1mm vertical surfaces too thick. (Page 1 of 3)

I did try calibrating that once already. I believe the default from repetier was 99, which was actually very close, but actually under-extruding very slightly in my measurements! I tried adjusting based on the measurements I took, and set the esteps in the eeprom to 101. Not surprising, it didn't make much difference except that it appeared to be over-extruding a bit.  I also tried backing it off to 96, which seems to improve the quality a bit, but it doesn't fix my thick vertical layers problem. And redoing the estep calibration, it's even more obvious that it's off. When I told it to extrude 100mm, I came up with 94.2mm extruded.  Any other things you can think of that would affect this?

To elaborate a bit more, you should set the extruder steps by measuring the actual filament travel, then set the extruder multiplier with a calibration print. The best method I've found is https://github.com/alexrj/Slic3r/wiki/Calibration . This multiplier really needs to be set per-spool, as it varies across batches of filament.

Yes.

The extrusion multiplier in the slicer settings is essentially a way to compensate for the fact that filament diameter isn't always precisely correct and density can change before and after extrusion such that a cubic MM going in isn't precisely a cubic MM coming out.

It's good to do both, though. Set the EEPROM extruder steps per MM with the ruler method and never change it, then do a quick 40x10 print to set the extrusion multiplier for each spool you load. If you set the steps per MM correctly, your extrusion multiplier should always be between ~0.9 to ~1.1. If it's far off that, it's worth checking if your extruder is slipping or somesuch.

Results of this came out a little closer to 1mm for the entire structure, but still oversized at about 1.25mm actual. However, it was tricky to measure though, because the lines making up the vertical surfaces were thin and detached from one another, which is what I would expect since I made the extrusion multiplier .695. It's almost like it's still trying to make the width of the 1mm wall closer to 1.4 or 1.5, and just coming up thinner because of the lower extrusion rate. I tried bumping this up to 1.77, and now they are coming up 1.3-1.4mm thick, and still very detached, especially when I print larger things.

That's... debated. You can measure the filament itself and then calibrate the extrusion multiplier afterwards, or you can just assume the ideal 1.75mm and set the extrusion multiplier with that. The end effect is exactly the same because the slicer uses both values to calculate extrusion rates. As long as the walls of your 40x10 or whatever calibration print come out at the size the slicer thought they should, you're set.

pwlars, if the calibration is right and your 40x10 comes out correct, the next question is whether the slicer is actually generating the walls at the thickness you think they should be. That's a bit harder to answer, unfortunately. In general, the slicer is going to have a width it wants for external perimeters, and it's possible that you're seeing is actually the slicer making a double wall of a fixed size in place of the 1mm walls you asked for.  Unfortunately, how you address that depends on the slicer.

It is???

OP, ignoring the fact that "it's been debated".  Proper calibration is essential and there is science and fact to support it.

It's simple and Carl is correct about going back to the basics.  Ignoring the science and using a fudge factor (extrusion multiplier) to compensate is only going to give you a headache.  Anyone that wants to disagree simply doesn't understand all of the variables involved.

Step
1. Calibrate E:steps per mm (100mm called for 100mm pulled in)
2. Measure your filament along a good length of filament and enter the average as the filament diameter.
3. Print a single wall calibration cube and measure the wall thickness.
4. Use that information to adjust the extrusion multiplier.
5. Repeat steps 3&4 until you get a single wall thickness equal to the proper 120% of your nozzle diameter.  In your case (assuming a .4 nozzle) a thickness average of the four walls equaling .48
The 120% is a product of die swell as the plastic exits the nozzle.

There are a couple other various versions of the above process but they use the same basic principles.

So although this subject has been debated (a very long time ago), the idea of using extrusion multiplier to compensate for poor calibration lost the debate.

Let's strive to make each and every one of us more knowledgeable through factual statements.

In my years of printing since 2008 I have never modified my multiplier. It has always been on 1 or 100 depending on the software. If you do hardware cal correctly then your slicer should not need to compensate. He is not flaming you, but I assure you he and I as well as many others here know what we are doing and following proven practices is better than trying something new that may cause more issues. Keep it sweet and simple and always go back to basics when errors arise.

I also think you might be confusing filament width for filament diameter which are two different settings. The only thing diameter affects is the amount of filament being fed at any given speed. If this is set wrong however it throws off other variables that depend on this value for their calculations such as width.

