UM3 NOZZLE 2 always leveled too close to bed after active leveling

[CarloK 203]

V5.2.17 was released today and fixes most of the known Active Levelling problems. With the 5.2 release we introduced an optimization for the probing speed by re-using old probing data when this existed but we missed a few situations where this old data should have been discarded. 

Most stupid bug was that we did re-use the probing report of a 1-nozzle print for a 2-nozzle print. This created many of the problems reported here where the 2nd nozzle was grinding or printing in the air. Note that these optimizations affect the UM3 only, the S-line uses a different algorithm and probes always.

 

To my knowledge all blocking Active Levelling bugs in the UM3 have now been fixed, only a few exceptional situations remaining.

[V3DPrinting 11]

On 8/5/2021 at 12:51 AM, CarloK said:

V5.2.17 was released today and fixes most of the known Active Levelling problems.

Thanks @CarloK !

 

I have upgraded my two UM3E to 5.2.17 from 4.3.3 in two steps without any issues.

The dual extrusion now works properly with this FW.

 

I have also implemented the tweak you've provided on mesh leveling but kept 9 probing points, so I'll be able to reduce the first layer thickness for fine prints.

[ddrsarah 0]

This issue still persists in 5.2.18. After upgrading both my UM3 and UM3E, the UM3E started exhibiting this exact behaviour although the UM3 did not.

 

After performing manual levelling and running the active levelling procedure, Print core 2 would print WAY too close to the bed. Was pulling my hair out trying to troubleshoot thinking it was a mechanical issue. Also tried swithcing print cores for both slots (all stock) and the issue persisted. After finding this thread and setting active levelling to "Always" (it was previously set to once a week), I could resume printing with active levelling calibrating the values correctly.

 

These are the steps that introduced the issue on my UM3 Extended (again did not apply to my UM3)

 

1) Perform manual levelling

2) Run active levelling from the build plate menu

3) Try to print with print core 2 -- core 2 is too close to the bed

 

^ worth noting I repeated the above steps multiple times for troubleshooting so maybe an error was introduced due to redundancy

[AndersK 59]

Going from 5.2.11 to 5.2.17 has solved my issues.

However material profiles I had created for PETG couldnt be used after the firmware update.

Updating Cura from 4.7 to 4.10 resolved it though

[V3DPrinting 11]

On 8/5/2021 at 12:51 AM, CarloK said:

V5.2.17 was released today and fixes most of the known Active Levelling problems. 

@CarloK The upgrade to 5.2.17 has fixed the leveling issues. I'm printing dual extrusion since mid August with that version without any issue.

But it has introduced an error in the filament loading and unloading procedures with a wrong length for the Bowden tubes. In um3.json, the Bowden tube length now is 700 mm. I think it was 760 mm in the previous version of 5.2.

Can you confirm ?

[CarloK 203]

@V3DPrinting I already replied your same question in another thread. Summary: no, this value wasn't changed in the 5.2 versions. The last change is from April 2016 when it was reduced from 750mm to 700mm, that's the correct value for the unmodified printers.

[V3DPrinting 11]

11 hours ago, CarloK said:

@V3DPrinting I already replied your same question in another thread. Summary: no, this value wasn't changed in the 5.2 versions. The last change is from April 2016 when it was reduced from 750mm to 700mm, that's the correct value for the unmodified printers.

 

@CarloK Thanks for the reply and sorry for the double post. I couldn't remember if I had posted it. I'll check my Bowden tube length then.

[upatamby 9]

The problem is back in 5.3.0. 😡
The right extruder tries to print into the glass plate again after active leveling. It makes no difference whether I set the Active Leveling Frequency to Never or Allways. 
I have no time for this.
I now will try to find out if/how I can downgrade to firmware 5.2.17. All help is welcome. 
 

[gr5 2,168]

Did you figure out how to downgrade?

 

Here is 5.2.11 - is that good enough?  The first link is for USB install.  The second is for unbricking method install (first method is 100X easier).

 

[V3DPrinting 11]

3 hours ago, gr5 said:

Did you figure out how to downgrade?

 

Here is 5.2.11 - is that good enough?  The first link is for USB install.  The second is for unbricking method install (first method is 100X easier).

 

 

@gr5 the working versions on UM3E regarding that issue are 5.2.17 or 5.2.18

Downgrading to 5.2.11 will not solve the bug.

[V3DPrinting 11]

On 3/17/2022 at 3:58 PM, upatamby said:

The problem is back in 5.3.0. 😡
The right extruder tries to print into the glass plate again after active leveling. It makes no difference whether I set the Active Leveling Frequency to Never or Allways. 
I have no time for this.
I now will try to find out if/how I can downgrade to firmware 5.2.17. All help is welcome. 
 

