Trouble with prints staying down.

I’m a newb - somewhat competent with good guidance.

I’ve been going nuts printing things with decent results - and learning as I go.

However, I’ve been trying to print a 6x6 piece - and I just can’t keep the ends from eventually coming up sometime during the print. I’ve tried Abs and pla - still same results.

I have a Taz5 with the pei bed - i’ve been using the Cura stock setting for standard definition print for each filament. I have leveled the bed using a dial indicator - however with all 4 corners zero’d there seems to be bow in the center of the bed that goes from front to back. The bow pops up by about .010-.012". or .25-.30mm (i’m still trying to get the metric lingo down). - I put a straight edge on the bed, and can’t really see it, though this should be easy to see. So maybe it’s the X axis bowing in the middle or the slop that I do have in the x axis bearings.

I set my z height off the center of the bed, which is the highest height. It’s set about .005" or .13mm. I ran the dial indicator around what would roughly be the 4 outer corners of the 6" square to see the bed height difference and it’s about .005" or .13mm lower than the center of the bed at all 4 corners of the 6".

I’m kind of at a loss, I’d like to figure it out, as I have a ton of things I would like to make that will even be bigger than the 6" square I am making now with poor results. Any help or guidance would be appreciated!

I’ll post a pic.

Here is a pic of the print starting to lift. This corner is always the first to lift, and the worst.

This is the part I’m trying to make - I let it run overnight, didnt catch it lifting until it was done in the morning.

For anyone interested - It’s a collet tray, for R32 collets.

There “IS” an apparent bow with the taz, but it isn’t due to the bed flatness. Its actually caused by rod flex. The bed is heavier, but supported at wider points. The carriage is lighter, but acts as a point load, which leads to a .010 ish deflection from midpoint to edge that shows up as an apparent hump. There are modifications you can do to alleviate that later. For now, you want to concentrate on getting a good contact pattern and first layer. For abs in particular, start with 240c for the nozzle temp and. 110c for the bed. You want a 0.40mm or thicker starting layer. If you are too close to the bed, the abs will not bind as well to the pei sheet and it will lift. I am not sure why. The thicker starting layer works better. Bed leveling is key. If a corner is lifting, lower it if the starting layer looks too thin, raise it if it isn’t adhered well.

Run the bedcalibration Gcode multiple times until you get a good contact pattern across the entire print area.

Printing with a 5mm wide brim layer also is a good idea.

In my experience, using less cooling fan will reduce warping.

And lowering your infill percentage should also help, but you may need a few more top layers to bridge the gaps and get a smooth top surface.

I’d expect PLA would give you less warping than ABS. I don’t think I’d even try that size part in ABS without having a heated enclosure (which I don’t have).

But piercet’s certainly right – the very first thing to do is get the first layer stuck down as well as you can.

Thanks for the feedback! The part in the pic is abs - Using PLA, I ended up with the same amount of lifting and cancelled the print.

I’ll try some calibration gcode tomorrow after work and report back.


Here are some of the ABS settings that are stock in Cura and what was used on the print coming up in the first photo.

ABS
initial layer height .425
initial layer width .125
bottom layer speed 15

bed temp is 110
hot end temp is 240

layer height. .25
fill density 20%

print speed 50
filament diameter 2.85 ( not sure why this is not 3mm stock)


I pulled the outer extrusion ring off the bed and measured it. This is the outer pass it makes around the part to prime the nozzle i assume. It measure .018" or .46mm. fairly consistently around it. and about .026" or .66mm wide.

Again, time for bed, but I’ll post more up tomorrow, and follow any additional guidance I see here.

Thanks again!

Looking at your first picture again, I see that the curling up is happening with just a few layers there. So the infill issue I mentioned isn’t going to be a factor in the curling up at that point. As the build gets taller it may be, but for now I’d say just ignore that factor.

[edited:]
At some point you may want to change the design to have the hole walls almost completely isolated from each other except on the bottom and top surfaces (with sturdy-enough cylinder walls). Therefore having almost no infill at all between the holes trying to shrink and pull in the sides and lift up the edges. Just enough infill to support the top surface. But again that’s for later, not now.

I don’t print in PLA, but these ABS tips might help. Give the Z-endstop a quarter turn so the nozzle is a tad lower. This should help with bed adhesion. Alternatively you can lower the nozzle height through the slicing software, look for a setting called Z G-Code Offset… try -.1 or -.2.

Another trick with ABS is a brim… 3-5mm should help seal the edges.

Give the bed a good wipe down with some Isopropyl Alcohol. Can also try a light sanding with a fine grit sandpaper.

Can you post a pic of the bottom of the print?

Lastly, it could be the PEI. Exhaust the suggestions everyone has provided then call LB support. Another user on the forum had a similar issue, and a replacement bed was “night and day”.

Hope these help.

