Retraction not retracting?

TAZ 4. Bought in November '14. In an effort to solve my glop problem, I am varying my retraction length. I get results that I did not expect.

Printing ABS 230/85, sliced in Slic3r v1.2.4, 0.20 layer heights, keeping all parameters the same except retraction length, here’s what I get (I deliberately stopped two of the following prints before completion).

Retraction length: 0 mm (ie, no retraction).


Retraction length: 1.75 mm

Retraction length: 3 mm

I am certain about which picture is which. Am I correct that it is unusual that the print that used no retraction looks the best out of the three?

Here are the gcode files.
flower pot0mm retraction.gcode (4.6 MB)
flower pot 175mm retraction.gcode (4.68 MB)
flower pot 3mm retraction.gcode (4.68 MB)

I’m gonna be little to no help at all other than ask a few more questions. :blush:

Are you using a Lulzbot provided ABS config? Which one?

It also appears you’re using an “unstable” release of Slic3r. I don’t know if that’s an issue or not.

Are you using Pronterface after you slice? If so, is there a setting in Pronterface that’s conflicting with the changes you’re making in Slic3r?

I’ve never seen the “glop” issues you’re seeing. I use the ABS profiles from Lulzbot religiously. I’m also using the stable release of Slic3r and Proterface. You’re sure your filament diameter is correct? What nozzle size are you using? .35 or .5 or something else?

I didn’t even know where the retraction settings were in Slic3r until this post prompted me to go looking for them.
So whatever Lulzbot has in their configs has been working really well for me. :blush:

hey man
the pot you printed again Looks exactly the same as the ones coming out for me, with the same Glop issues.

Just wanted to throw in: When watching the Printer doing it’s thing, I also thought recently that it used to do a lot more retracting, constantly making the noise of the Motor pulling Filament back.

I feel it’s not doing that anymore, as if retraction is not Setup properly, and thus we get a glop each time a new layer is made.

I would agree the best results you have are without retraction, which is very strange.

I will do more testing when i get some time, past few days have been hell, no time for doing any prints :frowning:

Edit: do you see your Printer doing the retractions at all?

See also: https://forum.lulzbot.com/t/show-off-your-printer-and-printing-station/48/1

edlink,

>>I’m gonna be little to no help at all other than ask a few more questions.
I appreciate your eyeballs on my problem–thanks!

>>Are you using a Lulzbot provided ABS config? Which one?
I start with the standard Lulzbot “ABS fine” profile, change layer height to 0.20, print “external perimeters first,” and vary retraction distance as noted in the pictures. I take out the custom starting gcode that sets the temps (I set temps manually 230/85 c). I also add custom gcode to set my extruder calibrated steps / mm.[/quote]

>>It also appears you’re using an “unstable” release of Slic3r. I don’t know if that’s an issue or not.
I have since resliced with Slic3r 1.1.7 (stable) and reprinted. I get the exact same problem.

>>Are you using Pronterface after you slice? If so, is there a setting in Pronterface that’s conflicting with the changes you’re making in Slic3r?
I use Octoprint. I don’t think there could be a conflict. Please correct me if I am wrong here.

>>I’ve never seen the “glop” issues you’re seeing. I use the ABS profiles from Lulzbot religiously. I’m also using the stable release of Slic3r and Proterface. You’re sure your filament diameter is correct? What nozzle size are you using? .35 or .5 or something else?
I am sure filament diameter is correct. I use the stock extruder (drive, hotend, nozzle) that shipped. The nozzle diameter is 0.35 mm. You probably always print the exterior perimeter last. That will hide the problem. The problem is with retraction. Specifically the re-insertion after a retraction creates a glop if the previous move is large.

>>I didn’t even know where the retraction settings were in Slic3r until this post prompted me to go looking for them. So whatever Lulzbot has in their configs has been working really well for me.
Yup. Like I said, the stock profiles put the glop next to the infill. Try printing with a single perimeter. I bet you’ll see them then.

zhd,

If you want to get rid of the glops, try this: print at least two perimeters and print external perimeters last. I know you can do that in Slic3r, don’t know about S3D. I don’t know if our two issues are exactly the same, but I’m pretty sure that this will hide your glops next to the infill.