I've stopped upgrading my UM3E printers, as nothing new under the sun except that bug fix. So still in 5.2.17 and I won't upgrade them anymore for the time there is no compatibility issues with my S5 as a group host or Cura.

[upatamby 9]

9 hours ago, gr5 said:

Did you figure out how to downgrade?

 

Here is 5.2.11 - is that good enough?  The first link is for USB install.  The second is for unbricking method install (first method is 100X easier).

 

I don't think version 5.2.11 does solve the problem. I really would like to get version 5.2.17, but I haven't found a way to  get or restore that version. Until then I just quit using active leveling because it causes more trouble than benefit.

[gr5 2,168]

@robinmdh can you please give us the link for the  swu file for UM3 version 5.2.17?  And maybe also for S5 but I think he only needs it for UM3.  5.2.18 might work if you only have that.

[JohnInOttawa 104]

I haven't upgraded since I rolled back to 4.3.3 several releases ago.  UM3 was OK but UM3X was not. 

 

For UM3X users, which firmware version is giving you better results?  5.2.17 or 5.2.18?

 

Thanks!

John

[gr5 2,168]

@upatamby - here you go!  🙂

 

9066 = UM3

9511=UM3 extended

9051=S5 Release 1

214475=S5 Release 2 (I think also S3 uses 214475)

 

 

 

5.2.18: https://software.ultimaker.com/releases/firmware/9066/stable/5.2.18.20210820/um-update-5.2.18.20210820.swu

5.2.17 for UM3E: https://software.ultimaker.com/releases/firmware/9511/stable/5.2.17.20210723/um-update-5.2.17.20210723.swu

5.2.17 for UM3: https://software.ultimaker.com/releases/firmware/9066/stable/5.2.17.20210723/um-update-5.2.17.20210723.swu

[V3DPrinting 11]

12 hours ago, JohnInOttawa said:

I haven't upgraded since I rolled back to 4.3.3 several releases ago.  UM3 was OK but UM3X was not. 

 

For UM3X users, which firmware version is giving you better results?  5.2.17 or 5.2.18?

 

Thanks!

John

Hi John

If I remember well the history, the version 5.2.17 fixes the leveling issue for UM3X (that's the version I'm on with my two UM3X), version 5.2.18 fixes another bug related to the display, but nothing more.

[gr5 2,168]

Release notes for 5.2.17 and 5.2.18

https://support.ultimaker.com/hc/en-us/articles/360011957840-Ultimaker-3-firmware-release-notes

 

[upatamby 9]

On 3/22/2022 at 4:47 AM, gr5 said:

@upatamby - here you go!  🙂

 

9066 = UM3

9511=UM3 extended

9051=S5 Release 1

214475=S5 Release 2 (I think also S3 uses 214475)

 

 

 

5.2.18: https://software.ultimaker.com/releases/firmware/9066/stable/5.2.18.20210820/um-update-5.2.18.20210820.swu

5.2.17 for UM3E: https://software.ultimaker.com/releases/firmware/9511/stable/5.2.17.20210723/um-update-5.2.17.20210723.swu

5.2.17 for UM3: https://software.ultimaker.com/releases/firmware/9066/stable/5.2.17.20210723/um-update-5.2.17.20210723.swu

Thanks for the link!

I installed firmware version 5.2.17 for my UM3E. Unfortunately it didn't solve the problem. In fact, both extruders were now printing too low. I tried with Active leveling frequency set to Never and also to Manual. I also tried rebooting before each active leveling (this has helped in the past), but nothing worked. Because I didn't feel like spending more time on this, I set the leveling frequency to Never and did the manual leveling again (this was the 'solution' I also used for firmware 5.3.0). But to my unpleasant surprise, this also didn't work properly anymore: extruder 2 was now at the perfect height, but extruder 1 printed too low. If even manual leveling no longer works this really is a problem.
Extruder 1 was a 0.4mm extruder and 2 was a 0.25mm extruder. This might have to do with it, but I did not have this problem with firmware version 5.3.0. However, I tried again with both 0.4mm extruder extruders. After manual leveling they printed fine, but after active leveling extruder number 2 was too low again. So firmware version 5.2.17 doesn't solve the problem for me.
For now: I will wait for a real solution and, in the meantime, try to forget that there is such a thing as active leveling with this printer.

Just a thought: perhaps the Ultimaker firmware programmers could add an option under Calibration to manually adjust the Z-offset (per extruder). I think that would be very usefull as long as the active leveling problem is not solved properly.

