Sometimes PCB klicky probe is unreliable

lord-carlos · Oct 17, 2023

Hello

I have a PCB klicky probe and most of the time it works great.
Just now I turned the printer on and did a few PROBE_ACCURACY tests on different places on the cold plate. It went great.

Code:

probe accuracy results: maximum 4.755000, minimum 4.750000, range 0.005000, average 4.751875, median 4.751250, standard deviation 0.001505
probe accuracy results: maximum 4.522500, minimum 4.517500, range 0.005000, average 4.519250, median 4.518750, standard deviation 0.001392
probe accuracy results: maximum 4.425000, minimum 4.416250, range 0.008750, average 4.421750, median 4.422500, standard deviation 0.002385

I turned on build plate and extruder heater for PLA values and redid the test.

Still all gucci
I did a QGL and it went through without a hitch.

Then I started a print and I get this on QGL

Code:

// probe at 250.000,25.000 is z=8.810541
// probe at 250.000,25.000 is z=8.803041
// probe at 250.000,25.000 is z=8.800541
// Probe samples exceed tolerance. Retrying...
// probe at 250.000,25.000 is z=8.801791
// probe at 250.000,25.000 is z=8.804291
// probe at 250.000,25.000 is z=8.821791
// Probe samples exceed tolerance. Retrying...
// probe at 250.000,25.000 is z=8.828041
// probe at 250.000,25.000 is z=8.828041
// probe at 250.000,25.000 is z=8.830541

Similar problems on all corners.
wat.jpeg

With the probe still attached I run an probe accuracy test again and it's great.

Code:

probe accuracy results: maximum 8.857435, minimum 8.849935, range 0.007500, average 8.853435, median 8.853685, standard deviation 0.002222

Other times everything just works™

eriks · Dec 1, 2023

Hey lord-carlos,
could you figure it out? I have similar problems and im out of tries. (it was working fine for a long time...)

NoGuru · Dec 4, 2023

eriks said:
Hey lord-carlos,
could you figure it out? I have similar problems and im out of tries. (it was working fine for a long time...)

I have seen the part that attaches to the toolhead come loose. Make sure that part cannot move.

eriks · Dec 12, 2023

Thanks, will have a deeper look into that. Did a probe test with unsatisfying results.

Code:

                                      min       max     first      last      mean       std  count    range    drift
test                                                                                                               
1: corner 30 samples (25, 255)   5.406603  5.447853  5.406603  5.447853  5.445228  0.007481     30  0.04125  0.04125
2: corner 30 samples (275, 255)  5.420353  5.446603  5.420353  5.444103  5.443520  0.004581     30  0.02625  0.02375
3: corner 30 samples (25, 5)     5.432853  5.460353  5.432853  5.460353  5.456853  0.005215     30  0.02750  0.02750
4: corner 30 samples (275, 5)    5.395353  5.419103  5.395353  5.417853  5.416686  0.004824     30  0.02375  0.02250

Repeatability test:
Take 20 probe_accuracy tests to check for repeatability
Test number: 20...19...18...17...16...15...14...13...12...11...10...9...8...7...6...5...4...3...2...1...Done

                            min       max     first      last      mean       std  count    range    drift
test                                                                                                     
01: center 10 samples  5.440353  5.475353  5.442853  5.467853  5.465259  0.007725    200  0.03500  0.02500
02: center 10 samples  5.441603  5.474103  5.441603  5.467853  5.465141  0.007643    200  0.03250  0.02625
03: center 10 samples  5.440353  5.474103  5.444103  5.467853  5.464997  0.007524    200  0.03375  0.02375
04: center 10 samples  5.441603  5.472853  5.442853  5.466603  5.465009  0.007554    200  0.03125  0.02375

Your probe config uses median of 10 sample(s) over 4 tests
Below is the statistics on your median Z values, using different probe samples

       mean       min       max       std    range  sample_count
0  5.444103  5.444103  5.444103  0.000000  0.00000             1
1  5.453322  5.452853  5.454103  0.000598  0.00125             2
2  5.461603  5.461603  5.461603  0.000000  0.00000             3
3  5.463634  5.462853  5.464103  0.000598  0.00125             4
4  5.464572  5.464103  5.465353  0.000598  0.00125             5
5  5.465353  5.465353  5.465353  0.000000  0.00000             6
6  5.466603  5.466603  5.466603  0.000000  0.00000             7
7  5.466603  5.466603  5.466603  0.000000  0.00000             8
8  5.466603  5.466603  5.466603  0.000000  0.00000             9
9  5.466915  5.466603  5.467853  0.000625  0.00125            10

Drift test:
Take 100 samples in a row to check for drift

                         min       max     first      last      mean       std  count    range    drift
test                                                                                                   
center 100 samples  5.435353  5.466603  5.435353  5.466603  5.465941  0.003293    100  0.03125  0.03125

lord-carlos · Dec 15, 2023

eriks said:
Hey lord-carlos,
could you figure it out? I have similar problems and im out of tries. (it was working fine for a long time...)