This number should be exactly what your fillament that you are using on your particular printer measures in diameter. A great many of the 3.00mm fillaments on the market actually measure 2.85mm, some measure 3.12mm. It varies by manufacturer, quality control process and whatnot. it can sometimes be variable throughout a given roll of filliment. If you specify 3.00mm, but your fillament is actually 2.80mm in diameter, you are underextruding by that 0.20mm worth of filliment constantly. If you have 2.85 specified but actually have 3.10mm in the machine, you are constantly overextruding, etc.

You almost never want any fan with ABS, if that is on, turn it off.

PLA isn’t prone to lifting due to contraction. If you are getting similar lifting, chances are that bed corner that is lifting is off.

Edit: recommended above, still try it…

Try sanding the surface of the PEI bed with 2000 or 3000 grit sandpaper. You aren’t looking to do much, but rather rough up the surface. If you do this, wipe the bed down several times with Isopropyl Alcohol (99% or better) before you print. It’s is very important to remove the part from the newly sanded bed at the proper temperature and not let it cool down to room temp.

Ditto what piercet said, but I wonder if you have a draft of air near your printer or is the room cool? ABS is very sensitive to anything that might cool the bed down. I even used an enclosure in the summer time when printing large prints. The enclosure was simply a light blanket over the machine with holes cut for the filament and control panel.

Here are a couple pictures of the underside - this is the abs part that was shown lifting in the first picture.

Not sure how to interpret these settings yet - but do recall seeing the fan running at 40% whenever I seem to be looking at it.
(Stock setting for abs).

  • fan full on at height = .5
  • fan speed min = 40%
  • fan speed max = 60%
  • min speed(mm/s) = 10

Looks like I have a bunch if things to systematically work on tonight. Will work through each of these and report back - Thanks again for the support!


I would kill the fan entirely for ABS except for bridging and very small layers. It will cause warping and lifting otherwise.

Ok, quick question.

I’m trying to download the calibration gcode from Lulzbot website. However; it doesnt seem to want to open in Cura. I know it’s already sliced and gcode, but how do I know what it is actually supposed to print?

Just save to SD and print it regardless?

It prints an X shaped outline twice, along with a Plus shaped middle section. It might also do an outline pass around the whole thing. I forget.

Something else that should help, after you’ve done all you can with getting the first layer to stick well:

Round the four sharp external corners (as seen from above the part), if you can. Make them the largest radius that fits your requirements for the part. It’ll help some with the lifting problem. Sharp corners are the hardest to keep down.

Ok, off to a late start.

Cleaned up the bed with Isop Alc…

Heated everything up for Abs 240 & 110 and have been running the bed calibration. and making tweaks.

I’ve been measuring various heights of the extrude layer and using that as my guide. As of right now - it’s a heck of alot more consistent now - but there is no way I want to put a dial indicator on it - With all the adjustments I made there is no way the dial indicator is going to indicate anything sane.

I am getting the impression that I am chasing repeatable and explainable perfection but through a combination of system errors nothing maybe repeatable at this level of measurement I’m trying to measure to…


This is it for me tonight - I’m going to try another print where I am and turn the fan off completely and see where we are at.


But - does anyone know how high these test extrudes should be coming out for the calibration in abs? I got to .23mm in the middle of the bed. and about .28mm at the highest in one of my corners. ??

When printing ABS the only fan that should be on is the heatsink fan. Never any fan with ABS, and read my other posts about drafts and enclosure.

Ok, ran the same print last night - the differences being:

  • cleaned the bed well with Iso Alc.
  • Turned off the cooling fan completely
  • leveled the bed several times using the Calibration gcode. I took the time to measure the height of the extruded lines around the table to assist. I was able to get the bed pretty darn close this way.
  • I slung a blacket over both sides of the taz frame, kind of leaving the center open, hopeing to eliminate a little bit of draft.

All looked decent when I went to bed, but woke up to the corners pulling again - but this time the corner that has always pulled the worst, was the corner that pulled the least( it was also the corner best protected by the blanket now) The skirt extrude around the exterior measured about .43mm - .47mm in different places. Initial layer height was set at .425 in the stock abs profile.

Note - I had the initial z height set lower - but what i experienced was that as the first layers went down they smooshed so much that the layer being laid down pushed up the layers it was being laid next to . I raised it until it stopped doing this.

Looking at the parts Lulzbot made it sure looks like their initial layers are very fat and spread apart more than mine are… hmmm could
be nothing… I’m to new to know.

Anyway come morning all 4 corners pulled - but the part was better overall - but not truly acceptable results.

So - I have a couple more things to play with -

  • going to make a drop over cardboard box cover to ghetto hot box it
  • going to try a brim to see if it will help hold the ends down

There is still a ton of suggestions here for me to work through - thank you to everyone who has been helping.

piercecat love the open rail conversions you have done - I’m probably going to venture into those mods next.

Will report back with more results as I go -

Glad you like them, let me know if you have any questions!