Extrusion calibration

[Deereengineer]

So I was working on calibrating my machine per the instructions at the beginning of this thread. I measured my filament, averaged about 1.72 mm which I entered . Extrusion multiplier set to .9 and the manual width set to .40 . Cube side wall came out to be .05 mm thick. In order to get down to a .4 mm thickness I had to back down to a multiplier of .7 .Then when I printed something else it came apart with no adhesion between layers . Any idea what I'm doing wrong?

[Jules]

My best guess would be the gap between the bed and plate is still too large.

I discovered something when I switched to the dual extruders - if you have the gap set correctly to start with, you can pretty much leave the filament diameter at 1.75mm and leave the multiplier at 0.90 provided that you have that first layer gap set correctly. (Yep - repeated that twice.)

Do this calibration first:

viewtopic.php?f=3&t=3124&p=20258#p20258

Then try this one with 1.75 mm filament diameter and 0.90 multiplier and see if you get a better bonding on your print. You should still get 0.4 mm thick walls on the print. (If not, then run this calibration with your actual diameter.)

[Deereengineer]

I thought that the z height would only affect the bottom couple layers. How does this affect the adhesion of the print many layers up ?

[Jules]

If your first layer doesn't go down correctly, the rest of the layers are going to have a hard time sticking to it, and might not bond correctly. Without going into too much detail, ('cause i've got something else cooking at the moment) , the ideal profile shape for an extruded thread is a slightly squished oval, with a ratio of 1.8 width to height. (Ed's got some good explanations on it around here somewhere, but I haven't got time to go search for them now. )

I suspect, (since i haven't seen an actual picture of the threads), that the gap is too large, because you haven't calibrated the Z-Offset yet.

The reason I suspect that is that when I first got started, I did the same thing, and had a lot of the same problems. (By reducing the filament diameter and multiplier before correctly setting that first layer gap - I think I actually exacerbated the problem.)

So put the filament diameter back to 1.75 mm, and the multiplier ratio to 0.9, then correct the Z-Offset. After you have done that, then measure the width of the walls on the cube to get the extrusion multiplier without changing the filament diameter first. You'll probably find that it's right on the money with a multiplier in the 0.89-0.91 range.

I think you're going to get better results. I've stopped measuring the filament diameter on the spools now, it's a crap shoot what the diameter is going to be depending on where you measure it. It swings up, and it swings down, and 1.75 mm seems to be a pretty good average, so i just use that and don't bother with measuring the filament diameter any more. (Filament is also a lot more uniform now, unless you're buying cheap knock-off stuff.)

[scampa123]

I'm seeing different thicknesses on different "walls" of the printed calibration cube....What would cause one side being thicker than another? (Except maybe a bad caliper!)...

I've set my Z and calibrated that...

It's the wall thickness that seems to vary as I measure each side...

Thanks!

[ednisley]

different thicknesses on different "walls" of the printed calibration cube

Preview the actual G-Code paths (perhaps using http://gcode.ws/ to make absolutely sure that each wall has exactly one thread.

The slicer will do its best to fill your model with plastic and, for example, may decide to use one thread on the X walls and two threads on the Y walls. That can happen if you use a "single wall" model created for (say) a 0.45 mm thread width, then let the slicer pick whatever thread width it wants; it may pack one fat thread into a wall, then use two skinny threads on an adjacent wall, for reasons having to do with mysterious algorithms and numeric accuracy.

Remember that you must re-slice the model whenever you change any slicing parameter; I've gone around the loop a few times wondering why nothing changed, before discovering that, mmm, I was changing only a few slicer configuration numbers, not the final G-Code results.

[scampa123]

different thicknesses on different "walls" of the printed calibration cube

Preview the actual G-Code paths (perhaps using http://gcode.ws/ to make absolutely sure that each wall has exactly one thread.

The slicer will do its best to fill your model with plastic and, for example, may decide to use one thread on the X walls and two threads on the Y walls. That can happen if you use a "single wall" model created for (say) a 0.45 mm thread width, then let the slicer pick whatever thread width it wants; it may pack one fat thread into a wall, then use two skinny threads on an adjacent wall, for reasons having to do with mysterious algorithms and numeric accuracy.

Remember that you must re-slice the model whenever you change any slicing parameter; I've gone around the loop a few times wondering why nothing changed, before discovering that, mmm, I was changing only a few slicer configuration numbers, not the final G-Code results.

I use S3D for my slicing...I'll try bringing the gcode into gcode.ws and see what it shows...thank you!

[scampa123]

different thicknesses on different "walls" of the printed calibration cube

Preview the actual G-Code paths (perhaps using http://gcode.ws/ to make absolutely sure that each wall has exactly one thread.

The slicer will do its best to fill your model with plastic and, for example, may decide to use one thread on the X walls and two threads on the Y walls. That can happen if you use a "single wall" model created for (say) a 0.45 mm thread width, then let the slicer pick whatever thread width it wants; it may pack one fat thread into a wall, then use two skinny threads on an adjacent wall, for reasons having to do with mysterious algorithms and numeric accuracy.

Remember that you must re-slice the model whenever you change any slicing parameter; I've gone around the loop a few times wondering why nothing changed, before discovering that, mmm, I was changing only a few slicer configuration numbers, not the final G-Code results.

I brought it into gcode.ws but I've never used this before. Can you tell me what I should be doing/looking for?

Thanks again!

[willnewton]

If you are using S3d then just look at the print preview window and scroll through the layers to see what it is doing.

[scampa123]

If you are using S3d then just look at the print preview window and scroll through the layers to see what it is doing.

Yes I did that and from what I can tell it seems like its a single thread going around the cube. I even played with gcode analyzer however it's view is a bit strange to me since it makes the extrusion look thick on the screen....