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

[ultradryan 9]

I was doing manual leveling and follow it with active leveling once a time every month, mostly.

 

No matter how i manual level the nozzles, whenever active leveling is done, the printer will always press nozzle 2 too close to buildplate once print is started. 

I can see the nozzle was so close that it got pushed back a bit upwards, so close that filament cannot print out of it.

 

Core 1 will always be fine and well calibrated, while Core 2 will always be too close to bed, after active leveling.

Either the active leveling is buggy, or the result of Z calibration between the 2 nozzles are not memorized by the machine.

 

Is there a fix to this? I am not sure why this is happening.

 

I suspect the distance between Nozzle 1 and Nozzle 2 somehow was wrong.

Now I have to set the machine never to do active leveling, and rely on manual leveling.

 

It will be great if there's a way to manually set the Z distance difference of the two cores in cases like this when the active leveling somehow failed?

 

 

Edited July 9, 2019 by ultradryan

[raulfilipe127 0]

I have exactly the same problem... Still no fix?

[RudydG 62]

I had just the same problem, after installing Z height plug in in Cura, which i gave 0.15mm my problem is solved.

 

[CarloK 203]

I have an idea about where the problem comes from. Do you people have set the Active Leveling frequency to 'never'?

Setting it to 'never' means that firmware 5.x will use the manual leveling settings. But when you insert an unknown print core I noticed some people use the Active Leveling to measure this print core.

 

My hunch is that there is a problem in this combination of Active Leveling and Manual Leveling where both methods use slightly different data.

Until we fix this the workaround would be to either:

- re-level with Manual Leveling

or

- Always use Active Leveling (don't set the frequency to 'never').

[bhalter69 1]

I have exactly the same problem. I use frequency "weekly".

 

Now I did the manual leveling and set the frequency to "never". I think this issue started with the last Firmware-Update. Before I only did the manual leveling only once, when I unpacked my the Ultimaker 3 extended.

 

I print quite a lot and use since 2 years the same two extruder. Could it be a issue with the extruder?

The print bed level sensor reports "Signal good".

 

[gr5 2,173]

There's 2 ways to fix this.

 

One is to just turn off leveling.  It's really a waste of your time as manual leveling is so simple and easy and you only have to do it when you swap out cores and usually not even then.

 

If the nozzle levels too close usually it's because the spring in the core is weaker versus the spring in the bed.  You need the core spring stronger because the leveling doesn't know you've hit the bed until it starts moving so if the core moves first then there is a problem.

 

So loosen those 3 springs by turning the 3 knobs about 5 or 10 turns CCW (as seen from below).  That might be too loose to level but try that first and do a manual level (mandatory if you move those screws that much) and then see if auto level is better after that.  Don't loosen them too much.

 

Another solution is to remove the "bad" core.  Remove that one screw in the clear plastic part, take the spring out and stretch it such that it is about 5mm longer.  Put it all back together and it will work great for another few years.

 

[bhalter69 1]

On 9/27/2019 at 8:20 PM, gr5 said:

There's 2 ways to fix this.

 

One is to just turn off leveling.  It's really a waste of your time as manual leveling is so simple and easy and you only have to do it when you swap out cores and usually not even then.

 

If the nozzle levels too close usually it's because the spring in the core is weaker versus the spring in the bed.  You need the core spring stronger because the leveling doesn't know you've hit the bed until it starts moving so if the core moves first then there is a problem.

 

So loosen those 3 springs by turning the 3 knobs about 5 or 10 turns CCW (as seen from below).  That might be too loose to level but try that first and do a manual level (mandatory if you move those screws that much) and then see if auto level is better after that.  Don't loosen them too much.

 

Another solution is to remove the "bad" core.  Remove that one screw in the clear plastic part, take the spring out and stretch it such that it is about 5mm longer.  Put it all back together and it will work great for another few years.

 

 

I tried both methods: I turned the bed to a higher position and stretched the spring in my core. No change at all 😕

Further I changed cores right to left and left to right. The problem remains the same:  With auto-level the right core gets to low.

Could there be a problem with the firmware 5.2.11.20190503 on the Ultimaker 3 Extended?

Now I switched back to Manual Leveling.

One point can't be solved with manual leveling: During the XY calibration the right head prints hardly visible.

[CarloK 203]

I had a long look into this problem, but it is a difficult one to tackle since I can't reproduce it myself.

My feeling is that the problem is caused by a sequence of actions involving the combination of manual leveling and Active Leveling.

 

In the V5.0 release for the S5 we completely overhauled the Active Leveling software. Since the S5 build plate is larger than for the UM3 the undulations became too big to be solved with a simple 3-point leveling, that's why we introduced the grid based probing. The probing algorithm itself was also rewritten to be more accurate and faster.

