Strange change to Z-axis calibration between prints  

RE: SuperPINDA adjustment and test

Just few thoughts...

First, when asking for solution, it's always good if you also tell if printer is new (just assembled) and if your trouble started recently or it's there since beginning.

I assume all screws are tighten properly and that there's no wobble (print bed, Z and X axis). I also assume printer geometry is correct (all axes perfectly perpendicular!). Also check again, that SuperPINDA is fixed, so it can't move up/down.

First, make sure SuperPINDA height is correct. Manually move printhead down (by rotating Z shaft) to the point where nozzle just touches the print sheet. Now check distance between bottom of SuperPINDA and print sheet: it should be 1.5mm!

You can also check if SuperPINDA is working correctly (at least to some degree). Manually move printhead up so the distance between SuperPINDA and print sheet is about 5mm. At this point LED on the top of SuperPINDA will be lit. Now slowly move printhead downward and watch the LED. As soon LED goes off, measure the distance between bottom of SuperPINDA and print sheet: it should be about 2.3mm (for normal smooth PEI sheet). Difference of about 0.1mm doesn't matter and so distance of 2.2mm is still acceptable. More important is, that the distance to bed is constant (repeatable). That is, if you do that 100-times, LED must go off always at the same distance.

Now make 1st layer calibration. I expect Z-offset will be at about -0.800 for acceptable first layer. Repeat 1st layer calibration between each print. If it happens you need to make big adjustment, repeat process in previous paragraph and see if distance is still about 2.3mm. If not, I would assume SuperPINDA doesn't work correctly.

I hope that was of some help

Sure sounds like a loose Superpinda probe. Give the screw that's holding it in place a bit of a turn. And I assume the nozzle/printer block are tightly secured and not loose.

Pinda's tight. Thanks for the input.

RE: SuperPINDA adjustment and test

Thanks for taking the time to post!

My Mini's been working well for over a year, not a new assembly.

Everything is tight, and the SuperPinda is at the right height. The SuperPinda LED goes off at just over 2mm to the bed.

I did a first layer calibration. The z axis offset is -1.542. Though I have done first layer calibration in the past where the offset was closer to -0.750. Weird, eh? I'll remember to measure the height the SuperPindo LED goes off the next time I get bed crash, or if first layer calibration comes out drastically different.

RE: Strange change to Z-axis calibration between prints

Did you ever resolve this? I have exactly the same problem. There is no rhyme or reason to it and the axis fluctuates between two values that’s around .75mm apart. My bed is also getting a beating. I even upgraded to super pinda and it made no difference (it isn’t loose either).

Curiously I am also using Ocotoprint but am sure this happened before too.

It makes printing a very anxious procedure because I can print the same print 3 times ina  row no problem and then the next print the head will be too high. I’ll adjust the z-axis and then then retry and it’ll either work or the head will be too low and destroy the bed surface.

Posted by: @jrock

After first layer calibration my z-axis offset will be something like -1.500. 

I'll do a few prints and they are awesome.

I'll do another print, sometimes immediately after a successful print, and now the nozzle is too high. I'll do a live z-axis calibration and fix it up - now the offset is something like -0.750. Weird.

Then I'll do another print, and now the z-axis is way too low and the nozzle crashes into the bed. I can fix it with live z-axis calibration to raise it up (back to about -1.500) and the print continues and succeeds.

So this is pretty annoying. My bed is taking a beating from nozzle getting dragged along it.

There seems to be no rhyme or reason or pattern behind this.  It is suspicious that the change in z-axis height is about half - from -1.500 to -0.750 (approx) and it is always been about this much for the dozen or so times this has happened.

I am using Octoprint, but I am pretty sure this was happening before I started using Octoprint. I do not think it is temperature - I have the SuperPinda and I will often let things heat up before print start.

Anyone else have similar experience or know how to fix this?

Thanks!

 

RE: Strange change to Z-axis calibration between prints

No I haven't solved this, unfortunately.

I've gotten into the habit of always being ready at the start of a print with my thumb on the z axis live adjust.

I am pretty sure it is nothing mechanical or temperature related since the height change is so large and nothing is loose.

Because the height change is so systematically the same, even if the occurrence seems random, I suspect an elusive software bug. I downloaded the code to take a stab at understanding how the nozzle height is set during the initial purge at the beginning of the print, but eventually gave up after a few hours of inspection. I was looking specifically for an uninitialized variable but the abstraction in the code makes it pretty complex and time consuming to review. But who knows, if this is a rare problem maybe it's some other obscure cause.

RE: Strange change to Z-axis calibration between prints

Here is another update. 

I did a print last night turned out beautiful. No live z axis adjustment was necessary at the start of this print. It is currently set to -1.545.

The printer has been on but idle all day.

Today I start another print and BOOM. Nozzle crashes into bed.

So I cancel the print (through Octoprint), press the reset button on the printer and let it reboot.

I reconnect and start the exact same print a few minutes later and... print proceeds normally.

I think I'll try printing from USB and take Octoprint out of the equation. 

RE: Strange change to Z-axis calibration between prints

Curious if you have an update a year later? Did removing OctoPrint from the equation resolve the issue for you?