[gr5 2,168]

Stupid question but when you manually level the second core, are you turning the knob on the display?  (versus the screw at the back of the bed).

 

I don't see how the firmware can do anything wrong with manual leveling.  I do it all the time.  you need to make sure the nozzles are hot and clean and the bed is hot and I use the calibration card and make sure the friction is exactly the same for both left and right cores and when I'm done I start a print and I usually do some final adjustments as it prints the skirt (turning the 3 metal screws the exact same amount) but then the right core is always (always!) perfect.

 

Maybe you are doing it with a cold nozzle and there is a tiny bit of plastic on the tip?  I do hot nozzle and use the calibration card to clean the nozzle a little as well as to level it.  You don't have to use the official card - you can use other paper but at least the 2 nozzles will be consistent.

 

You can experiment and ssh into the printer and send it gcodes like "G1 Z0" which moves the Z axis to position "zero".  And you can use "T0" and "T1" to switch between the cores (I never tried T0,T1  - I assume it works but I'm not 100% sure it will work through ssh as I think it's sending these directly to the second computer that controls the servos but normally I think T0 is intercepted by the main linux computer and then converted into more than one gcode command).  Google "sendgcode" and "ssh" and "site:ultimaker.com" for examples.

[upatamby 9]

23 hours ago, gr5 said:

Stupid question but when you manually level the second core, are you turning the knob on the display?  (versus the screw at the back of the bed).

 

I don't see how the firmware can do anything wrong with manual leveling.  I do it all the time.  you need to make sure the nozzles are hot and clean and the bed is hot and I use the calibration card and make sure the friction is exactly the same for both left and right cores and when I'm done I start a print and I usually do some final adjustments as it prints the skirt (turning the 3 metal screws the exact same amount) but then the right core is always (always!) perfect.

 

Maybe you are doing it with a cold nozzle and there is a tiny bit of plastic on the tip?  I do hot nozzle and use the calibration card to clean the nozzle a little as well as to level it.  You don't have to use the official card - you can use other paper but at least the 2 nozzles will be consistent.

 

You can experiment and ssh into the printer and send it gcodes like "G1 Z0" which moves the Z axis to position "zero".  And you can use "T0" and "T1" to switch between the cores (I never tried T0,T1  - I assume it works but I'm not 100% sure it will work through ssh as I think it's sending these directly to the second computer that controls the servos but normally I think T0 is intercepted by the main linux computer and then converted into more than one gcode command).  Google "sendgcode" and "ssh" and "site:ultimaker.com" for examples.

Thanks again for your answer.

I love stupid questions, because sometimes they lead to new insights 😉 . In this case the answer is that I do turn the knob on the display for setting the second extruder. The new insight is that this rotation must be stored somewhere and that seems to me to be a task of the firmware (just like storing the data during active leveling). So firmware does play a role in manual leveling. 

Manually adjusting the buildplate while printing a skirt or the first layer is something I already do (although it should be redundant - why else would you need active leveling?).

I always make sure the nozzles are clean. I have always done manual leveling with cold nozzles (default behaviour of the firmware). But after your advise I did manual leveling again but now with warmed up nozzles. A level test print showed that extruder 1 was printing correctly everywhere, but extruder 2 was now slightly too high. You can't fix that during printing by turning the screws up, because then extruder 1 will be too low. So you have to do manual leveling again (although a working active leveling would solve this automatically).

Again: if you could adjust the Z-offset manually (via the printer menu) at that moment, that would be a great option.
I have now (against my better judgement) tried active leveling again, but extruder 1 was printing too low as usual.

I'm sorry, but I don't feel capable of experimenting with sending gcodes. I have to skip that.

 

At the moment, I have no choice but to carry out manual leveling and then test with a test print whether the Z offset is correct. Unfortunately, it is very time-consuming to have to carry out the entire manual leveling procedure again just for an incorrect Z offset.

 

As for the use of the Calibration card, I wonder if you can replace it with an ordinary piece of paper. The thickness of that card seems to me to determine the initial distance from the nozzle to the buildplate. Only if you also do an active leveling (which must work correctly!) then this distance will automatically be adjusted to the correct value. 

 

In short, a real solution is still very welcome.

[V3DPrinting 11]

2 minutes ago, upatamby said:

I have now (against my better judgement) tried active leveling again, but extruder 1 was printing too low as usual.

@upatamby

I am using the FW 5.2.17 on 2 UM3E since it has been issued. That have fixed my dual extrusion issues. And I print a lot in dual extrusion and a lot as it's my business.