 

The v5.2 release integrated the UM3 and S5 to be one software package again which for the UM3 has a lot of changes under the hood that you won't see but will be beneficial like:

- A year of software development, including lots of bug fixes.

- A new Linux version

- A new file system on the internal Flash memory for improved long term stability

- The overhauled Active Leveling

- A much improved version of Cura Connect, including Cloud connection

- etc

 

But.... for the users in this thread the v5.2 firmware breaks Active Leveling on the right print core. And I can't reproduce it which annoys me a lot after all the time I spent looking for the cause.

The problem seems to be caused by switching between manual and Active Leveling.

 

Could one (or more) of you please sent me the log files from your printer when it fails? Sent the files to me by private email (hover your mouse on my name avatar picture, a pop-up will show with email option). Add a short description of what you did and what you saw happen.

[Renly77 0]

I have the same problem on my UM3 Extended bought 3 days ago. I currently solved by installing the z offset plugin.
Can I send you the log files?

[V3DPrinting 11]

I have the exact same problem on two UM3E that are 2 and 1.5 years old, running firmware 5.2.11.20190503 since it has been released.

 

The nozzle in slot 2 is squeezed by nearly two layers, making the extrusion nearly impossible, jamming the core and grinding material.

 

I also have an S5, but without any issues.

All printers are printing with 3d Solex Hardcores since the beginning.

 

I print approx 20 hours per day on my printers, and haven't noticed the problem since last Thursday, when I upgraded to Cura 4.4.

I print a lot with a different support material so it should have appeared before.

 

I have checked the gcode file generated for the first layer, and it seems fine, both tools are printing at the same height.

 

I have rechecked the bed height and adjusted it to 14 mm (recommended height by Ultimaker to perform Autobed leveling), increased the spring length by 5 mm on the Hardcore, done a manual leveling, disabled the Autobed leveling.

 

But the issue is still there.

 

[V3DPrinting 11]

I have some updates on this issue on my 2 UM3E.

 

I still have the nozzle in slot 2 squeezed by nearly 2 layers.

 

For some reasons, I have reverted to Cura 4.2.1 (too many bugs on 4.3 and 4.4)

 

I have done a factory reset, rechecked the bed height and adjusted it to 14 mm (recommended height by Ultimaker to perform Autobed leveling), increased the spring length by 5 mm on the Hardcore, done a manual leveling, disabled the Autobed leveling.

 

Then doing a dual extrusion print sliced with Cura 4.2.1. Z offset plugin is present, set to 0. and Extensive Z Offset processing disabled.

 

The result is the same : nozzle in slot 2 is again 2 layers too low.

 

As there is no active leveling involved, it's not a question of springs or sensors.... So the issue is in the firmware 5.2.11.20190503.

 

Remember, I had no issues printing dual extrusion with my 2 UM3E until this new firmware.

 

I understand Ultimaker wants to focus on new products, but at least they should fix what they broke.

[V3DPrinting 11]

@CarloK Is it possible to downgrade the firmware of the UM3E to the latest 3.xx version ? How and where to find the firmware.

 

This would solve my issue as the printers were running perfectly before the firmware upgrade to 4.X and 5.x.

 

I don't mind having the sensor test, nor an improved bed leveling, as it is not working.

 

I just want to get back a reliable dual extrusion, reason why I've bought 2 Ultimaker 3 couple of years ago.

[uncle_bob 21]

Hi

 

I do not see this issue on either of my UM3's. They both run with "stock" AA cores. Last time a manual level was done ... hmm .... errr .... two years ago maybe? Its been all auto level since then. Prior to the manual leveling one of them had issues with the auto level. The problems do not sound like what's going on above.

 

Bob

Edited January 30, 2020 by uncle_bob

[V3DPrinting 11]

1 hour ago, uncle_bob said:

 

I do not see this issue on either of my UM3's. They both run with "stock" AA cores. Last time a manual level was done ... hmm .... errr .... two years ago maybe? Its been all auto level since then. Prior to the manual leveling one of them had issues with the auto level. The problems do not sound like what's going on above.

 

 

@uncle_bob You are lucky. That was my day to day life with my UM3 before upgrading the firmware. Which Firmware version are you running ?

[uncle_bob 21]

Hi

 

I've always run on whatever the most recent firmware is. I check for a new version about once a week. Right now I'm on the latest and greatest, version 5.2.11. The S5 is on 5.4.27.

 

Bob

[V3DPrinting 11]

12 hours ago, uncle_bob said:

I've always run on whatever the most recent firmware is. I check for a new version about once a week. Right now I'm on the latest and greatest, version 5.2.11. The S5 is on 5.4.27.

 

So you are lucky.

 

I'm one of those people having and reporting issues on UM3 since the 5.2.11 on the active leveling / leveling issue.

 

My concern is that with this firmware, even not using the auto bed leveling but doing a carefull manual leveling the right nozzle is not properly leveled and is always below the left one, not allowing any extrusion.

 

I have been through the gcode, and there is no problem with it. So it's not linked to Cura.

 

So the issue is in the 5.2.11 firmware on UM3.

[CarloK 203]

@V3DPrinting I'm sorry to hear about your problems.

First, yes it is possible to downgrade the software to v4.3.3. Attaching the document failed for some reason, but it was also discussed in this thread.

 

The problems you experience are most likely related to the firmware in the printer and I really like to figure out what is going wrong. I tried reproducing all the steps you mentioned but in my printer that worked correct. So, my guess is you are doing some detail in a different way than I did when trying to reproduce.

 

What printcore type and diameter are you using?

What material type are you printing with?

Do you use the default first layer height in Cura of 0.27mm?

When you say you disabled the Active Leveling on the printer, then how did you do this? Did you achieve this by setting the bed leveling frequency on the printer set to 'never'?

Is it possible you mail me the log files of a failing print? (hoover your mouse on my avatar picture, a pop-up will show where you can send me a message).

[tinkergnome 927]

On 1/31/2020 at 11:17 AM, V3DPrinting said:

the right nozzle is not properly leveled and is always below the left one, not allowing any extrusion

 

This sounds more like a problem with the lifting switch than with the leveling procedure?

Does the lifting mechanism for the second print core work at all?

And: there's a calibration function for the lifting switch in the maintenance menu, it's worth a try.

 

 

[V3DPrinting 11]

58 minutes ago, tinkergnome said:

This sounds more like a problem with the lifting switch than with the leveling procedure?

Does the lifting mechanism for the second print core work at all?

And: there's a calibration function for the lifting switch in the maintenance menu, it's worth a try

 

@tinkergnome That's the first thing checked. Lift switch had been calibrated and checked right after a factory reset prior to a test print.

And the problem is on my two UM3E, so it's likely not to be an hardware issue.

[V3DPrinting 11]

12 hours ago, CarloK said:

@V3DPrinting I'm sorry to hear about your problems.

First, yes it is possible to downgrade the software to v4.3.3. Attaching the document failed for some reason, but it was also discussed in this thread.

 

@CarloK Thanks for the link. That was the type of information I was looking for.

 

12 hours ago, CarloK said:

The problems you experience are most likely related to the firmware in the printer and I really like to figure out what is going wrong. I tried reproducing all the steps you mentioned but in my printer that worked correct. So, my guess is you are doing some detail in a different way than I did when trying to reproduce.

 

What printcore type and diameter are you using?

 

Both print core are 3d Solex Hardcore. The left one with a 0,6 nozzle, the right one with a 0,4 nozzle

12 hours ago, CarloK said:

What material type are you printing with?

 

FormFutura TitanX on the left core, Formfutura Helios Support on the right core.

12 hours ago, CarloK said:

Do you use the default first layer height in Cura of 0.27mm?

 

No, as the layer height was only 0,1 mm, I reduced the first layer height to 0,2mm.

See the project file attached with a dummy model for privacy reasons.

 

I had the same issue with prints using the default first layer height.

 

12 hours ago, CarloK said:

When you say you disabled the Active Leveling on the printer, then how did you do this? Did you achieve this by setting the bed leveling frequency on the printer set to 'never'?

 

I first made a factory reset, then lift switch calibration, then set bed leveling frequency to 'never', then checked the distance between the bed and the tray at 14 mm, then done a careful manual bed leveling using a sheet of standard paper.

 

12 hours ago, CarloK said:

Is it possible you mail me the log files of a failing print? (hoover your mouse on my avatar picture, a pop-up will show where you can send me a message).

 

I don't have the logs anymore as I've done some prints since my post on Friday.

I need to make the whole procedure again to provide the logs.

 

Attached some pics of the issue, the print cores used.

UM3E_3DBenchy.3mf

[CarloK 203]

Ok, I see some differences in how I performed my test. I will re-try too.

 

How do you perform the Welcome setup steps? Do you execute the steps where you have to load the printcores? I always skip all the steps.

 

I suspect a problem with the switching bay settings. I know you calibrated this but I've seen it fail and couldn't reproduce it.

In the maintenance/diagnostic menu there is a 'switching test'. Can you execute the test for switching the second nozzle? Are these running good? Is the right nozzle actually moving down and up?

[V3DPrinting 11]

13 hours ago, CarloK said:

How do you perform the Welcome setup steps? Do you execute the steps where you have to load the printcores? I always skip all the steps.

 

Yes, I skip the whole Welcome setup steps. My printers are connected via Ethernet, so no need to go through that process.

 

13 hours ago, CarloK said:

 

I suspect a problem with the switching bay settings. I know you calibrated this but I've seen it fail and couldn't reproduce it.

In the maintenance/diagnostic menu there is a 'switching test'. Can you execute the test for switching the second nozzle? Are these running good? Is the right nozzle actually moving down and up?

 

Here's the video of the switching test. Works like a charm.

 

https://www.dropbox.com/s/y2yacheprqknsd2/IMG_4842.MOV?dl=0

 

@CarloK I am currently making a print with the benchy attached previously. The only change will be the first layer height set to 0,27 mm (default).

I let you know and provide the logs.

[V3DPrinting 11]

@CarloK

I have made the test print of a Benchy with the same materials as before, but with a 0,27 mm first layer height.

Project file attached.UM3E_3DBenchy.3mf

 

Before printing, I have done the following :

- 3d Solex Hardcore AA with 0.6 mm nozzle set in slot 1

- 3d Solex Hardcore BB with 0.4 mm nozzle set in slot 2

- Made a cold pull in both print cores -> no material left in the nozzle.

- Factory Reset, skipped the steps.

- Cura Connect Reset

- Lift Switch calibration

- Switching Test done -> OK, see video from my previous post

- Bed leveling Frequency set to 'never'

- Set Build plate Temp to 80°C (working temp for the job) waited it to stabilize

- Set the 2 printcores temp to 240°C (close to working temp for the 2 materials)

- Manual bed leveling with a sheet of paper, just gripping slightly the paper. I made a specific attention to the Z offset height.

 

I've started the job.

First layer for material 1 in slot 1 is perfect

 

 

First layer for material 2, in slot 2 is too close to the bed, but thanks to the very thick layer height, it has extruded.

 

 

I've aborted the job after the second layer with the material 2.

 

 

I have measured the first layer thickness for each material

 

Material 1 is very well calibrated (0,24 to 0,27 mm actual thickness for a 0,27mm first layer height)

 

Material 2 is nearly 0,2 mm lower than the first layer with only a 0,08 mm thickness.

 

As the bed leveling was done with the print cores and bed hot, dilatation of the hardware has been already accounted.

 

As only a manual bed leveling was done, no sensor, nor springs stiffness is involved.

 

And the test is not brand dependent, as all the calibration was done manually with the printcores and bed at working temp.

 

To me, the only way to introduce this bias is the firmware.

 

Here's the printer logs

UM3E Logs.zip

 

Thanks for your feedback !

[V3DPrinting 11]

@CarloK

I have made the same test print of a Benchy with the same materials as before, but with a 0,27 mm first layer height, 2 Ultimaker Printcores

 

 

Before printing, I have done the following on the same printer as in my previous post

- Ultimaker Printcore AA with 0.4 mm nozzle set in slot 1

- Ultimaker Printcore BB with 0.4 mm nozzle set in slot 2

- Made a cold pull in both print cores -> no material left in the nozzle.

- Set Build plate Temp to 80°C (working temp for the job) waited it to stabilize

- Set the 2 printcores temp to 240°C (close to working temp for the 2 materials)

- Manual bed leveling with a sheet of paper, just gripping slightly the paper. I made a specific attention to the Z offset height.

 

I've started the job.

First layer for material 1 in slot 1 is perfect

 

First layer for material 2, in slot 2 is better.

 

Material 1 is again very well calibrated (0,27 to 0,28 mm actual thickness for a 0,27mm first layer height)

 

Material 2 is nearly 0,1 mm lower than the material 1 first layer with only a 0,2 mm thickness. It is better than with the 3d Solex Hardcore

 

So, actually, with no bed leveling, the situation with the Ultimaker printcores is better.

I haven't done any other print in these conditions to confirm.

 

Still the issue from my previous test with 2 Hardcores manually calibrated at working temperature being not leveled is there.

 

Remember that it has been working properly before firmware 5.x.

 

 

 

[V3DPrinting 11]

I have redone the same print jobs with 2 Ultimaker Printcores, but with active leveling set to 'always'.

The print went well for the two first layers

First layer height on material 1 is between 0.25 and 0.28 mm for a 0,27 mm target.

First layer height on material 2 is 0.23 mm

Which is fine.

 

I have also redone the same print job with 2 3d Solex Hardcores (0,6 mm nozzle for material 1), also with active leveling set to 'always'

First layer height on material 1 is between 0.25 and 0.28 mm for a 0,27 mm target.

First layer height on material 2 is only 0.09 mm

 

So I must admit there is something slightly different in the two brands of print cores, around thermal expansion of the nozzles or print core.

 

It was better handled in the previous firmware and not noticeable when active leveling was on, which is my usual way of working, always swapping printcores.

 

I also use the same combination of 3d Solex Hardcores / materials in my S5 without any problems.

So it is linked to the UM3 with the 5.x firmware.