[Solved] Z-level calibration varies with temperature  

RE: Z-level calibration varies with temperature

Prusa's got a good history of fixing these things. See those GitHub issues opened by Stahlfabrik for examples. 

RE: Z-level calibration varies with temperature

Posted by: @bobstro

Prusa's got a good history of fixing these things.

I hope so. This thread was started nearly 6 months ago.

RE: Z-level calibration varies with temperature

@loglow

i don't think that there will be any solution delivered by Prusa. And honestly i would ask myself, why i should pay more money if they would do it in any kind of upgrade kit. The delivery quality of my printer was just as good or bad as from a printer which only cost the half. I had to fix issues before i could use it and the z-level calibration is just crappy. The "G-Code-Waiting" only works when you start from the same point in temprature, so if your Printer was cold before or cooled down again. And the waiting really annyoing me, since i'm usually printing small things and to wait there for 3 minutes, before the start, to check if the calibration is really working is just stupid. And the scratches in my springsheet reminds me every time on it.

Use a self made "hacking" solution which is voiding the waranty is nothing i would like to do, when i'm buying a Prusa. The support was also not helpfull, stateing that there will most likely be a solution for it by the developer or coumminty soon. I couldn't even return the printer (everything took longer than 14 days).

So why should i buy a Prusa?

• Quality: The mini had delivery / assembly mistakes and operational issues -> no better than other brands
• Community: other communitys are just as helpfull
• Support: eager but not helpfull at all
• Price: 1,5x - 3x as much.

I was a big fan of Prusa before, but there continuous show of unwillingness just cured me. As long as they are not changeing the behaviour i could absolutly not recommend Prusa to anyone anymore.

RE: Z-level calibration varies with temperature

It’s the same for me. Can’t recommend prusa any longer. I suggested a preheat routine on GitHub which would have fixed this issue 100%. It was denied months ago due to „safety“ concerns. Total BS in my opinion. 

makes me really angry. I wasted 100s of hours to get consistent results on the mini. But this was ok. Their reaction on GitHub was not. 

RE: Z-level calibration varies with temperature

I don't understand Prusa's lack of urgency about this.

Isn't this an issue that every single customer of the Mini is facing?

How did this issue even get through development/QC? Am I missing something here?

RE: Z-level calibration varies with temperature

maybe because it can be fixed by simply adding some lines to the start GCODE (heat up the bed and wait for some seconds) I agree that it is strange they swapped the PINDA for a MINDA (and thus creating this issue) but that is again something they can fix in the MK1 of the printer (it seems the buddy board has an extra pin...

RE: Z-level calibration varies with temperature

I have been using the added GCODE for months. I turn on the printer, start printing and get a a first layer like this every time. There are a few lines that result from the way it fills in around the round nubs. The center is filled in pieces. That is not a result of first layer height. The slicer could use a better algorithm for filling in large areas with small ones off the sides.

I have not adjusted my first layer height in probably a month and a half and have printed almost every day since then with both PETG and ASA.

I have had some issues when switching from ASA back to PETG with under extrusion. They were solved by cold pulls with a piece of PETG at a high temp of 275 and a low of about 135 to clean out the remaining bits of ASA.

Regards,

Mark

RE: Z-level calibration varies with temperature

Posted by: @marklgoldberg

I have been using the added GCODE for months. I turn on the printer, start printing and get a a first layer like this every time.

How stable is your ambient temperature? With a recent hot spell, my room temp went to about 33 deg C, now it's back down to about 25 deg C.  I have to adjust the Z height accordingly, even with Gcode to wait for stabilization.

I guess people with HVAC are more likely to say they have no problems, vs people with no HVAC. It's easy to say "there's no problem" if you don't take all use cases into account.

RE: Z-level calibration varies with temperature

@bobcousins

You are correct, I am in Arizona and have HVAC. It also is in an enclosure. I did not state that "there is no problem" and I certainly believe some do have a problem. Someone else stated "Isn't this an issue that every single customer of the Mini is facing?". That is not correct either.