I’m pretty sure I see my extruder doing the retractions when it’s supposed to.

So pardon me for the stupid question, but isn’t that the “fix” to your problem? :blush:

Make sure the “external perimeters first” box is NOT checked off, and don’t change the number of perimeters from the default, which my “fine_ABS_no-support” shows as set to 4.

edit to add: I can understand if you’re just trying to figure out if you have a problem or not. I can surely see when I’m printing if the herringbone gears are reversing for retractions. I could also see where retracting too far could cause a glob of something when the extruder is forced to really advance material faster due to having retracted material farther. Sort of the more tension on the rubberband when it releases effect.

edlink:

>>So pardon me for the stupid question, but isn’t that the “fix” to your problem?

It depends, right? If you’re willing to use Slic3r forever, always have more than one perimeter, don’t tick that checkbox, and the problem never manifests in some other yet undiscovered way, then yes I think you could call it a successful work-around. But there are a number of considerations that I think motivate a more complete and considered reaction.

  1. Unintended consequences. What if this glop-issue manifests in other ways, more subtle, more insidious? If we just throw our hands up and say “solved” now, without fully understanding it, then we could be leaving a figurative rake on the ground to smack us in the face later. If you don’t fix a problem at the source, once, you can make a lot more work for yourself as you scramble to get around the problem when it presents in other contexts.
  2. Too narrow scope. This ‘fix’ doesn’t apply to Cura users. Many people want to use Cura. Lulzbot wants people to use Cura. You’ll see these glops in Cura too (it looks a little different, more about that later), and Cura doesn’t have the option of controlling which perimeters print first or last.
  3. Helping others. The more general a solution, the more users are helped by the fix. At the very least, if you identify the cause and articulate it in the most general way, you will help the largest number of people.

So it’s not just a philosophical question of whether a problem avoided is a problem solved. There are real, broader, considerations.

I said I would update this thread with info about how this issue manifests with Cura. So here it is.

First off, the following models are one-half the linear size of the previous ones. The layer height is the same. But the 50% scale makes the blobbiness look worse than it is. I had to do it to cut the print time. But it’s going to be the relative quality that’s important.

Looking at the G-code generated by Slic3r and Cura, it looks to me like Cura makes a more continuous extrusion path, with fewer retracts and less non-extruding distance travelled. That’s good, and it shows. Here is the flower pot. Remember, in comparison to the previous pictures, the general surface quality looks worse than it is.
hc8a1733.jpg
So you see some blobs here and there (not the holes), but not many. I think that’s because the extrusion path is so continuous that there are very few retracts. And when there are retracts/advances, half the time the advances are on an inner perimeter, and (because Cura really optimizes the path) the move before the advance is short.

But look what happens when you put four copies on the print bed.
hc8a1732.jpg
I changed nothing in the Cura settings (except to print four). Now you see lots of blobs. That’s because each layer has a retract, followed by a long move to the next island, followed by an advance. Half the time Cura puts the advance on an inner perimeter. But that still makes a lot of blobs. Unfortunately, Cura doesn’t have an option for setting which perimeters you print first, inner or outer.

But Slic3r does. So compare the last picture (Cura) to Slic3r’s G-code for 4-up on the bed:
hc8a1736.jpg
I tried to match the parameters (layer height, retraction distance, speeds, it’s basically the ABS fine config). But Slic3r gives you the option of printing exterior perimeters last. So that puts the blobs next to the infill where you can’t see it. The result is much cleaner, and although the pictures exaggerate the effect, you really can tell the difference with your naked eye.

So, I’m still hoping that somebody will chime in with “you dope, you’ve got your ________ set too _________,” and that will make all the blobs go away. But really, any ideas about this would be greatly appreciated! Thanks!

I will try to print the same flower pot again using my improved s3d settings. (since on the more rectangular prints i’ve been doing since fixing some issues with your help I didn’t see any blobs at all)

Maybe we can then compare the s3d settings with your cura / slic3r settings.

For the record, I did always print 2 perimeters but never printed the ouside perimeters last so far. As you stated, if I get the same issues on the same flower pots in S3D I could do that as a quick fix but that’s not really a solution to the root of the problem.

Will check back hopefully tomorrow when I have printed another pot :wink: