imperfections

Hi,
My first print was great...
Then I tried 'fixing' things, like Al bed and X axis

Now, I'm having difficulty getting a consistent first layer.

for large X values, the layer seems to stick.
for small X values the layers seem to coagulate..blobs dribbling...

The Z plane seems to be flat, only 0.1mm compensation in one spot
(I've got the Al triangle - so 3 fixing points)
and zeroing Z seems reliable and repeatable...

I'm trying to print the front anti-vibration mount.
printing on to Kapton.

My initial thought was to ask if anyone knew of a picture gallery of 'mistakes' that I could learn from?

Or can anyone suggest a diagnostic methodology?

And....

How do you guys ever get a complete perimeter?
every time I print, the first side, typically X, doesn't stick?


DaveS

1. Have you cleaned the Kapton, with acetone or isopropanol, and are you careful to avoid finger-marking it after cleaning?

2. Does the gcode file print a sufficiently long (in terms of extrusion amount) skirt before it starts your print, to get the filament flowing well?

3. What first-layer temperature are you using?

Blobbing on the first layer typically means either that the filament isn't flowing well yet (in which case you can increase the minimum skirt extrusion amount), or the nozzle is too close to the bed.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42
Blobbing on the first layer typically means either that the filament isn't flowing well yet (in which case you can increase the minimum skirt extrusion amount), or the nozzle is too close to the bed.

Ah...that's probably it.
I've just noticed that the X-runner can't be made to clamp the X-beam...so the extruder is drooping and since I use the 'dragging on paper' method of Z=0 setting...that's going to make it too close until it gets to the area of the bed which is slightly lower....

Time for some 'persuasion' of the X-runner...
...maybe some redesign as well...I really would prefer a screw adjustment instead of the clamp-a-slot-in-the-same-direction-of-force technique.

thanks.

The X runner is not intended to clamp the X plate. The print head needs to be free to rise, so that the nozzle can ride over bumps. The x runner position should be adjusted so that the nozzle is vertical.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Mystery solved.
Loose X-runner - now the print head is vertical.
Loose top nut on the bolt near the hot bed connections.
and the main culprit..

I've been doing manual setting of Z=0 when the bed was COLD.

rechecking the Z plane with the bed at 65 shows that it swells by 0.5mm and is slightly barreled
(no change along the y axis, but curves on the X axis)

I have the axis compensation turned off for now (to avoid any possible z backlash issues)
and a manually adjusted bed that is as flat as I can measure.

I'm setting G32 Z0 at X=100, Y=100 when paper is pinched, and then Z up 0.2 and down 0.1

I've now got a snowman which is almost blob-free.
The main thing is that the first layer stuck properly, although the permimeter is rather thin and wide and incomplete.

This is with an unchanged snowman.g , so temps are 185, 65 with white PLA from filamentprint that came with the kit from RS.

at last I'm making progress again

DaveS

Quote
gnarly
rechecking the Z plane with the bed at 65 shows that it swells by 0.5mm and is slightly barreled
(no change along the y axis, but curves on the X axis)

That usually means not that the bed is barrelled, but that there is a slight wave in your X plate, causing the nozzle to rise a little near the centre of the X axis.

If you are still using the original MDF bed support, then the bed does tend to move with temperature.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

hmm, I'm using DD Metal Products Al bed, x-rib and x-plate, so that's less likely.

It struck me that if both sides/ends of the bed support and hot plate are rigidly bolted,
then the hot plate will expand more and cause doming...
looking up some numbers...for linear expansion coefficients in 10^-6 m/mK:
Al 22.2
glass 9
Cu 16
Kapton 20
plastics >40

so for 200mm length and say 45 degree rise, we get Al expanding by 0.2 mm with is unhelpfully close to
the tolerances we are trying to work to.
I guess that with MDF it is pliant enough for the bolts to bend it a bit when hot.
This might explain what force is delaminating the MDF around the screw holes (for those with MDF)
ABS bed temperatures are going to double the problem.

At least it is enough to say that it's worth doing your axis compensation at your working temperature.

returning to my increase in Z, my extra 0.5mm is obviously not coming just from thermal expansion
of the FR4/Cu/Al/Glass sandwich either.

time to sleep on things again.

DaveS

If you turn the glass over, does it dip in the middle, or does it still bulge?

