Bed calibration, Z, 1st layer not sticking/too flat
Just assembled an MK2 yesterday.
I'm having problems with printing a proper extruded filament, and the filament isn't sticking, raising at the corners. I'm printing a 150x150 single layer file, so I can play with the Z.
The following is a pic of the overly flat filament that isn't sticking:
And another:
You can see that the extrusion is too flat. It will print 4-5 rows and then lift up the entire thing.
Here's another print, after thoroughly cleaning the bed with rubbing alcohol and recalibrating:
Same thing, different angle:
After moving the bed temp to 65, it almost seems worse in that it will print only 2-3 rows before lifting everything up.
I've upgraded to 3.0.4 beta firmware. I've gone through and ensured that the threaded Y rods are bottomed out in the Z frame, the space between the inside rear Y frame and the Z frame is exactly 100mm. If I print from SD card the Checkpoints1and9.gcode, the PINDA head is perfectly centered inside point at the front left (during the first two up and down motions but while waiting for user input it shifts slightly to the front left); it is perfectly centered inside point at the back rear.
I've used the Settings->Bed Calibration. On the very last pass, it appears to show Z levels at each stage? I'm getting some very different ones at each calibration point:
If I select settings->Calibrate Z, the readouts are the same. I've reset the bed calibration and redone the bed calibration.
Does that mean that sites 6-3 are completely erroneous?
At this point I can't get the V2Calibration.gcode to print properly. It'll lay out a perfect line at the front, a good line at the very back but will fail to stick by the first corner. I've wiped the surface clean first with Windex, then with rubbing alcohol. With glue, it'll stick better, but not perfectly.
Interestingly enough, when I first printed something after just finishing the build, I dropped the PINDA probe down below where recommended in the instructions (i.e. 1mm below the lower reference of the PINDA cable guide on the extruder head) and this is what we printed:
Why in the world would having the PINDA probe down lower yield such good results? And what is causing the waves in the top right?
Re: Bed calibration, Z, 1st layer not sticking/too flat
Do you clean your bed, if so with what? I had the same issue before, and got rid of it when I stopped using the window cleaner that I was using (after watching instructions videos). Some window cleaners contains stuff that make the adhesion worse. Now I'm only using pure alcohol and it sticks really nice.
Re: Bed calibration, Z, 1st layer not sticking/too flat
Hi Curtis
If the probe is set too high then the nozzle will hit the bed before the probe triggers. When the nozzle hits the bed, it pushes the bed down and therefore prevents the probe triggering. The firmware is programmed such that it will not go too far downwards to prevent damage.
Sounds like that may be happening, especially in the middle range and it would also explain why you had a better layer with the probe lower.
Every probe and bed are slightly different, so the instructions have to cater for most setups.
Peter
Please note: I do not have any affiliation with Prusa Research. Any advices given are offered in good faith. It is your responsibility to ensure that by following my advice you do not suffer or cause injury, damage…
Re: Bed calibration, Z, 1st layer not sticking/too flat
Do you clean your bed, if so with what? I had the same issue before, and got rid of it when I stopped using the window cleaner that I was using (after watching instructions videos). Some window cleaners contains stuff that make the adhesion worse. Now I'm only using pure alcohol and it sticks really nice.
I was using Windex. I've been using household rubbing alcohol applied with a makeup applicator. Happy to use something else if it would help.
If the probe is set too high then the nozzle will hit the bed before the probe triggers. When the nozzle hits the bed, it pushes the bed down and therefore prevents the probe triggering. The firmware is programmed such that it will not go too far downwards to prevent damage.
Sounds like that may be happening, especially in the middle range and it would also explain why you had a better layer with the probe lower.
Every probe and bed are slightly different, so the instructions have to cater for most setups.
Thanks, Peter. That makes sense. I moved the PINDA probe about 1mm below the bottom of the guide and recalibrated the bed and Z. Now, it is so close to the surface that it catches on any filament that might have lifted, so that wasn't the trick.
I did another V2Calibration test. It runs a big, fat filament at the front left, but then those at the very back are so paper thin as to be nearly invisible. In fact, the brass feed guide is chugging and slipping as it can't feed the filament.
(ignore the feathered remains of the previous way-too-thin prints at the front right of the bed. And ignore the gouge leading to the back left--tech support was saying Z needs to be lower if it isn't laying down. Turns out lower was too low. 🙁 )
Re: Bed calibration, Z, 1st layer not sticking/too flat
Did you adjust the Live Z value after moving the probe?
You may want to do a Factory Reset and then adjust the bed again. To do that, click and hold the selection knob and then hit reset. You should hear an audible beep to let you know it reset.
Re: Bed calibration, Z, 1st layer not sticking/too flat
Did you adjust the Live Z value after moving the probe?
You may want to do a Factory Reset and then adjust the bed again. To do that, click and hold the selection knob and then hit reset. You should hear an audible beep to let you know it reset.
Thanks for responding, Anderson. I moved the PINDA probe so the back of the probe is flush with the bottom of the cord cord slot extrusion, as suggested in the manual and in the forums in the build section. I've done the factory reset as you suggested, then Settings->Reset bed calibration then Settings->Calibrate bed.
When I do Calibrate Z, the Z values displayed are:
What is with these numbers in the middle and back? Is that a bad PINDA probe, maybe?
Printing Checkpoints1and9 prove the bed is perfectly aligned, as before. Ran V2Configuration from the SD card 2x, each failing to stick so adjusting greater Z. The third pass, the first 5 long horizontal lines stuck (albeit seeming somewhat thin), but the 6th didn't stick around the corner. My Live Z is currently at -0.957mm. I worry about going too much deeper with the Z.
Here is the result of the V2Configuration showing the last pass failure.
And this shows how very thin the peeled up extrusion is:
Should it be this thin? It is rather translucent.
Re: Bed calibration, Z, 1st layer not sticking/too flat
Sounds like a deep Live Adjust Value... I've had the same problems as you but I never exceeded -0.500mm
It sounds like your build plate isn't level.
Have you tried the newest firmware?: http://shop.prusa3d.com/forum/original-prusa-i3-mk2-f23/beta-fw-3-0-4-with-automatic-calibration-point-ali-t1108.html
My solution when I had the same issues was to rebuild the Y axis, and making sure that the Y rods really sits in the frame. I had to sand away some paint to get them all the way down. Although I still had the issue with the filament on the left side of the bed not being squeezed to the plate enough, if I tuned the live Z calibration to squeeze it enough on the right side. After the new firmware and the automatic calibration point align function, the probe was centered in all 9 circles and the bed leveling seems to work as it should now.
Anyway you are way to low for sure, what happens if you raise the live Z value to around -0.4mm? Doesn't it stick at all then?
Re: Bed calibration, Z, 1st layer not sticking/too flat
I ran into a similar situation when building mine and I had to tear apart the Y-axis and rebuild. I thought I had everything right but the Y-axis was not exactly square. Once rebuilt and measured, re-measured and measured again I am able to print flawlessly.
In theory the bed-leveling/mesh-leveling should compensate for this but it hasn't done the best job of fixing this problem.