My point is that is does work for a lot of people. Our HVAC is set to two temperatures in summer and winter, about  3 C apart. I would estimate with some additional variation the most change we get is about 4 C in the room the printer is in, about half of the variation you see. Maybe that bounds the issue somewhat. It is also possible that some MINDA sensors are more temperature sensitive than others. There is always variation in manufacturing. Maybe I have a good one.

Regards,

Mark

RE: Z-level calibration varies with temperature

Hello Everyone, @markgoldberg, @bobcousins, @oxygen, @loglow, @avrcore, @bobstro, @texy, @thunder1312, @dave1234, @tracy, @vital-g, @malcmcm, @jan-p29, @benji, @ryanoserous, @joantabb, @rcfishhunt. 

Thank you all for commenting and keeping this conversation moving along. It seems we've found two solutions; a temporary one (Pre-Print GCODE) and a permanent one (P.I.N.D.A). Because this is a significant hardware change, it's unlikely PrusaPrinters (Care to comment @jakub-d, @josefprusa ?) will want to perform this change before the Mini Mk2 as most fixes/changes they incorporate revolve around printed part revisions and/or software updates.

This doesn't mean we should stop, if anything it means we need to keep pushing and asking for a solution because if we don't it might never happen, even on a Mk2. I urge you all to please fill out the PRUSA MINI Survey! (Check your spam email folders in case you missed it), please be honest yet kind, communicate effectively and perhaps we can get a solid solution. 

In the mean time, since I've referenced two Admins, I'll archive these forum pages in the unlikely event they get removed.

I can't thank you all enough for your help through this, because of the advice of many of you I have been able to print hundreds if not thousands of COVID-19 related PPE (Mask Extenders) that is in use today in nearby hospitals and businesses, almost 1/2 of which was printed on a Prusa Mini (PETG). 

Archive links:

Page 1: http://archive.is/nlg01

Page 2: http://archive.is/0PVDK

Page 3: http://archive.is/3XUFP

RE: Z-level calibration varies with temperature

Quick question for those still experiencing issues even with the warmup pause: Are you willing to try heating the bed to 80C for the warmup, then setting it back to printing temp after leveling? I found that a higher warmup temp gave more consistent and faster results for my PINDA v1 issues across seasonal changes. 

The bed temp usually drops from 80C to 65C very quickly, so prints are unaffected.

RE: Z-level calibration varies with temperature

Just a note that a fix does not necessarily require a PINDA or other hardware. If Prusa modifies the M109 gcode command to simply wait for a nozzle temp without heating, the temp of an unheated nozzle (block thermistor) closely mirrors that of a PINDA. This way, positioning the nozzle close to a heated bed while cold, waiting for the nozzle temp to hit 35C, then doing leveling and deferring nozzle warmup (usually fast) should work.

I don't have a Mini to test on, but it looks workable from testing I've done on a Mk3.

RE: Z-level calibration varies with temperature

A hardware approach is a much better solution. If it wasn’t why did they introduce the temperature compensated Pinda for the mk3. 
Texy

RE: Z-level calibration varies with temperature

Using the block thermistor is a hardware approach, just a different one.

The Mk3 cost nearly 3X as much, so I'm sure many components could cost more while still maintaining profits.

Using the block thermistor is viable on the Mini due to the program space available. The Mk3 is very tight on space for new code.

If anybody has Octoprint running on a Mini, I'd be curious to see if a plugin could be used to emulate this fix.

RE: Z-level calibration varies with temperature

@bobstro

First I want to thank you for coming up with a fascinating alternative to the long wait for the current GCODE solution. 

However there is a small issue with thermal expansion due to the ~200C change in temperature at the hotend. If it shifts by as much as 0.05mm, it could cause print issues. Will it change the same every time? depends on filament and other factors.

Also there is a much easier solution. Just use the $2 more expensive probe (P.I.N.D.A.), or even letting us mod the printer without voiding the warranty. Such as jerry-rigging a thermistor to the last pin on the M.I.N.D.A. connector and attaching the thermistor to the M.I.N.D.A. probe.