RE:

Me estaba volviendo loco, me pasa lo mismo hago 2 impresiones bien recién ajustado Z y a la siguiente impresión no está bien calibrada, baja o alta aleatoriame.

Tendrían que revisar esto porque es frustrante.

RE: Strange change to Z-axis calibration between prints

Posted by: @tumtumtum

Curious if you have an update a year later? Did removing OctoPrint from the equation resolve the issue for you?

Not the OP but I had a similar issue. I wasn't using OctoPrint. I did a factory reset and re-flashed the latest firmware and it seemed to behave after that...except once, after which a power cycle "fixed" it, and ever since it's been working great. Over a month since it first printed now, and I stopped expecting it to go to the wrong height long ago. It just...works. So the OP's idea this was an elusive software bug seems plausible.

But try upgrading to the latest firmware. If you keep having issues try a factory reset and re-flash the firmware.

RE: Strange change to Z-axis calibration between prints

Posted by: @jseyfert3

Posted by: @tumtumtum

Curious if you have an update a year later? Did removing OctoPrint from the equation resolve the issue for you?

Not the OP but I had a similar issue. I wasn't using OctoPrint. I did a factory reset and re-flashed the latest firmware and it seemed to behave after that...except once, after which a power cycle "fixed" it, and ever since it's been working great. Over a month since it first printed now, and I stopped expecting it to go to the wrong height long ago. It just...works. So the OP's idea this was an elusive software bug seems plausible.

But try upgrading to the latest firmware. If you keep having issues try a factory reset and re-flash the firmware.

Hmm, I did upgrade to the latest... but I don't think I've tried a factory reset. Not sure why that would help, but it's worth a try.

RE: Strange change to Z-axis calibration between prints

Posted by: @tumtumtum

Curious if you have an update a year later? Did removing OctoPrint from the equation resolve the issue for you?

Still a problem. I don't think I ever tried printing without Octoprint - I forgot 😀 and Octoprint is just so damn convenient.

RE:

Posted by: @jrock

 

Posted by: @jseyfert3

Posted by: @tumtumtum

Curious if you have an update a year later? Did removing OctoPrint from the equation resolve the issue for you?

Not the OP but I had a similar issue. I wasn't using OctoPrint. I did a factory reset and re-flashed the latest firmware and it seemed to behave after that...except once, after which a power cycle "fixed" it, and ever since it's been working great. Over a month since it first printed now, and I stopped expecting it to go to the wrong height long ago. It just...works. So the OP's idea this was an elusive software bug seems plausible.

But try upgrading to the latest firmware. If you keep having issues try a factory reset and re-flash the firmware.

 

Hmm, I did upgrade to the latest... but I don't think I've tried a factory reset. Not sure why that would help, but it's worth a try.

Me either, but when I told the support rep I had upgraded to latest firmware already they were insistent I do a factory reset followed by re-flashing the latest firmware again, to rule out a bad firmware flash. And both of those, not just one or the other. 

Be sure to note your statistics before the factory reset, if that’s something you care about, because the factory reset will wipe them. 

RE: Strange change to Z-axis calibration between prints

Sadly this did not work for me.

I did a factory reset and re-flashed with 4.4.1. All seemed ok for about 15 or 20 prints. I was starting to think all was well.

Then I got complacent and distracted and did not watch the first layer start. After about a half hour, I took a look and oh my it was gruesome. I'd guess only a few layers were lay down, less than 1mm, but man it was stuck to the plate real good and took some patient tedious work to peel it off with a sharp dental tool. I am certain it was the same problem - crashing into the bed. But this time I only noticed after a few layers.

I am sad and disappointed.

Undeterred I gave the bed light a sanding (yes it was that bad) and cleaning to remove the stuck ASA bits, put down a layer of glue and started a new print.

And now it is printing as expected. I did not even reboot the Mini or change settings at all. I've been using Octoprint so I would have disconnected but I think it is unlikely to be a factor as Octoprint just executes the commands in the gcode and the only custom commands I've used were the necessary ones for Bed Visualizer.

The mystery is unresolved, imo.

RE: Strange change to Z-axis calibration between prints

Just chiming in, I also ran into this weird issue a while back. Just a single time, on FW 4.3.4. Going through z-height calibration, I would make adjustments that would take mid-print, but the very next print it was as if they weren't saved whatsoever, even though the z-height setting was still at the newly adjusted value. Only fix for it was a complete power cycle, has worked flawlessly ever since.

I was also using octoprint.

RE: Strange change to Z-axis calibration between prints

I'm still getting this problem. Forgot to wait for my print to start yesterday and my poor print bed got quite terribly scratched. I'm almost at the point of just buying a new main board to see if it fixes the problem.

RE: Strange change to Z-axis calibration between prints

I've got more data.

I started a series of prints in the last few weeks. The z-axis offset started at -1.7 or so for the first print. No surprise there.

As I prepped for the next print, waiting and watching to see if I needed to do a live z-axis adjust... I did, but I adjusted only about +0.300. This is unusual. If I have to adjust, it is normally by +0.750 or so.

The next print I was again cautious. I found that I needed to adjust another +0.300 to the z-axis. And the next. And the next.

So for several prints, the z-axis had to be adjusted by approx +0.300 each time. When I got to around 0, I did a first layer calibration.

The calibration put me back to -1.7 or so.