I can confirm that the FW is working fine. But active leveling is not only FW, HW is also involved and may have a significant impact.

 

Have you checked the print cores springs ?

If the spring is too weak, then the z offset for that printcore would be too high during the active leveling procedure, therefore it would print too low.

Maybe a first test would be to swap your printcore 1 by another AA printcore.

 

You can also do a test on a scale : you hold the printcore by its body and press the nozzle on the scale until it starts moving. Then note the force applied. It should be somewhere above 0,9 kg for a genuine printcore, to be confirmed by @gr5

 

Have you checked the buildplate height against the printbed frame ? It should be 14 mm or so. Make a search on the forum, you should find some gauge for an easy adjustment.

If the distance is too high, the springs would not be compressed enough and the build plate would move when doing the active leveling. Same result as for the printcore spring.

 

Also checking the sensor would be a good thing (in maintenance menu).

[gr5 2,168]

Somewhere around 0.9kg, yes.  That's typical.  If the 3 screws on the bed are too tight that means the spring pressure of the bed is on the high end and that can mess up active leveling also but... but... the problem is also in manual leveling which is so hard to believe.  We would have thousands of people talking about this issue if the firmware couldn't store the value.  so something is very strange.

 

Oh!  So there is supposed to be some grease on the core on the metal flange.  Just above the heat break.  Such that the flange slides in perfectly into that anodized fork.  The fork makes sure the cores are aligned perfectly in X,Y and Z.  Every time.  So that you only have to calibrate once ever (well anytime you mess with the 3 bed screws I guess the calibration is in question again.

 

So some advice for doing manual leveling in under 30 seconds.

 

When it talks about "level so it's 1mm off bed" - ignore that and click continue 3 or 4 times until you can do the calibration card.  Then once you get the back adjusted, skip the front two corners (you haven't touched the screws so they should be fine still) and then do the second core.  Entire procedure (after heating nozzles to 180C or hotter) should take well under 30 seconds.

 

The point is you are setting the height - not really leveling as of course it's already level (or another way to say it, the bed plane is parallel to the gantry plane).

 

Another thought - are you printing in the same spot on the glass?  Sometimes the glass is a bit curvy/hilly and if you are printing in one area with the left core and a completely different part of the glass bed with the right core - that may be the overall issue and not the leveling procedure.

 

[gr5 2,168]

Oh - the grease - so it might be that the core isn't always sliding all the way down into it's seat (when the right core is in the down position).  I've seen this before.  Fixed with grease.  The grease should last many years but if you keep changing the right side core (or left side core) then you might get inconsistent leveling on one or both cores.

 

Lift the right core up and then down using the switch on the head.  Then jiggle the nozzle (push nozzle gently towards the front and back of the printer).  Does it suddenly go farther down?  Try it a few times.  If it does move down due to a little jiggle then you may need some high temp grease.

[upatamby 9]

The magic trick was a factory reset.

 

Since then, manual leveling works again as it should, with the Z offset being stored correctly.

Active leveling also works fine now, but only when the frequency is not set to Never. If you perform an active leveling and the frequency is set to Never, or you set it afterwards to Never, then the printcores are invariably too low. So I suspect that @V3DPrinting has set an Active leveling frequency on his printers that is not equal to Never. Is this correct?

So I have to choose: either set a periodic active leveling or do manual leveling only.

Thanks to @gr5's tip (the quick manual leveling procedure), I choose for manual leveling only.

I would like to add to that tip that when using the quick method for setting the back with the calibration card, you should not turn the rear buildplate screw but turn the wheel on the front (contrary to what the display says).

Things I have checked:

- springs of the print cores: ok (around 1100g)

- buildplate height: ok (that check was already in my program)

- sensor: "Good"

- grease: not needed; print core is stable at its place

- Curvy glass plate: test print gives no indication of this. An extra check could be: a bent X- or Y-axis (the thin axes that go through the printhead)

[V3DPrinting 11]

1 minute ago, upatamby said:

Since then, manual leveling works again as it should, with the Z offset being stored correctly.

Active leveling also works fine now, but only when the frequency is not set to Never. If you perform an active leveling and the frequency is set to Never, or you set it afterwards to Never, then the printcores are invariably too low. So I suspect that @V3DPrinting has set an Active leveling frequency on his printers that is not equal to Never. Is this correct?

 

As I swap quite often printcores and material, I have the leveling frequency set to Always.

To be honest, I don't see the point of using the Active leveling when set to Never, and a Factory reset + Manual leveling is the best option then.