Layer height and solid layer dragging

Hi,

We have problems with tuning our Prusa Mendel when it comes to layer height. We have a 0.5 nozzle (Arcol hotend). When we print with a layer height of 0.5, solid layers and prints in general look ok. As soon as we start lowering the layer height, the nozzle tip starts dragging through the extruded filament (see the linked image):

[nicolinux.org]

Using Slic3r we made sure that the filament diameter is set correctly. Speed is at the default settings (and also for the first layer at 30%). The Extruder microstepping is set to 1/16 and the DEFAULT_AXIS_STEPS_PER_UNIT is set as follows (Marlin firmware):

#define DEFAULT_AXIS_STEPS_PER_UNIT   {80,80,3200/1.25,544}

We have also used Rich's method of measuring the extruded filament and tuning the last value (with our old microstepping of 1/14) but it didn't yield good results. Which is not to say that it is Rich's fault - his method of measuring the filament extrusion height is pretty nifty
[richrap.blogspot.de]

We are a bit at loss here. We have no idea what could cause this issue. I have posted something simmilar here about failing tower calibration. I think the issues are related. Just mentioning it for reference:
[forums.reprap.org]

I know for sure that printing at lower layer heights is possible. 0.5 layer height isn't even the right thing to do, but that's the only height that works for us right now.

Help!
Thx.

Stefan

Your extruder is calibrated right? Meaning that when it gets told to extrude 100mm at what it does? If that is correct what are you printing with, ABS or PLA? What temps? Yu might also want to print a 20mm or 40mm square and measure, making sure when completed you measure the x,y, & z. If wrong then fix in firmware. Maybe your z axis thinks its going up less then it thinks.

I have a 0.5 nozzle and I have printed at 0.2mm layer heig quite a bit with great results.

I think the extruder ist calibrated right. When it is told to extrude x then it extrudes x. We are printing with PLA at 180° - 185°. We have also tested the x,y and z axis. They move by the right amount. That's why I think that we must have missed something more fundamental...

Thanks for the answer.

Stefan

Measured your filament in several places, averaged and inputed into your slicing program. Having the wrong number in there will have the slicing program calculate the amount of filament wrong, too much or not enough.

Ok, will do. I have measured it in three places. I have read that bad filament can mess up the print and this behaviour could explain it...
Are there any other settings or components that could be related to this issue?

thx

Try printing with honeycomb infill and choose maybe a 0.4 layer height

Shot in the dark here, nicolinux: Is your extruder block made out of PLA? Is your hotend set directly into it?

When I first started printing with my reprap, I could answer yes to both the above questions, and it caused me continual grief. As the hotend would get and stay warm, it would ever-so-slightly begin to melt the extruder block. The mounting point on the block would begin to sag down (maybe .5 to 1 mm) through the hole in the x-carriage. My first print would be ok, with some head dragging, and my second print couldn't start correctly because the head would be smashed into the kapton tape.

I note that in your "failed tower calibration" thread, you got a first layer that barely came out. That would be consistent with the problem described above.

If this sounds like you, switch to an ABS extruder block, add a fan, or find some other solution to keep your hotend in place. And please let me know. This problem was so frustrating, and knowing that 1) others fall into this trap, 2) my frustration eventually helped someone else out of it, would make my day.

Good luck!

(And FWIW, I was able to use my own fails-the-second-time printer to produce the ABS extruder block. The printer is very well behaved now.)

Thanks for the idea. But I guess that's not our problem. We are using Wade's extruder with an Arcol hotend. They are pretty far from each other. See the attached pic.

[nicolinux.org]

We were able to somehow aleviate the problem by adding an z axis offset (0.2) in Slic3r. Printing with 0.4 and 0.35 layer height does not exhibit extreme dragging through layers. I still don't think that it is the ideal solution, but it helps right now. We will print a new x-carriage (right now using Greg's) and we will swap the x-end-idler. We have also bought aluminium couplings for the z axis (with a built-in "spring cutout form"). We will see if the problem persists after upgrading all those parts.

Glad to hear things are improved. As I understand it, if you've got a .5 nozzle, you really won't be able to effectively print with layer heights more than .4 or so. The extrudate needs to be [technical term alert] "smooshed" a little to bind well.

While you say that you've confirmed that your z-axis moves by the proper amount, I wonder if you are nonetheless losing z-steps. I sometimes find that my printer behaves differently when printing than it does when just taking the occasionally manual command. You might let a failed print continue, and see how high you are at the end, or just increase power on the z trimpot and see if your problem goes away. Another thing to try would be to reduce z speed, jerk, and acceleration values, which would make them less likely to lose steps.

And if you're upgrading parts, let me recommend Greg's Wade's reloaded with the herringbone gears.
It's a big improvement over where we each started.

And I personally like Triffid Hunter's "E steps fine tuning", here: [reprap.org]. It couldn't hurt.

Edited 1 time(s). Last edit at 06/10/2013 02:38PM by Karmavore.

Hey thanks for the great tips! I have printed a few calibration "20mm-box" objects and measured them with a caliper. They are spot on so I guess the z axis must be allright. Printing at .5 was an just a workaround because we got that layer dragging with solid infill at lower layer heights. It seems that entering a z offset in Slic3r made a huge difference. The dragging is almost gone. As I said - I still don't understand why this helped. All articles on bed leveling mentioned that the nozzle should be very close to the surface with only one sheet of paper fitting between. That's how it is leveld. Now adding a z offset seems to counter the fine leveling.
However our heated bed also forms a dome at the center - another less than ideal case. We are using a mirror plate with PVC tape. The mirror plate is clipped to the heated bed. We suspect that the heated bed expands when heated up and has no where to expand to because it is restricted by the three screws that hold it in place. One solution would be to use screws with a smaller diameter than the holes on the plate so it can expand horizontally.

So thanks for the recommendation about the extruder. Right now we have a mix of x-end pieces by Ecksbot, x-carriage from Greg, Wade's extruder with an Arcol hotend. No wonder this mix is troublesome. We intend to replace every piece by the Prusa Mendel counterparts. The problem was that up until now we weren't able to print objects that exceeded a specific height. Also the Arcol hotend in combination with Wade's extruder seems to make it impossible to fit a fan that blows on the extruded plastic and not on the nozzle (found nothing useful for our printer on Thingiverse).

I wonder though how other people are able to print at extreme low layer heights as I still don't understand how a .5 nozzle is able to produce .1 layer heights. Rich wrote it up here:
[richrap.blogspot.de]

Thanks again - there is so much to learn. This project had its ups and downs. How many times did I hover on the "order" button on Ultimaker's site, ready to ditch the Prusa...
However - finally I could print my first designed object. So happy
[www.thingiverse.com]

Stefan

Your problem could be from the bed expanding a little when it reaches its operating temp. Try leveling your bed at operating temperatures. Also, adding a Z-lift might help with the dragging.

Should level the bed and set nozzle when both are at normal running temp. I didn't do that the first few times I used my printer and had nothing but issues.

Hi,
i got a very similar problem. My prusa i3 Prints look allright as long as i got no solid layers. So when i print the 20mm thin wall object, it comes out very nice, even with layer height 0.2mm. but if i try to print the 20mm solid box, the nozzle starts dragging through the solid layers. Not only on the Top, allready at the beginning of the print, so i got this problem with every print which has a solid base. extruder is calibrated, bed is leveled, all axes move exactly the distance which they should... i just cant figure out what the problem is.

so maybe we got the same problem, have you tried printing the thin wall object?

what seemd to solve my problem: adding another 5% to fillament width. before i had 3.0mm now i have 3.15mm and everything seems to work fine, just printing :-)

torbenberger Wrote:
-------------------------------------------------------
> what seemd to solve my problem: adding another 5%
> to fillament width. before i had 3.0mm now i have
> 3.15mm and everything seems to work fine, just
> printing :-)


Great! You'll find that different rolls of filament come in slightly different sizes, and you'll want to modify that setting based on measurements you make with your digital calipers. So this isn't a particularly good spot to just be putting in "made up" numbers.

