New 7.0.3 level sensor diagnostic and unrealistic height difference between two extruders

[dmanthei 0]

I have had very frustrating "unrealistic height difference" errors on our Ultimaker S5 from multiple users.  Here's what I've tried:

- Manual bed level twice

- Unload and reload materials

- Unload and reload print cores

- Clean build plate

- Remove and reinstall build plate glass ensure strong contact to metal clips

- Power cycling

- Changing outlets

- Attempting to print with core #2 completely removed (not an option?)

- Running level sensor diagnostic test

 

My question lies with that last point: What are acceptable results for the level sensor test?  I see that Firmware 7.0.3 now reports a numeric result.  Is this in percent? Millivolts? Decibels?  The outlet I was using reported 44.9. I switched outlets, power cycled, and got 35.3 but I don't know if this is good or bad, an improvement or loss, or if the scale is linear or logarithmic.  

 

Any feedback on what these numbers mean would be greatly appreciated so I can continue to troubleshoot this problem.  And I apologize if this information is available somewhere but I couldn't find it. Since this numeric readout for the diagnostic test appears to be a feature just introduced in the latest firmware I suspect documentation is limited or not published.  I checked the release notes but there's no detailed explanation of the readout's meaning.

[glx 22]

I'm can't help with what exactly that value is, but I got 2.6 on my S5 (I honestly didn't run this anytime before now, so I don't have a clue). 

Nearly always when I get that error it is wisps or small debris inside the printhead an/or a dirty front fan/build plate. Once I had a damaged wire in the back of the printhead (the red/white one). 

Do you maybe have any electrical devices nearby thay may cause EMI problems? 

 

[SanneM 12]

We are still working on fine tuning at what numeric value the sensor should be considered broken, so unfortunately the message is really vague at the moment

Why we did choose to show this message on the display already, is so that you can share the number with support

 

I expect that the message will have more clear feedback in 2 firmware updates from now, until then if you are experiencing issues with bed measurement please contact support and share this number with them

 

That said I can say that 2.6 is really good, and 44 or 35 definitely is reason to contact support if you have issues with bed measurement, as your print head circuit board might need replacement

[gr5 2,167]

There are so many things that can give you the same error message.  First and foremost, watch the leveling when it does the first two levelings - left and right core at the same spot.  It's supposed to slowly move the bed until it touches the nozzle.  One failure mode is that it stops before the bed touches the nozzle.  Another is that it touches the nozzle and keeps going much too long afterwards.

 

Another failure mode is that the lift switch isn't raising and lowering the right core.  You might have to repair the print head or re run the lift switch calibration procedure (which takes less than a minute).

 

There may be electronic noise (obviously - discussed above).

 

There may be spring issues with the bed or a print core (basically you want the bed springs weaker than the nozzle springs but the bed springs can be somewhat stiffer and it still works fine because the entire cantilevering bends a bit also).  And of course there can be something on the nozzle tip (usually not a problem as the leveling is done while the nozzle is hot).

 

Oh - and you can have two nozzles that are not the same "height" (where height is from the nozzle tip to the shiny steel angled part that sits in the metal "fork" of the print head).  Tolerance here is 0.7mm.  And you can have nozzles that are not seating properly into said fork.

[dmanthei 0]

Wow, thank you for all the helpful feedback everyone!

 

glx: The area where we keep the printer is free of other devices for at least a 4 meter perimeter.  At the edge of that perimeter is a small office air handling unit on a different electrical circuit.  Beyond that, it's 4 offices with laptops and desktop LCD monitors which likely share an electrical circuit, but these shouldn't be noisy devices.  Would trying a surge protector or UPS help or hurt the situation?

 

SanneM: Thank you very much for providing your sensor value. I agree that's a significant difference and I assume this is a lower-is-better situation.  I will contact support at this point.

 

gr5: All good points and I tried to consider each.  I did a fresh calibration of the lift switch and that seems fine, though it also seemed fine before I did that calibration.  I wondered about spring issues so I checked during another manual-levelling to ensure there was sufficient pressure in the build plate springs. They seem good. I have a spare (but used) AA 0.4 nozzle on-hand so I changed that out to see if it helped with respect to your nozzle height comment.  Unfortunately it did not.