Not really.
I just let it soak for 10 minutes and then it works.
I mostly get the error when the printer is cold. Though this also can happen after soak, but less so.

eriks · Dec 17, 2023

there is no difference for my pinter between cold, hot, soaked... :-(
what i have found out today by accident during z-calibration (first the nozzle goes on the endstop, then the body of the microswitch -> after that i do not have retries. will print a new mount and body for clicky and see if it works.

lord-carlos · Dec 18, 2023

eriks said:
there is no difference for my pinter between cold, hot, soaked... :-(
what i have found out today by accident during z-calibration (first the nozzle goes on the endstop, then the body of the microswitch -> after that i do not have retries. will print a new mount and body for clicky and see if it works.

So it's the z endstop that has retries for you and not the clicky?

eriks · Dec 26, 2023

lord-carlos said:
So it's the z endstop that has retries for you and not the clicky?

Sorry, that was little confusing.
-> -no, when probing the endstop there are not retries due to tolerance, not with the nozzle, nor with the body of the clicky.
Will send a quick update how it is with a new afterburner / clicky combination. Currently building the filametrix mod.

tx_wombat · Jan 6, 2024

eriks said:

Thanks, will have a deeper look into that. Did a probe test with unsatisfying results.

Code:

                                      min       max     first      last      mean       std  count    range    drift
test                                                                                                              
1: corner 30 samples (25, 255)   5.406603  5.447853  5.406603  5.447853  5.445228  0.007481     30  0.04125  0.04125
2: corner 30 samples (275, 255)  5.420353  5.446603  5.420353  5.444103  5.443520  0.004581     30  0.02625  0.02375
3: corner 30 samples (25, 5)     5.432853  5.460353  5.432853  5.460353  5.456853  0.005215     30  0.02750  0.02750
4: corner 30 samples (275, 5)    5.395353  5.419103  5.395353  5.417853  5.416686  0.004824     30  0.02375  0.02250

Repeatability test:
Take 20 probe_accuracy tests to check for repeatability
Test number: 20...19...18...17...16...15...14...13...12...11...10...9...8...7...6...5...4...3...2...1...Done

                            min       max     first      last      mean       std  count    range    drift
test                                                                                                    
01: center 10 samples  5.440353  5.475353  5.442853  5.467853  5.465259  0.007725    200  0.03500  0.02500
02: center 10 samples  5.441603  5.474103  5.441603  5.467853  5.465141  0.007643    200  0.03250  0.02625
03: center 10 samples  5.440353  5.474103  5.444103  5.467853  5.464997  0.007524    200  0.03375  0.02375
04: center 10 samples  5.441603  5.472853  5.442853  5.466603  5.465009  0.007554    200  0.03125  0.02375

Your probe config uses median of 10 sample(s) over 4 tests
Below is the statistics on your median Z values, using different probe samples

       mean       min       max       std    range  sample_count
0  5.444103  5.444103  5.444103  0.000000  0.00000             1
1  5.453322  5.452853  5.454103  0.000598  0.00125             2
2  5.461603  5.461603  5.461603  0.000000  0.00000             3
3  5.463634  5.462853  5.464103  0.000598  0.00125             4
4  5.464572  5.464103  5.465353  0.000598  0.00125             5
5  5.465353  5.465353  5.465353  0.000000  0.00000             6
6  5.466603  5.466603  5.466603  0.000000  0.00000             7
7  5.466603  5.466603  5.466603  0.000000  0.00000             8
8  5.466603  5.466603  5.466603  0.000000  0.00000             9
9  5.466915  5.466603  5.467853  0.000625  0.00125            10

Drift test:
Take 100 samples in a row to check for drift

                         min       max     first      last      mean       std  count    range    drift
test                                                                                                  
center 100 samples  5.435353  5.466603  5.435353  5.466603  5.465941  0.003293    100  0.03125  0.03125

Whats the gcode to run this probe test? - I've only seen the basic one before...

lord-carlos · Jan 7, 2024

After a break I printed a bit again today.
Sooo much "Probe samples exceed tolerance. Retrying."

Even after it was warm / stable.

I lowered z speed to 10, did not help.
I wonder if it's because I use this probe:

Someone mentioned the normal probe without rubber is better. Forgot the details though.

lord-carlos · Jan 7, 2024

Omg .. that might have been it ...
standard deviation 0.000625

23:18:21 // probe at 260.000,240.250 is z=5.910000
23:18:22 // probe at 260.000,240.250 is z=5.911250
23:18:23 // probe at 260.000,240.250 is z=5.910000
23:18:25 // probe at 260.000,240.250 is z=5.910000
23:18:26 // probe at 260.000,240.250 is z=5.910000
23:18:27 // probe at 260.000,240.250 is z=5.910000
23:18:28 // probe at 260.000,240.250 is z=5.910000
23:18:29 // probe at 260.000,240.250 is z=5.910000
23:18:30 // probe at 260.000,240.250 is z=5.910000
23:18:31 // probe at 260.000,240.250 is z=5.910000

Ahhh yes. Looking fine.

sanketss84 · Jan 25, 2024

lord-carlos said:
Omg .. that might have been it ...
standard deviation 0.000625

23:18:21 // probe at 260.000,240.250 is z=5.910000
23:18:22 // probe at 260.000,240.250 is z=5.911250
23:18:23 // probe at 260.000,240.250 is z=5.910000
23:18:25 // probe at 260.000,240.250 is z=5.910000
23:18:26 // probe at 260.000,240.250 is z=5.910000
23:18:27 // probe at 260.000,240.250 is z=5.910000
23:18:28 // probe at 260.000,240.250 is z=5.910000
23:18:29 // probe at 260.000,240.250 is z=5.910000
23:18:30 // probe at 260.000,240.250 is z=5.910000
23:18:31 // probe at 260.000,240.250 is z=5.910000

Ahhh yes. Looking fine.

can you further simplify and explain if you got it this resolved.

NoGuru · Jan 25, 2024

sanketss84 said:
can you further simplify and explain if you got it this resolved.

Carlos is saying the rubber piece over the probe switch is creating the deviation.

sanketss84 · Jan 25, 2024

NoGuru said:
Carlos is saying the rubber piece over the probe switch is creating the deviation.

got it. thanks

eriks · Oct 15, 2024

tx_wombat said:
Whats the gcode to run this probe test? - I've only seen the basic one before...

Sorry for the late reply, did not find any time to work on it. Probe accuracy tests: https://github.com/sporkus/probe_accuracy_tests

And yes: Problem still existis with CW2, Filametrix mod, new klicky... when hot QGL aborts due to too many retries..

NoGuru · Oct 16, 2024

eriks said:
Sorry for the late reply, did not find any time to work on it. Probe accuracy tests: https://github.com/sporkus/probe_accuracy_tests

And yes: Problem still existis with CW2, Filametrix mod, new klicky... when hot QGL aborts due to too many retries..

Switch to Beacon and have FAST and more accurate probes.

Laserbea4k43 · Oct 16, 2024

NoGuru said:
Switch to Beacon and have FAST and more accurate probes.

This. I didn't have a nice klicky like yours, but I had to buy a pack of switches and swap them out every couple of weeks or so. It was annoying. Went to TAP, and eventually Cartographer. Not looking back.

eriks · Oct 17, 2024

somehow i keep buying new parts for this machine, without ever coming to a point where i can say: now i can print without watching it

Challenge now: im running an EBB36 with USB (as i had massive issues getting it up and running with CAN), which works wonderful.

thanks!

eriks · Nov 27, 2024

Laserbea4k43 said:
This. I didn't have a nice klicky like yours, but I had to buy a pack of switches and swap them out every couple of weeks or so. It was annoying. Went to TAP, and eventually Cartographer. Not looking back.

Said. Done!
So upgraded to Cartographer USB, works like a charm. But i guess i found out now that i have some other problem, trying to figure out but kind of helpless now. Doing a scan of my bed, without print plate comes out quite well (my guess), hightmap shows a range of 0.109mm.

But when i attach my spring sheet, tried all 3 i have, i get results i do not understand. -> Do i need a new magnet? Or are all my spring sheets garbage? Or is this normal and i am good to go.
Got 2 Times perfect first layer, but now suddenly again disturbing results (5 patches, some are perfect, some are squished into the flexplate) and that even while using bedmesh > that does not make sense at all to me.

argusz · Nov 28, 2024

Have you guys tried adjusting the bed screws? One should be tight and the other three not tight, but not loose either. If the bed has play because the screws are loose it can give errors when probing with klicky...

Sometimes PCB klicky probe is unreliable

Well-known member

Member

Well-known member

Member

Well-known member

Member

Well-known member

Member

Member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Member

Well-known member

Well-known member

Member

Member

Active member

Similar threads