3D Printers

Hey all.

Our team just got an Airwolf 3D printer thanks to one of our mentors (he’s letting us use it this season).

I wanted to post to get a little bit more understanding from the CD community about the intricacies of 3D printing, and maybe an address to some of my problems.

Firstly, could you tell us a little more about the various settings in Slic3r (if you use it) and Pronterface? They’re new software to us, and while we have printed a few objects out and familiarized ourselves with them, we’re not experts.
(Particularly Slic3r, and the various settings from layer height, to temperatures, perimeters, etc).

I’d also really like to know about the kind of speeds we should set, which speeds (there seems to be a lot of “speed” settings in Slic3r), and what kind of speeds should we set relative to layer height (higher permissible speeds at higher layer heights, lower at lower layer heights, etc).

As for my problems… we’ve tried to print out a few parts, and essentially the very bottom layers of the print are… well… not even remotely wholesome. They look like a pile of spaghetti rather than a solid layer of the part. This happens for the first few layers, and it appears that this is indeed a layer print of the part, not some kind of “pre-layer” to make sure future layers stick.

We want to have the entire part come out solid and perfect as it should, rather than a stringy bottom. I’ve posted pics below of the top and bottom of our part, maybe you’ll get a gist of what I’m talking about.

Top shot of smooth(ish) layers:

Bottom shot of spaghetti layers:

Side shot of increasingly better layers as we go up the z-axis:

I’ve run a Mendel, ReplicatorG with Skeinforge, PLA as material… but software doesn’t look like it’s all of the issue here.

If I got a part like that, the first thing I’d do is check Skeinforge’s “this is what I should be doing” window (Skeinlayer and Skeiniso–if Slic3r has any sort of analyze section, open that up and see if you can find a layer viewer) and verify that that bottom is in fact showing up right. If I couldn’t do that, I’d assume that it was correct.

Looking at the picture, I’m seeing the following possible issues, any one or all of which could be the issue:
–Not enough width/thickness ratio (on ours, we run very high at 2.6 and don’t really have issues with that, but lower and we do). To be point-blank, I’d put that as culprit #2 based on the top view–another couple decimal points up won’t hurt you. Look under whatever settings can be found for Fill. (You’re showing gaps on top, some not at the edges. The stuff at the edges can be tweaked with another ratio–I think it’s overlap with edge.)
–Too much speed for the stickiness. I don’t know what surface you’re printing on, but whatever it is, if the material can’t stick, you may want to run elsewhere on the plate, slow down, or replace the plate covering.
–But the #1 thing I’d say on seeing that: Your head is flying too high above the platform. Have someone grab a feeler gauge and check that you’re not much higher than the tip thickness above the platform when the head is all the way down. You’ll see a marked improvement if you bring the head down to the platform slightly.


Thanks for the quick reply.

We use Slic3r Configuration Files provided by Airwolf 3D for use with the printer. I’d like to post them to see if you notice anything that should be tweaked, but i’m not at our workspace right now. I’ll try to get them later.

As for the surface, it’s a piece of glass sitting atop the hotbed coated in acetone.

Speed may be an issue; it seemed to certainly be moving quite fast. I may not be recalling correctly, but 30 mm/s seems to stand out in my head.

What height above the platform should our nozzle be? What do you mean by tip thickness? I think our nozzle was something like half a centimeter or less above the platform during prints.

**EDIT: **Again, this is my first time using a 3D printer so I know very little about them, any explanation will help :P. The website for our printer is http://airwolf3d.com/, if maybe it offers some information that would be helpful for your diagnosis.

Here’s a suggestion. When milling parts on a cnc we use some material to hold it in a vise and mill it off after one side is complete. Try a similiar concept by making a starting base to get the material flowing through the nozzles then start making the part a little higher on the part then mill or sand off the base part. Use micro joints to hold the part then break off the micro joint once the part is made.


That’s actually what we were assuming the noodle-like initial layers were. We thought it was automatically generated g-code made to prepare for the actual printing of the part. But it seems like that isn’t the case.

I’m not familiar with Slic3r, so I’m not sure I could understand them. I noted that I use Skeinforge, which would probably have similar things, but in different places.

Speed may be an issue; it seemed to certainly be moving quite fast. I may not be recalling correctly, but 30 mm/s seems to stand out in my head.
That’s… pretty fast. reaches for Replicator/Skeinforge calls up the “fast” profile On the other hand, we’ve been traveling at 38 and feeding at 34 (both mm/s).