There are many slicers out there and not all work the same. If they did you would not have people in droves spending $150 bucks for Simplify3D. It has an extrusion width setting and as Ward said for a .4 nozzle that setting by default id .48. Yes you can change it but I have yet had to. My machine is calibrated to almost perfect values.. When I ask for 100mm I get exactly that. When I specify 1mm walls I get exactly .99 to 1.001 walls. This is all done by hardware cal only. I have not made any changes in the slicer or firmware which is Marlin.

So first you explain to me how you are justified in patronizing me because I joined this forum recently and haven't considered your experience...

I never mentioned extrusion width and I never meant extrusion width.

Take a look at the slic3r source at https://github.com/alexrj/Slic3r/blob/m … truder.cpp at line 15. It's calculating the extrusion distance per cubic MM, and the formula is extrusion_multiplier * (4 / (filament_diameter ^ 2 * pi)).

My point here is that you can measure and set the filament diameter and then set the extrusion multiplier or you can just assume the ideal filament diameter and set the extrusion multiplier. The end result will be exactly the same because the actual number that matters is the extrusion distance per cubic MM. These two numbers we keep debating are really just two knobs controlling the same final result.

This is factually incorrect. Slicers vary extrusion widths by increasing or decreasing the extruded length relative to the traveled length. There are of course physical limits to how wide or narrow the extrusion width can be, but try setting the "Perimeters" or "External perimeters" parameters within slic3r, print a single-walled object, and observe that the parameters do change the width of a single wall.

If you don't want to change any parameters at all, simply slice an object in slic3r itself instead of Repetier Host or another software that embeds it. Slic3r colors the gap fill lines in your model red, and you can pretty easily observe that it's automatically varying the extrusion width for you so the gap fill comes out at the width it needs. Note that you do have to have a speed set for "Gap fill" as this feature is currently incompatible with autospeed.

In the source, this corresponds to https://github.com/alexrj/Slic3r/blob/c … Region.cpp starting at line 22, where the width of a "Flow" (basically a width and height of an extruded path with some metadata) is created, and later at https://github.com/alexrj/Slic3r/blob/m … r/Flow.cpp around line 69, where a Flow is converted into a ratio of cubic millimeters of filament to extrude per millimeter of distance the extruder travels. Flow.cpp also contains a variety of formulas to calculate flow widths.

pwlars, sorry to hijack your thread in this way, but I suspect these parameters "External perimeters" and "Perimeters" are what you need to modify. However, you definitely have some other issue if your multiplier came out to ~0.7 - there's no way your filament is 1.75/0.7=2.5mm thick (+/- 10%), so there has to be some other reason you're getting such massive overextrusion. It's apparent, though, that my presence here is hurting more than it's helping, so it's probably best that I make this my last post.

Rather than responding to everything indivudually, let me try to do it all at once.

Actually, 1.75 looks like a very reasonable average. I get a pretty narrow range of results over a long segment of filament ranging from about 1.73-1.77, with most coming right out at 1.74-1.75. nozzle is at .4

I did try printing the _40x10 calibration part, but I don't see any place that says what the expected width of the walls themselves should be? I *looks* like they should be .5mm thick, but I don't see it mentioned anywhere. I did this print with the extrusion multiplier reset back to 1, and I'm measuring mostly between .75mm - .8mm for the walls of the part. The part itself seems to come out to 20x30, which seems a little weird given the name of the part is 40x10, but I loaded the part into a cad program though and it looks like it should indeed measure about 20x30.

Otherwise, it prints looking very nice! The walls are all very straight and even, the first layer is not squished out, no gaps, strings, or anything.  Just that the walls are too thick.

What's the best way of getting this, just make something that is .4 wide? Or was that 40x10 calibration part supposed to have a .4mm wall?

The easiest way to do a calibration cube of any dimensions is draw a simple solid cube.  For example a 20mm cube then set up your slice profile (save it as Cailbration or similar) as Carl suggested.  Single bottom and Single perimeter with no top layer and zero infill.  I like to use a 20mm cube but any dimensions will work.  The important part is to not define a wall thickness in the design as it is used only for calibration.  What your slicer will do then is create a hollow cube with a single perimeter (independent of the model) that you can then print and measure the single perimeter wall to set the extrusion multiplier accordingly.  Once you've achieved the proper values you will then use that information to set up the profile that you actually want to print with.  So if you then want to print items with more perimeters or more infill for example, just reuse that extrusion multiplier for that different profile.
You may want to take a look and see what the extrusion widths are set at.  Using 0 as a default where possible is just fine so the screenshot is just for reference. But you will want to at least set the top default extrusion width to .48

No you need to adjust extrusion multiplier until the wall is .48.
.48 is a constant that is determined by the 120% die swell of your .40 nozzle.

You might double check that E:steps is still accurate because something isn't adding up here.