What it sounds like you really should do is calibrate your E-steps value, which (if a 5% increase in your filament width is an indicator) is too high by about 10%. There are a lot of ways to skin this cat, but a link to my preferred method is above.

torbenberger Wrote:
-------------------------------------------------------
> [...]
> so maybe we got the same problem, have you tried
> printing the thin wall object?

Yes I did, and it came out right. The layer dragging exhibits only with horizontal, solid layers.

And like Karmavore said, entering bogus filament width is not a good idea. I find that rich's method of determining the E steps is really easy. You can also calculate it easily:
#define e_steps_per_mm = (200 x 1/(1/4) x 39/11) / (pi x 8) = 112.855 =~ 113

The first value describes the motor steps (200), then you have your motor microstepping (1/4). This is set via a jumper beneath the polulu driver. Then comes the gear ratio (39/11 for Wade's Extruder).
See here: [reprap.org]

We now have another problem... Printed objects have some very nasty "ridges". It seems like one set of layers is out of place and is not flush with the rest. It is too irregular for z-wobbling and I don't know what could be wrong this time...

[www.flickr.com]

i also think you should not just add some random number to the filament. the problem is i got no digital caliper ;-) so just to test if this could solve my problem i added this 5% and it came out pretty well.

my E steps are already calibrated with the method you mentioned. if i want it to extrude 100mm of filament, it extrudes exactly 100mm. so i dont think this is my problem ;-)
i'm going to get a digital caliper to measure my filament right.

Digital calipers are a must for,3d printing.

I found another clue for this issue. Maybe the nozzle tip which has a sharp edge causes the dragging. Of course adding the z-offset makes it less of a problem because the hotend will be further appart from the heated bed. Take a look at this blog post. Near the end where the gear is printed. He uses a J-head hotend and did touch-up a bit the sharp edges on the nozzle tip:

[blog.lincomatic.com]

And another idea. Maybe (like in our case) the heated bed is not perfectly flat and forms a dome at the center. While printing solid layers (especially the first ones), some layers have less space between the nozzle and the heated bed while others have enough space.

I still don't like the solution with z-offset, especially since all bedleveling tutorials suggest to have only a tiny amount of space between the nozzle and the heated bed, but it works and I am happy

Hi, I already had that problem in the past.
It happened to be as simple a setting my extrusion ratio to 0.9 (or 90%).
The problem occurred exclusively with pla.
This is because the slicer(slic3r, cura, ect) calutate with the density of abs plastic or something like that.
But even if I asked it to extrude 100mm and it extruded 100mm, it would overextrude with lower layer height.
Tough, I do not experience the same priblem with abs plastic(maybe because of the shrinkage?????)
Hope I helped you.
Have nice days guyz!!!1

Was it the nozzle itself hitting the work? That would be odd as it implies the Z-lift isn't enough to compensate for the layer thickness. Was there build up on the nozzle? check Z-stepping / offset?

If you had to increase the width of your filament past what the actual dimension of the filament is (measure in multiple places) - then you are jerry-rigging a solution which will cause problems in the future. If you're telling it you have more material going in to the print, than you actually have (ie 3.00mm set at 3.15mm) then chances are you're feeding too fast. That could cause build up and oozing in the hotend, and also backpressure on the incoming filament. that was the first problem I encountered. the excess would ooze during moves, and create a blob which would damage other parts of the print. Check that the volume of filament feeding in (diameter, length) is equivalent to what is going out, the printer has no pressure or visual guide to know that the glass state melt is correct volume, so it has to be juggled/calibrated manually.