error in the terminal at the end of the calibration routine:

Is this something different, or related to the incorrect parameters in the subroutine?  I wasn't able to find any "calibration_offset" widget in either my installation or on PB/QTPYVCP github pages.

Check the stickies, big changes that require config updates, the sim folder has the files with all the changes and a libido to the docs page. All subs require an update, var file, yaml file, ini file, post gui hal file

Lcvette wrote:

Check the stickies, big changes that require config updates, the sim folder has the files with all the changes and a libido to the docs page. All subs require an update, var file, yaml file, ini file, post gui hal file
 

I did see that, and overwrote all my old subs, var, etc. when I reinstalled everything a month or so ago.

With the most recent cal_round_pocket sub installed the calibration routine started working.

Unfortunately, the erratic probing is still happening.  Fiddling with debounce from 2 all the way up to 25 didn't seem to make much difference, although at 25 it was doing single-touches at every point.

Started halscope to see if the input - upstream of the debounce - was acting flakey... nope.  The input signal was clean and not jumping or fluttering.

Changed the radio channel on the receiver and that stopped a couple of complete disconnects.  But still didn't resolve the erratic behavior described in my first post.

So I'm still stumped.

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