Off round circles...

I purchased the TAZ4 kit and it went together pretty smoothly aside from the usual issues when building things from scratch. I managed to get it together in just about 15 hours. I had some binding on the right z axis due to the lead screw coupling attachment being out of alignment(3d printed part), but I quick fixed it by removing the two lower screws so it wouldn’t “pull” on the right lead screw. So I ended up getting things up and running, did the obligatory few days of failed prints etc. and learned and have learned a ton about 3D printing due in part to building the kit from scratch. Overall, I’d say I’m fairly pleased with the product and it offers a great intermediate platform to learn and build from.

Overall, I am fairly pleased with my prints, but there’s one thing I’d like to get figured out and that is circlular objects come out with two creases and slightly off-round. I’m not sure whether to chalk this up to just normal pit falls of 3D printing on a reprap machine or should I expect perfection? I’ve attached pictures of what I mean. It’s almost like half the circumfrance prints slightly skewed, which logically I’d figure has something to do with the x,y axis. but just want to hear others opinions…
OffroundIssues.jpg
testprint3.jpg
testprint2.jpg
testprint1.jpg

I’ve not had any trouble printing perfectly round objects.

Starting off simply:

Are you using Slic3r & Pronterface?
Are you using the Lulzbot provided profile settings?

Which way did that print? Large funnel on bed? Or smaller base on bed?
Can’t tell from picture, but it looks like smaller irregular base does not have the offset issue the larger funnel has. Is that correct?

Oh snap, I remembered I wanted to add my gcode in my post but forgot…

Here were the settings…

T:200
BT:60
LH:.2mm
FLH:.3mm
speed: 50mm/s
BRIM: 6mm
skirt: no
raft: no
Slic3r and Pronterface.


Yes, now that you mention it, anything that’s not “round” throughout the build comes out fine, but as soon as there is a circular form it comes out like that. It’s weird… it’s very minor, I can live with it but if you’re saying you get perfect circles than something’s up. Smaller base on bed. It does it with other files also, the larger the circle, the less pronounced. Just trying to pinpoint if it’s hardware or software related, I’m leaning towards hardware. If you think about it, it’s at the point where the Y axis bed would shift directions while going through the motions of completing a circle.

and the weirdest part, is the rest of the model comes out fine. Some slight z banding but I can work on that.

In Pronterface - when you view the “build preview” where you can go slice by slice, do you see that offset in the larger diameters?

I had this problem quite bad with my prebuilt TAZ4, went through all sorts of trouble shooting with support. In the end I returned it for replacement. The replacement was better but still not perfect, really annoying when trying to print items with a diameter or holes. Keen to see if anyone can come up with an answer to this issue.

You can see in this thread: https://forum.lulzbot.com/t/taz4-printer-enclosure-heated/900/2

I’ve done some large diameter, thin wall parts.

If you click on the attached pictures you should see there’s no offset in the walls.

Makes me wonder if your parts file, the stl file, or maybe the gcode file was somehow corrupted.
I’d also check the belts on the stepper motors.

I’d print a simple solid cylinder, with infill of about 30%, to see if that cylinder has same offset.
Use the brim feature to see if the brim has offset. Something like a one inch diameter and 1 inch high cylinder.
Then maybe do the same part, but put thru hole in cylinder. 1 inch OD x 0.5 inch ID, and one inch high.

You could also make your skirt height something like 1 inch high. Does the skirt print with offset?

In my quest for perfect circles I tried Slic3r, Cura and Simplify3D including gcode and stl files supplied by support. None of these worked. Still think this is a hardware problem. Going to try belt tension and part you have suggested.

First of all, I appreciate all the replies and support. That’s one of the reasons why I bought this printer, because I knew there was a fairly large, and growing community of users of this printer. It sucks when there are problems, but it’s so cool to know that there is a huge database of information to help troubleshoot things! :slight_smile:


Edlink, I glance at my files using a web based gcode viewer, obviously it doesn’t show the offset, so actually that sort of rules out gcode related issues, which I’m just realizing now.

Also good suggestions for troubleshooting tests. I will definitely try them when I have time, just havn’t done much printing lately because I’ve been designing a custom sheet metal enclosure for the printer. Sort of seperate topic but I’m wondering if others would be interested in such a thing, instead of using a hobbed together cardboard box, having something a little more “polished”. If I can get a small group together for a bulk buy, it will bring the costs down somewhat. I’m trying to keep it to a few hundred dollars, all sheet metal, powder coated, integrated fan ports for airflow when cooking PLA, nice acrylic door with seal, bolt holes for optional fan heater and a bunch of other little bells and whistles, a nice polished little box to give your Lulzbot a home. It’s sort of naked as is.

Also, I’m glad to hear others have had this issue but I’m also somewhat concerned because it still seems like no one know’s what it is! lol…

My belts and pulleys are as tight as beans, I think one of my tests will be to swap out the Y-motor for a new one, there seems to be a little bit of chunkiness when the bed moves towards the Y end stop.

PMA, please post results with pictures when you have them…

When we evaluated PMA’s exchanged printer we found that the X axis belt was rather loose. After tightening the belt it printed well. We recommend Medium well to well done belt tension:

When tightening the belts, use this method to judge the tension:

http://www.leggebeef.com/finger-test-meat-doneness.png

Loosen one of the belt clamps on the axis, and pull the belt tighter, then tighten the belt clamp to re-secure the belt.

An X belt tensioner retrofit does exist if it’s of interest to anyone.
http://www.thingiverse.com/thing:438318

Jonathan

I have tried tensioning my belts, even using the devices you can print and add to the belt. Still the same.


What is up with the belt tensioning hand depiction stuff. I’ve got a hand gesture for you! :slight_smile: It involves only one finger… :wink:


Guys, I’m an engineer/designer, the first thing I’ve done when troubleshooting is go through the list of possible things, so belt tension was checked along time ago, if anything, I might try loosening the belt tension somewhat. I have them very snug, absolutely ZERO backlash, play or anything. But I will experiment and report back. Ofcourse I have to get the issue that hasn’t be solved and has a mystery surrounding it :stuck_out_tongue:

Plus, PMA has gone through the whole experience and still hasn’t solved his off round circle dilema, I have a feeling it’s going to be one of those mystery issues.

PMA, thanks for sharing, you’re getting basically exactly what I’m getting except a little bit worse than mine. I have two creases opposite side each other and the two hemisphere’s print accurately but it’s almost like one side of the circle becomes offset. Logically that would tell you that when the Y switches directions, there is a step lost or something or a shift that stays continuous.

One thing I’ve noticed that compounds the weirdness of it all is that I’m pretty sure my brims and skirts are PERFECT! WTF!!! hahahaha.

Let’s keep this thread going! gotta figure it out. Also, I’m willing to bet it has something to do with the 3D printed parts being off whack, one of the downsides to this printer but eh, if everything was CNC’d, you’d be looking at a whole different price point! :slight_smile:

What about the Lulzbot people? What do you have to say? any ide’ers? Any help is appreciated.

Jonathan - If you print a 2 inch by 2 inch square, a quarter of an inch high, and put a 1 inch diameter thru hole in it, does the hole have the offset?

Ed,

I’ll try that troubleshoot and post up the results…within a few days, just busy atm.
Thanks.

Try a couple of small simple parts like these. Nothing big. Simple shapes.

Print two of each, but rotate 1 of each set so the flat side of the square in each part isn’t running parallel to x or y axis.

I’d assume if there was a problem with a stepper motor the offset would appear adjacent to a flat side on one print, and then adjacent to a corner on a rotated print.

Here’s my test print for the 2" square with .6" Radius hole in center. Notice the small crease at the top, it’s barely noticeable but the whole circle is off round. I got a little skewing towards -x on this one but I think that’s because I havn’t leveled the Z axis screws to the bed in a while, usually I will readjust every few prints, just forgot this time.

Disregarding the skewing, which normally isn’t an issue, the square measures 2.021" left/right along X, and 2.01" along Y. The circle diameter is about 1.184", it should be 1.2", but still pretty close.

Printing out a barrage of test prints isn’t going to do much else but confirm the same problem. I’m not sure I understand? Would those test prints oriented that way isolate a certain variable, ie the motors vs. pulleys etc…


Before I got this printer, I had ZERO hands on experience with 3D printing hardware and software and how it all fit together. Now that I’ve had a few weeks to play with it, I’m now enlightened. I know I’ll have it printing perfectly at some point but I think that one downside of this design, which I believe is based on the i3, prusa models and similar styles of rep-raps, is that all the planes are isolated from each other, it makes it harder to dial things in correctly and they’re susceptible to going out of level relative to each other. I really think the best style to go with would be X on Y carriage, and Z just straight up and down. That way, X and Y are locked to each other so it’s impossible for them to get out of level relative to each other, so you just have to worry about leveling between (XY) and Z planes. And Z should be on a single large lead screw so it’s consistently level. Sort of common sense stuff. I plan on building one like this…eventually…:slight_smile:


Sorry - I’m not sure I see the problem in your test print. Nothing at all like the pictures you posted in first post of this thread.

These printers are for prototyping. If you need accuracy to 3 decimal places you get parts machined.

There’s shrinkage with ABS. That kind of thing can be designed into parts if size is critical.

I also wonder if some of your “out of round” might be due to extruder temp, bed temp, and extrusion parameters?