What would cause zits on curved walls only?

I have a Taz 5 and S3D. I am not sure if this issue is slicer related or hardware related.

I get perfectly clean straight walls on my prints, but curved walls have blemishes on them (from starting and stopping layers). I have cured these on straight walls using coasting and negative restart after retractions and they are perfection But on curved walls the blemishes remain.

Is this some sort of issue with the hardware trying to drive 3 steppers at one time? X Y and the extruder?

Do you have a picture?

I get some of exactly what you are describing with Slic3r as well. I don’t think it’s a software issue, and I don’t think it is a motion segment issue. I think it has to do with the extruder itself actually, and maybe the controller. On a straight run any pressure variances tend to not show up because outflow is constant. As soon as it hits a curve and that direction change though, it seems like any excess pressure from that layer, especially on a really well dialed in machine that shows no other blemishes, is going to come out at that location. Possibly a pressure wave, possibly a belt handoff issue, maybe a nozzle wobble, but that shouldn’t be an issue on my machine at all at this point. I’m not really sure yet. Slowing the speed does seem to diminish the effect. I’m not convinced it isn’t some sort of minor overextrusion. I really need to get a high speed camera now to figure out exactly what is going on there for sure though. It doesn’t seem to be taz specific either. You can see similar artifacts on really well dialed in ultimaker and mendelmax prints in similar locations. It definitly varies with layer height as well. It’s every other layer, or every two layers or so, but intermittantly too, which would make me think it’s extrusion related.

My next testing phases are as follows:

  1. Finish the sensor project (or at least the width sensor) and see if its due to variances in filliament width
  2. Finish the belted bondtech standard filliament extruder to see if this is a filliament feed issue
  3. If that doesn’t fix the issue, build a pen plotter insert with 3 pens in a triangle shape around the center of where the nozzle would be and an automatically raising writing surgface of some sort to see if the motion path through corners is accurate and precise on each pass
  4. Get a high speed camera to see what the hell is actually going on (the cheapest possible option of which seems to be about $170…)

I’ll post pics tonight. But trust me the straight vertical walls are perfect in every direction, and every curved wall is covered with blemishes.

I think (my guess) it is related to the printer pcb. I think that it has trouble keeping three hard working steppers in perfect sync and it effects extruder function negatively (in the most obvious manner, during retracts and restarts). This behavior happens to all Taz-like printers that use the same controller as far as I know.

It can’t be 1, because that would effect all walls the same wouldn’t it? The effect would be random in appearance.
2 seems unlikely because the feed shouldn’t really be effected by the other axes’ motion

I think it’s data choke of some kind.

It’s unlikely that filliament width or feed is the issue, but it might be. On long sections any variance might just be small enough that it gets buried in the part or the extrusion wall. It may only be showing up because of the effective compression of the corners. It’s the whole oval racetrack effect. The car wheels on the inside travel less than the wheels on the outside, because the inner track is shorter. Same thing with filliament extrusion in corners. but software seems to be handling it properly in most layers, but not all of them. or software is handling it just fine and it’s something else entirely.

I get the same thing using S3D. I attribute it to Marlin / slicers not using native arcs when slicing the models. I’d be interested to see the results from someone using native arcs (G02 / G03) insteard of G1 segments.

Overhang or curved exteriors? Overhangs hard to cure, but curved exteriors can be mitigated. I still get zits in random and usually in groups.

I had a challenging custom 1911 grip that helped me understand how different settings affect the final finish. Here’s a few that I feel that made a difference:

  • Slow the print down - usually print at 90mm/s (5400mm/min), I’ll drop to 70mm/s (4200mm/min)… with a target perimeter underspeed of 40mm/s (2400mm/min). Drop the infill speed a bit also…
  • Check your extrusion rate… I find 90-95% flow multiplier helps with print accuracy and top layer. It may also help with the exterior zits, but I haven’t compared in a while.
  • Try rotating the object on the bed. A 45deg rotation in the Z-axis can help.
  • In S3D, try the different options for the “Layer” -> “Start Points”… for a while I chose a start point to put it on a seam or corner if possible.


    I ended up printing the 1911 grip upright and side-by-side (flat sides facing each other). The knife edge of the grip posed a problem, which I added a raft with “tack” points to minimize the bottom edge from warping / curving. The client was pretty happy with the time and effort… I probably went through a spool of black ABS trying to perfect the print.

Hope some of the tips help.

Well I don’t think that’s even possible for a slicer to use true curves since they use STL files which are all straight line polys. You’d have to develop a slicer that used a true solid model and boy my brain hurts thinking about the complexity of trying to do that. It could be done obviously but by somebody a lot smarter than most.

I am talking about curved vertical walls. Nothing corrects this that I can find. I am fairly convinced it is hardware related.

Overhangs are pretty bad in general on the Taz. I cure overhang problems by making sure there’s a lot of time between layers, either because it’s a large object or by printing multiple objects.

I can print great overhangs with my current setup, and I don’t see any other blemishes on anything I print except for corner pieces. Whats wierd is that full cylendars often don’t seem to have as many of them, and small cylendars have none at all. And they are so inconsistant and random I can’t find a pattern yet. They aren’t very noticible at all, but compared to the rest of the print it’s still something to work on.

I get perfect straight walls but any curve at all, regardless of size and I get zits all over the place. Overhangs are bad on the Taz in relation to my other printer, but I can get them to work really well if I pay attention to layer cooling.

To ask again, for somebody not knowing this problem: Somebody has a picture of this problem? I have no problems with overhangs and my curves are as clean as straight walls, so it’s possible or I can’t imagine you are taking about :wink:

So does are referring to walls that curve outwards (ie. hump), or inwards (ie. pocket)?

For better overhangs, try more perimeter overlap in S3D. I get the best results with 30-35%, and a lower extrusion multiplier…

I had to print something to get a good picture of the issue I have. You can see it here.


Every single other section of that print is flawless. There is a little dip before and after the hole in the flat section, but that seems to be normal. But as soon as it gets to the hump or the fillet before and after it. They aren’t major, and probably the only reason i am noticing them is because they are the only remaining flaws. But, I don’t know whats causing them. I know that the perimiters were not starting or stopping in those locations in most cases. And as you can see, not all layers have them. it’s worse at the bottom where there is an internal hole behind that, but the top 3/4 of the structure at that curve is identical, yet not an identical pattern.

/shrug? I dunno what causing it, but I intend to find out.

Very intresting, never had seen something before… Which version of Marlin are you running? The first thing I would try is to dramaticaly slow down the speed (to around Lulzbot default ~40-50mm/s) and see if it’s still there.
I think to solve this, we should be very precise in separating the different cases. You say it’s no start/stop point of the perimeter in your picture and I belive this, but the thread starter is refering to layer start/stop point. The last one should be a thing of tuning settings, your one I would say is something inside slicer or (more likely) FW.

Speed doesn’t make a difference at all. I can run it super slow or super fast and get the same exact results. It’s not the part, it happens on most curved surfaces. Marlin is at whatever the latest build was 2 months ago. It was also doing that on older versions. SLic3r is used for model generation, but models made with Cura seem to do the same thing. Given my machine setup, I’m 95% positive it isn’t a hardware motion glitch. It could be extruder related.

The thing that is just baffling to me is i’d expect it to be doing whatever it is at every layer, or at least with a consistent pattern if it was overextrusion or a hardware issue.

I’m no expert, but looking at the photo something occurred to me. Are those bumps on the curve places where retractions occur? Maybe adjusting retraction length or speed or coasting could help.

There shouldn’t be any retractions occuring in those areas, and I don’t remember any during the print at those locations.