 

I did just grab the log files in preparation for contacting support.  Now, I have no experience or knowledge of how to read a log file, but the file created during the capacitive sensor test was this:

 

capdac 0 value 4312
capdac 1 value 425
capdac 2 value -5981
capdac 3 value -12337
capdac 4 value -18761
capdac 5 value -24946
capdac 6 value -31088
capdac 7 value 32767
capdac 8 value 32767
capdac 9 value 32767
capdac 10 value 32767
capdac 11 value 32767
capdac 12 value 32767
capdac 13 value 32767
capdac 14 value 32767
capdac 15 value 32767
capdac 16 value 32767
capdac 17 value 32767
capdac 18 value 32767
capdac 19 value 32767
capdac 20 value 32767
capdac 21 value 32767
capdac 22 value 32767
capdac 23 value 32767
capdac 24 value 32767
capdac 25 value 32767
capdac 26 value 32767
capdac 27 value 32767
capdac 28 value 32767
capdac 29 value 32767
capdac 30 value 32767
capdac 31 value 32767
capdac 1 value 411
capacitive_baseline 427

o010F93

 

Now again, I don't know what I'm looking at, but as an engineer I suspect the repeated "capdac 32767" lines mean a capacitive digital to analog converter input is reading 32767 which is the maximum value of a 16-bit signed value.  That seems like a problem like an electrical short or accumulated static charge or something, but again I don't know...it just looks wrong to me.  So hopefully support can help me out from here.

 

Thanks again everyone! You helped me determine my problem is not some silly operator error and actually warrants reaching out to support staff.

[gr5 2,167]

5 hours ago, gr5 said:

First and foremost, watch the leveling when it does the first two levelings - left and right core at the same spot.  It's supposed to slowly move the bed until it touches the nozzle.  One failure mode is that it stops before the bed touches the nozzle.  Another is that it touches the nozzle and keeps going much too long afterwards.

1) Okay but what about this above.

 

2) And since you have the log file, do a grep for "peak" and "Preliminary" as shown here:

 

geo@geo-Wild-Dog-Pro:~/ultimaker/logs oct28 2020$ grep peak ultimakersystem-ccbdd3005b4c.5.7.2.boot0.log

 

Oct 29 00:13:45 ultimakersystem-ccbdd3005b4c python3[1166]: 2020-10-29 00:13:45,662 INFO     probingController Found

peak @ 8.623

 

Oct 29 00:13:59 ultimakersystem-ccbdd3005b4c python3[1166]: 2020-10-29 00:13:59,315 INFO     probingController Found peak @ 7.173

geo@geo-Wild-Dog-Pro:~/ultimaker/logs oct28 2020$ grep Preliminary ultimakersystem-ccbdd3005b4c.5.7.2.boot0.log
Oct 29 00:13:59 ultimakersystem-ccbdd3005b4c python3[1166]: 2020-10-29 00:13:59,507 INFO     probeResultValidationStep Preliminary z offset for second nozzle: 1.45

 

Actually first search for "height difference between nozzles is too large" as that is the message that occurs right when the error happens.  Write down the date and time and *then* search for "peak" and "Preliminary" and look to see if even the passing tests *almost* fail.
 

 

[dmanthei 0]

Quote

1) Okay but what about this above.

I didn't reply to this earlier because I'm not certain the failure mode has been consistent.  Most recently it has touched off on the first nozzle ok and makes very little attempt to touch off on the second nozzle before it fails (it's probably 4-8mm from the plate when it 'gives up' on the second nozzle).  This is why I hoped manual leveling would help but it doesn't.

 

Myself and other users in my workplace have also seen it touch off on both nozzles in what appears to be a successful process but it then fails after the second nozzle touches.

 

I'm searching for what you suggested in 2), but it takes some time...I'll report back.

 

[dmanthei 0]

I did not find any results for "Preliminary" case-insensitive.  

 

I found these results for "peak" and included the few lines after "peak" occurs:

 

cap_Ultimaker-S5-01_2022-07-20_13.58.52_N1_X264.5Y193.5.log
# peak found
# z 8.58191
8.09000 609 302525 0.00 3 9 7 2
# done
8.58191

cap_Ultimaker-S5-01_2022-07-20_13.59.06_N0_X286.5Y193.5.log
# peak found
# z 7.12554
6.61750 814 483325 0.00 3 8 5 3
# done
7.12554

cap_Ultimaker-S5-01_2022-07-22_13.26.20_N1_X264.5Y193.5.log
# z 13.86628
13.40250 288 178475 0.00 3 10 8 2
# done
13.86628

cap_Ultimaker-S5-01_2022-07-22_13.26.25_N0_X286.5Y193.5.log
# peak found
# z 13.79476
13.50000 284 146975 0.00 3 9 7 2
# done
13.79476

cap_Ultimaker-S5-01_2022-07-22_13.28.52_N1_X264.5Y193.5.log
# peak found
# z 14.24934
13.76250 268 239900 0.00 3 7 5 2
# done
14.24934

cap_Ultimaker-S5-01_2022-07-22_13.28.57_N0_X286.5Y193.5.log
# peak found
# z 13.09327
12.86000 327 163250 0.00 3 9 7 2
# done
13.09327

cap_Ultimaker-S5-01_2022-07-22_13.44.47_N1_X264.5Y193.5.log
# peak found
# z 8.65690
8.10500 608 357975 0.00 3 14 12 2
# done
8.65690

cap_Ultimaker-S5-01_2022-07-22_13.44.51_N0_X286.5Y193.5.log
# peak found
# z 14.39473
13.93750 286 101475 0.00 3 10 7 3
# done
14.39473

cap_Ultimaker-S5-01_2022-07-22_13.46.42_N1_X264.5Y193.5.log
# peak found
# z 8.52497
8.04250 599 349225 0.00 3 9 7 2
# done
8.52497

cap_Ultimaker-S5-01_2022-07-22_13.46.49_N0_X286.5Y193.5.log
# peak found
# z 12.19153
11.87250 364 167850 0.00 3 8 6 2
# done
12.19153

cap_Ultimaker-S5-01_2022-07-22_16.54.25_N1_X264.5Y193.5.log
# peak found
# z 13.36184
12.87500 324 181075 0.00 3 10 8 2
# done
13.36184

cap_Ultimaker-S5-01_2022-07-22_16.54.30_N0_X286.5Y193.5.log
# peak found
# z 13.79867
13.26000 294 136175 0.00 3 11 9 2
# done
13.79867

 

This was using "findstr" on Windows for all the log files on my USB drive.  I cannot do the same search exclusively on my log file "ultimakersystem-0030d6262a40.7.0.3.boot0.log.gz" because it's an *.gz file for which my organization doesn't have an archive manager installed on my locked-down PC to decompress it.  I can look at home later today if that's going to be helpful in some way.

[gr5 2,167]

All the values < 10mm look great.  All the values > 10mm are a serious problem.  Both your description and the log file are saying the same thing:

 

Sometimes it thinks it hit the glass when it clearly did not.  It never even got within 5mm of the glass.

 

So you can ignore:

1) spring isssues

2) filament on tip issues

3) bad cores

4) bad seating in print head

 

And concentrate on:

1) Bad hardware

2) Electric radio noise

3) Loose wire - I didn't mention this earlier but open the print head and there are 2 wires going to that base plate in the print head.  Tug very gently to make sure there isn't a wire that is barely connected.

 

I know a few resellers and most printers do active leveling just fine and some that fail are easily fixed but some can be quite the headache to fix (replace basically everything twice until it works).

[PieterLemmens 0]

Similar issue here. 

model S5, updated firmware and i have reset factory settings.

 

But i noticed: the colder it gets outside... the more often the problem...
I had a small heater near the printer ( about 0.5m away). I think it might be causig an interference. The warm air raising next to the printer may cause some ESD issues??? It's a large volume of moving air.

 

@dmanthei, you also have an air unit nearby? You might want to move that one and test.

 

Anyway. Just sharing my thoughts here.

Rgds

P