What height above the platform should our nozzle be? What do you mean by tip thickness? I think our nozzle was something like half a centimeter or less above the platform during prints.
5mm? That’s your problem right there. I looked at the Airwolf site, and the stock tip is 0.5mm (tip thickness). We run a 0.35mm tip (stock for our printer type), which they offer as an option. You want to run with the tip height above the printer at or below the tip thickness.

The other thing I’d look into is the tape. Kapton tape over aluminum over heating pad is what we’ve got on our Mendels; Airwolf suggests PET or Kapton tape. Kapton wears out its “stickiness” after a few hundred runs; I’m not sure about PET. If it’s not as sticky as it should be, move the center of the run a few millimeters to one side and see what happens.

Roy, the “raft” option might work–but in my experience, it’s not necessary for something that size. (The parts I sent you were all made without one, BTW.) I’m looking at the print head height as a primary culprit here because that part looks like the head is flying way high, and possibly not feeding fast enough in relation to movement speed.

If your print head is 5 mm above the bed that is really too high depending on the machine setup.

The “zero” spot for the z axis should be about a paper width thickness from the bed. I am not familiar with that machine, but all the ones I work with start with 0 right by the bed.

Check in slic3r is pretty basic in these setting, so there should not be much initial “preprocessing” movements.

In pronterface go to your slicing settings under the settings menu.

on the start / end gcode what are you seeing in the start section?

Is it just a G28 (home all axis) command?



How thick is the paper? We found a practice-field schedule card from CMP worked well, but a feeler gauge was more accurate.

I took a quick look at the printer, it looks like it has a screw adjustment for the z-limit switch for hte zero process. That may need some adjusting.

The first layer should be smooshed a little bit against the glass.

As for speeds, in the printer and filament section

Print speeds should be the normals speeds you can run at. I found on my Prusa I can do 40 mm/s on perimeters without and issue and a 50 mm/s infill.

But, in the bottom of that box you will see a setting for bottom layer speed ratio. Set this to .5 or higher as it will slow the initial layer and not go to fast building that initial layer.

The other speed to watch is the travel speed. If the travel speed is too high and the current is not high enough to the stepper motors they will skip steps.

60 mm/s is what I have had good luck with.

A feeler gauge would work. .2 mm is a typical layer height so somewhere between .1 and .2 mm would be a good starting point.

You will know when you get the first layer right is when is is smooth and consistent and does not “plow” through other plastic already laid down.

Sometimes it takes a little tuning to get just right.

Make sure you turn the skirt feature on as it gives a rough idea of the evenness of the bed. That is usually where I run into issues, but that printer may be fixed bed height.

Don’t be afraid to stop the print if the first couple of layers don’t look right.

Best advice I can give. We run a 3-layer shell, and if the first three layers come out, we’re good until much later in the build unless something goes badly wrong like a jammed tip. If the first three don’t come out, we’ll stop and tune.

Protip: find a small, 10-15 minute build. Maybe something you can do as a giveaway. If you make a machine settings change, or a machine change, run one and see if it turns out right. We use turtles.

Wow, the nozzle has to be that close?

You know… this is starting to make sense.

Perhaps the reason why the top layers turn out better is because at that point, the height of the object we’re printing is > 5 mm, and thus the nozzle can get super close to it. Whereas the first few layers are super far from the nozzle and thus turn out noodl-ey.


What’s happening is: You’re starting so high up that the material won’t stick. But, what does stick forms a base. As you get closer to the head, the material sticks better, and forms the shape better. So by the end of the part, it looks really good. But at the beginning, ugh.

I’ve seen that when I’ve pulled material from one part in a 4-part build–eventually, I got high enough that it wouldn’t act up, then it built almost normally, then other stuff happened (that I won’t go into here).

Use that height-adjustment screw set–it’s your friend–but once you get good parts, you don’t want to touch it unless you absolutely have to.

As you get into this a little further take a look at skeinforge. it is a bit more complex, but if has some more features over slic3r. One of the things it can do is print a raft which is a couple of layers of plastic that is used to offset the part. When I had some initial layer issues the raft really helped make some great parts. But after really tuning it, I don’t need it very often.

Good luck, looking forward to seeing your parts you produce.