help with lost adhesion please?  

RE: help with lost adhesion please?

@robert-rmm200

The bolt didn't feel tight.  And its fellow bolt was snugged as I usually do with metal in plastic issues.  Not torqued down but snugged up. 

I think this screw failed. Although that feels very improbable.  The housing that sits behind the heat sink fan and over the heat sink was only secured with one screw (the other being broken). And the play was in that direction.  I've reassembled it and will try a print in an hour once I've got a couple of other things done and gone through calibration again

RE: help with lost adhesion please?

Thanks for not giving up.

There is a good printer hiding in there - somewhere.

RE: help with lost adhesion please?

Posted by: @justin-adie

[...] agreed.  But I have already done everything possible to the steel sheet short of putting glue on it - and it still does not explain why from one moment to another the whole print fails when there is no external intervening change - that feels much more like an alignment issue to me.  Also to note, as I said, I've tried the underside of the sheet too, to zero avail. 

Ah, OK. Just re-read the entire thread. Partway through there was a disconnect when you sounded like you were throwing in the towel. OK, so having confirmed that, we can focus on other stuff.

[...] I really don't understand why my posts are not appearing for other people.  I can read them!  I posted on 13/3 at 8h51 that I had followed your advice and scrubbed the board with fairy liquid.

I read your response as "you were going to try" but didn't see the confirmation. Apologies. Will quit pestering you about those.

The only pics I've seen in this thread is this one:

Is that still an accurate representation of your results? 

Looking at that pic:

1. Your Live-Z is OK though a bit too close. It really looks like the nozzle dragging on the edge.

2. Your extruder is putting out even extrusions. The inner extrusions look good.

Under normal circumstances, I'd say your Live-Z is too close. You're saying you have bad adhesion though, so I'm having a hard time visualizing exactly what you're seeing. Can you provide a few more pics.

3

RE: help with lost adhesion please?

Is that still an accurate representation of your results? 

unfortunately not.  that was a 'lucky' print.  The print immediately following that (two days ago?) entirely failed - zero adhesion: just a plastic ball being dragged around by the nozzle.

After rebuilding the extruder assembly and tightening it all down I ran the wizard last night and tried some calibration prints.  Zero joy so gave up.

I tried again this morning, calibrating from scratch again.  these are the results:

z-0.25

z -0.5

z-0.75

the plastic bunches around the nozzle on the last two prints.  looks like this (although obviously far worse on longer prints left to their own devices).

any advice appreciated.   to be clear the base is scrupulously cleaned with high purity IPA and is hardly handled (just the edges).  failed prints are removed by pulling on a bit of filament that's above the surface or a light push with a palette knife.  I try to avoid touching the print surface so far as possible and always clean it immediately after removing a print come what may.  

RE: help with lost adhesion please?

Posted by: @justin-adie

[...] z-0.75

Looking at these pictures, your results look very typical for a Live-Z that is too high (less negative) to closer (more negative). The filament is beginning to stick on the right. I would keep lowering (more negative) from 0.75mm slowly in 0.02mm increments. My own smooth PEI sheets are at -0.755mm and every sheet can be a bit different. Anything below -1.0mm is well within normal expectations.

I'd re-run the test and just keep lowering until (hopefully) you start to see the lines stick to the point that you can poke them with your finger and they'll remain stuck, or you start smearing the filament on the surface.

the plastic bunches around the nozzle on the last two prints.  looks like this (although obviously far worse on longer prints left to their own devices).

Any time the filament fails to adhere, it will be attracted to hot surfaces. That includes the nozzle and heater block. You can reduce this somewhat with a silicone sock and coated nozzle, but that is not necessary to get a good Live-Z.

any advice appreciated.   to be clear the base is scrupulously cleaned with high purity IPA and is hardly handled (just the edges).  failed prints are removed by pulling on a bit of filament that's above the surface or a light push with a palette knife.  I try to avoid touching the print surface so far as possible and always clean it immediately after removing a print come what may.  

That sounds good. Now that we can assume the sheet is clean, it's back to Live-Z. Again, your last picture at -0.75 looks like you're getting close. I suspect you'll wind up at around -0.8mm. I'd post pics of my own, but my printer is down awaiting a replacement LCD panel. You might bump the temps up a bit (10C) to increase PLA adhesion just for this test.

If this fails, we can eliminate a lot of items from consideration:

1. You've cleaned the surface.
2. You've tried Live-Z at a range of settings.
3. The extruder is pushing filament cleanly. Raise Z and extrude a bit to ensure the filament flows evenly out and coils underneath. If it is uneven or way off to one side after a few seconds, that can indicate a partial clog.
4. The printer seems to be moving cleanly in both X & Y. The lines look even though they're not sticking.

PEI doesn't normally "crap out". I've had 2 Prusa sheets and various other PEI sheets for nearly 2 years with no issues. In very rare cases, I've used a 3M 7445 pad (used for chrome and glass polishing) to scuff up the print area to improve adhesion. I'll re-read the thread, but don't recall you having done that.

RE: help with lost adhesion please?

thanks for the reply. 

I ran a few dozen calibration prints. these started working across a range of z-offsets from -0.75 and an ideal looking extrusion was around -0.55.  I re-ran that a few times and all was good.

I then ran a test print using the prusa logo with PLA.  Perhaps 20 minutes after the last calibration print, no intervening event other than cleaning the bed.  failed dismally (zero adhesion - balling of the filament around the nozzle).

started calibration again (from -0.8 to get some squidging) and I'm now back to no calibration prints coming out acceptably at all.  

I will have a cup of tea and start all over again.

RE: help with lost adhesion please?

Is your Live-Z drifting between prints? Can you confirm that the value used at the start of a print is the same you settled on for LivehZ adjustment?

There were reports of the Live-Z setting drifting if the on-board routine wasn't completed. Not aware of other issues with the firmware value going off.

I assume you've given the PINDA mount a close inspection and that there's no chance it's moving?

Did you resolve the hotend movement? It should be solidly in place once everything is tightened down.

Any possibility the X carriage has loosened? Check the screws on the back.

Trying to think of things that might cause physical movement. 

RE: help with lost adhesion please?

Is your Live-Z drifting between prints? Can you confirm that the value used at the start of a print is the same you settled on for LivehZ adjustment?

confirmed that the setting seems to be sticky between prints.  

I assume you've given the PINDA mount a close inspection and that there's no chance it's moving?

yes.  no chance. it's solid. 

Did you resolve the hotend movement? It should be solidly in place once everything is tightened down.

yes - that was yesterday afternoon's task. the hot end is solid now. 

Any possibility the X carriage has loosened? Check the screws on the back.

it seems rigid.  no play/wobble and calibration wizard reports squareness.  

I ran more calibration prints and found that there was a reasonably good result at 0.70.  a bit too squished for my liking but the 0.65 print was dreadful.

so thinking that I had calibrated correctly, I _immediately_ ran the prusa logo test print and it failed.  i cleaned up the bed, set z-offset to 0.80 and tried again.  this time i got a bit better results but I'd say the first layer was fairly awful.  your view? 

I know I'm bleating like a donkey, but all of this is so disappointing after having a week of perfect and trouble-free prints that were genuinely helping my daily life!   I've now spent more time recalibrating this printer than the total 'successful' print time of the machine.  

RE: help with lost adhesion please?

Posted by: @justin-adie

[...] I know I'm bleating like a donkey, but all of this is so disappointing after having a week of perfect and trouble-free prints that were genuinely helping my daily life!   I've now spent more time recalibrating this printer than the total 'successful' print time of the machine.  

I can understand your frustration. You may have noticed, I get testy if I detect any "I'm taking my ball and going home" or "Prusa sucks!" notes. Those really kill any desire to help people. I apologize if I misinterpreted your earlier comments. I'm glad you're sticking with this (if only because we're all locked up indoors for a while) and am happy to contribute in any way I can. You bought the right printer, but we need to figure this one out.

One thing that I don't think was verified before: Have you enabled 7x7 mesh bed leveling? Also, unlikely culprit, but have you done a PID tune? 

RE: help with lost adhesion please?

Yup.  7*7 is still enabled.  

I've not done a pid tune but I think the hot ends are pre-tuned no?  The test sheet suggested this. 

Could a hot end thermistor go so far out of tune over night that a successfully and stable printer becomes as it has?  I guess thermistors can break but then I'd expect all sorts of craziness with the heat.  

No harm in doing a tune though.   Will do and report back.

RE: help with lost adhesion please?

I'm just trying to think of things that could "drift". More checking than fixing at this point. PID tuning at your actual print temps is recommended.

Just to be thorough: Maybe check bed leveling? No idea why it would go bad overnight, but ensuring the bed is flat enough is a good check.

If nothing else, you're building a good list of things tried to update Support online chat.

RE: help with lost adhesion please?

it's PID tuning now; at 230C rather than 210C to cover the bases. 

I will check bed levelling after but doesn't the mesh bed levelling + PINDA automatically compensate for out of flat beds?  and alert if the whackiness is too far gone?

RE: help with lost adhesion please?

PINDA can compensate to a degree, but in the (extremely unlikely) event the bed warped, you could be out of the range over which it is effective.

This does however remind me of something: The PINDA itself is sensitive to temperatures. Did you by any chance experience a temperature change recently? I've got some notes here on my startup gcode that includes a documented routine to ensure that the PINDA is at a consistent temperature before doing mesh bed leveling. There are good reference links on that page as well. I've got a set of profiles here that you can try. Be aware that the warmup can take 5-10 minutes to start printing on a cool day. (There is a calibration item in the printer menu that is supposed to remove the need for this, but I like knowing it's done and heating the bed evenly before prints.)

I'd start by slicing your own 1st layer square (75x75x0.2mm) print with the PINDA warmup. Adjust your Live-Z using that print. Then try another print using the same startup routine.

Bit of a long shot, but it might explain your variations between prints.

RE: help with lost adhesion please?

PINDA can compensate to a degree, but in the (extremely unlikely) event the bed warped, you could be out of the range over which it is effective.

This does however remind me of something: The PINDA itself is sensitive to temperatures. Did you by any chance experience a temperature change recently? I've got some notes here on my startup gcode that includes a documented routine to ensure that the PINDA is at a consistent temperature before doing mesh bed leveling. There are good reference links on that page as well. I've got a set of profiles here that you can try. Be aware that the warmup can take 5-10 minutes to start printing on a cool day. (There is a calibration item in the printer menu that is supposed to remove the need for this, but I like knowing it's done and heating the bed evenly before prints.)

I'd start by slicing your own 1st layer square (75x75x0.2mm) print with the PINDA warmup. Adjust your Live-Z using that print. Then try another print using the same startup routine.

Bit of a long shot, but it might explain your variations between prints.

RE: help with lost adhesion please?

Posted by: @bobstro

 

(There is a calibration item in the printer menu that is supposed to remove the need for this, but I like knowing it's done and heating the bed evenly before prints.)

I have performed the calibration in the printer to which Bob is referring.  Note that this process heats the PINDA to (IIRC) 60 C.  I found it necessary to "tent" the extruder with folded paper to help the heating along so that it would complete in a reasonable time.

Having said that, I use a small variation on Bob's process to pre-heat my PINDA prior to mesh bed leveling and have experienced very consistent results in my first layer prints.

Regards,

RE: help with lost adhesion please?

Did you by any chance experience a temperature change recently?

yes.  the ambient temperature in my electronics lab increased from friday to saturday by 1.5C.  

I tend to pre-heat for PLA and then make a cup of tea and start the telegraph crossword before going back and starting a print in the morning though.  happy to try variations to that if advised!  
I will try calibration again tomorrow.  i've done the PID tuning and MBL and scrubbed the bed again with fairy.  fingers very much crossed.

RE: help with lost adhesion please?

I've raised the initial bed temp for PINDA warmup in my startup gcode  to 80C due to cold winter mornings here. I immediately set it to the printing bed temp after the PINDA hits 35C, so by the time mesh bed leveling completes and the nozzle completes heating, it's dropped into the 65-69C range , cool enough not the deform PLA (IME).

RE: help with lost adhesion please?

I did the PID tuning last night, as I mentioned.  after calibration I reran the first layer wizard and got the usual garbage.  So poured myself a g&t, drank it and went to bed.

in the morning I tried another first layer calibration (no other intervening external change) and it came out well.  So then I tried the Prusa logo and that was sound, albeit with some very minor stringing. 

so, chancing my luck i tried printing a 2h45 slice for an enclosure for a circuit board.  I got mixed results that I'd like your collective wisdom on, please.

this was the first layer.

the deformation on the left got worse but I was able to cut away the really gnarly stuff when the gantry was busy printing on the right.

here is the finished print

you can see the odd artefacts on the right and left hand printed areas.

on the right hand side:

and on the left

and here is the left hand piece, flipped over

so the good news is that I am now getting reasonable prints again.  For which many many thanks!  

the bad news is that with no intervening change it went from garbage to mostly acceptable.  That's worrying from an engineering perspective...  

and the active conundrum is that odd behaviour at the low and high x values.  at high x there looks to be nozzle scraping across the print.  and I'd say potentially the opposite on low z on the first layer but towards the end there are definitive scraping patterns across the diagonals of the left hand piece.  Thankfully this is not a decorative print and should be completely fine for my requirements. 

many thanks as ever!

RE: help with lost adhesion please?

I should add that I have not yet tried the PINDA heating suggested above.  happy to try that if that's a suggested solution to this inaccuracy. 

the interface says that temp calibration has to be at ambient of 21-26C.  Unfortunately I keep my house way below that!  I'm not sure whether it would affect the more generalised heating that is suggested here though?  If the induction probe is that sensitive, though, is that advice to move to a different design of z-probe?  

RE: help with lost adhesion please?

Whatever you did with your pictures - those last ones came through well.

If the spirit moves you, could I ask you to print another calibration square and post it?

From the picture, it looks like Z may be too low now and the nozzle is dragging.

It should be just low enough for the square to peel up as a sheet - not separate threads.