I'm simply convinced that GCODE options are band-aid solutions that won't hold up over time. While a GCODE solution is what we're stuck with now, I'd like to see if we can get a permanent 99% reliable solution.

RE: Z-level calibration varies with temperature

Posted by: @drkmstr

[...] First I want to thank you for coming up with a fascinating alternative to the long wait for the current GCODE solution. 

It's very much like the situation we found ourselves in with PINDA v1. I've found that warming up the printer has benefits overall, so while I could upgrade to a PINDA v2, I find no need. I'd rather ensure the best possible print conditions by waiting a few minutes up front for everything to warm up.

However there is a small issue with thermal expansion due to the ~200C change in temperature at the hotend. If it shifts by as much as 0.05mm, it could cause print issues. Will it change the same every time? depends on filament and other factors.

I'm not sure I'm understanding what you mean. The heating of the nozzle after mesh bed leveling is no different if the sensor is at 35C, regardless of whether the nozzle is at 35, 160 or 230C. So long as the PINDA/MINDA itself is at a consistent temperature, results during MBL should be consistent. Or have I missed something?

Also there is a much easier solution. Just use the $2 more expensive probe (P.I.N.D.A.), or even letting us mod the printer without voiding the warranty. Such as jerry-rigging a thermistor to the last pin on the M.I.N.D.A. connector and attaching the thermistor to the M.I.N.D.A. probe.

Every time I sit in my car and look at all the blank switch panels that could have been installed to resolve problems that every driver encounters, I feel the same way. 😀 

I'm simply convinced that GCODE options are band-aid solutions that won't hold up over time. While a GCODE solution is what we're stuck with now, I'd like to see if we can get a permanent 99% reliable solution.

Hey, no arguments there. If Prusa elects to change the approach they use on the Mini, I certainly don't object. Meanwhile, rather than fixing blame, I'm concentrating on fixing the problem. Griping at Prusa is all fun and games, but for those who are satisfied or are simply looking for a fix, it's worth looking at working with what we have available today.

The Mini hasn't been out a full year yet, so many users have yet to use it in winter temperatures. As with the Mk3, those in colder climes may find that a warmup period ensures an evenly heated bed as well as providing an opportunity for warming the MBL sensor. Even if Prusa upgrades the Mini with a PINDA and implements temperature compensation in there code, you may still want to wait for your bed, nozzle and MBL sensor to warm to a consistent temperature. Understanding the gcode "workaround" (I call it print prep) is still worthwhile.

RE: Z-level calibration varies with temperature

Does anyone know what Prusa's current position on this issue is? We can speculate in the forum what Prusa will and won't do all day. Has someone contacted support recently on this issue?

Regards,

Mark

RE: Z-level calibration varies with temperature

Ugh... "their code". I hate the short edit window. I know these things are frustrating. It would be awesome if everything worked perfectly, but 3D printing is still an evolving technology and hobby. I'm reminded of one of @joant's stories of problems dealing with MINTEMP errors on Mk3s on cold winter mornings. One sensible individual solved the issue by setting their cup of tea next to the lowered extruder for a few minutes on cold mornings. I'm not British, but it does seem that a lot of problems can be solved with a bit of patience and humor.

RE: Z-level calibration varies with temperature

Posted by: @marklgoldberg

Does anyone know what Prusa's current position on this issue is? We can speculate in the forum what Prusa will and won't do all day. Has someone contacted support recently on this issue?

Regards,

Mark

Yep, I did last week. They said the MINDA was faulty and sent a replacement. 

I fitted it a couple of days ago, and no improvement.

RE: Z-level calibration varies with temperature

Posted by: @marklgoldberg

Does anyone know what Prusa's current position on this issue is? We can speculate in the forum what Prusa will and won't do all day. Has someone contacted support recently on this issue?

The "official" (public official in any case) response is on the GitHub issue page. I believe #636 is the most active. #561, #367 and and #309 seem to be duplicates.

Edit: Worth noting that some users assuming this is a MINDA issue have actually found other issues.