If it is the glass that isn't flat, get some replacements from a local glass shop. If you think the glass is bending, try 4mm glass, which is about twice as stiff as 3mm glass.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
gnarly
hmm, I'm using DD Metal Products Al bed, x-rib and x-plate, so that's less likely.

It struck me that if both sides/ends of the bed support and hot plate are rigidly bolted,
then the hot plate will expand more and cause doming...
looking up some numbers...for linear expansion coefficients in 10^-6 m/mK:
Al 22.2
glass 9
Cu 16
Kapton 20
plastics >40

so for 200mm length and say 45 degree rise, we get Al expanding by 0.2 mm with is unhelpfully close to
the tolerances we are trying to work to.

No, that's not quite so. The expansion in the X and Y direction may be 0.2mm (over the 200mm lengths), but we are concerned with expansion in the Z direction, which will be over the plate thickness not its length. The XY expansion is a good thing, because it results in the print releasing when the plate cools down.

To see whether your glass is domed, put a metal straight-edge across the plate when it is at temperature and see if it rocks. It did slightly on my glass plate which told me that the plate was indeed domed (by about 0.15mm), but I am now using a 5mm aluminium build plate that is completely flat - and together with the aluminium bed support and X arm the nozzle is at a constant height in all places, though I have found that I occasionally have to tweak the bed levelling screws as they seem to go out of adjustment over time. I probably need to replace them with spring-loaded bolts - I currently use a bolt and 3 nuts (bolt head, heater PCB, nut, bolt shaft, nut, bed support, nut).

Dave
(#106)

@dmould Sorry, I conflated two observations, and I didn't explain very well.

In considering the expansion of the various materials, I should explicitly say, that the thickness of the glass and Al, means that the amount of expansion is well below the 0.1mm tolerance that we are working to.

Howerer, my other observation, was that the top plate is moving 0.2mm for PLA and 0.4mm for ABS
relative to the bottom support, when it gets hot.

if something expands and the ends are constrained (such as the Al plate and it's bolted side through the PCB to the bottom bed support (MDF/Al)
then it will try and move into the unconstrained direction (z axis) - i.e. buckle or dome.
In practice, I think the bolts will get pushed apart on the top, no longer be absolutely vertical
-------------
\______/
and just mangle the holes on the bottom MDF/Al, or loosen the bolts.

I was wondering if it was worth not fixing one of the 3 bed bolts to the bottom plate, but just let it rest on top,
(or working out some sliding fixing) to release the stress from the expansion.

I may be over-thinking this

DaveS

OK understood. Yes, the reason I want to change to spring loaded adjustments is so that the heater PCB is not constrained by tight nuts & bolts and so has a bit of wriggle-room (with the holes being slightly bigger than the bolt diameters) to expand laterally as well as vertically. As I am printing in ABS my temperatures are higher, I have changes of 90 deg C or more (twice as much as you used in your calculations). I just haven't got around to sourcing suitable springs yet.

Dave
(#106)

I've been waiting for the upgraded IR sensor, and so held off any calibrations.

But now the sensor has arrived (thanks dc42), I've started printing again.

Having waded through old posts on the forum I decided to tweak the extruder multiplier in slicer...
and set it to 0.9...
and verily...I have my first complete perimeter without blobs

and a perfect print of the sensorpad - only my 5th ever print

Dave

Quote
gnarly
..Having waded through old posts on the forum I decided to tweak the extruder multiplier in slicer...
and set it to 0.9...
and verily...I have my first complete perimeter without blobs

Nice! - I guess you made a single wall print and adjusted accordingly to get the measurements right and that is the correct way to do it AFAIK

..but a faster way to adjust the EM for blobs and other visual things would be to adjust in software on the fly with dc42 splendid firmware but you prolly know that

Quote
dc42
4. Print not going well? Adjust print speed and extrusion factor on the fly!

You can adjust the print speed on the fly by sending M220 S### where ### is the percentage of the normal print speed you want. For example, M220 S50 will print at half speed. You can’t override the maximum speeds with M220, you’ll need to use M203 to do that.

Similarly, you can adjust the extrusion factor by sending M221 S### where ### is the percentage of the normal extrusion amount that you want.

Even if you can’t save your current print with these commands, by experimenting with different values during a single print, you may be able to determine the optimum speed and extrusion factor for when you restart the print. Bear in mind that if you adjust the extrusion factor using M221, then the print end time estimated from filament consumption will be inaccurate.

[miscsolutions.wordpress.com]

Erik