Circles not coming out Circles

I recently printed the Simple Spool Hub from Thingiverse.com and noticed that my circles aren't coming out perfectly circular; they come out egg-shaped to some degree. It's probably hard to tell in the attached picture.
Does anyone have any suggestions as to how to fix the non-circle circles?
I'm printing with PLA and am using skeinforge.

Thanks,
Chad C

It is probably caused by backlash or eccentric pulleys.

---

[www.hydraraptor.blogspot.com]

I agree with nophead, check your pulleys

---

---

reprapworld.com
PLA, ABS, PETG multiple colors (3 and 1.75 mm) €16,49 / kg
Megatronics 3.0 €79.99 / Minitronics €37,19

I had the same problem. Was caused by the friction between the X-Axis belt the various waschers that help keep the belt level. Solution was to lubricate the the part between the belt and washers. Been working perfectly ever since

Giving the stepper more torque might help as well...

---

---

reprapworld.com
PLA, ABS, PETG multiple colors (3 and 1.75 mm) €16,49 / kg
Megatronics 3.0 €79.99 / Minitronics €37,19

might also be your x and y aren't perfectly perpendicular.

Now that the X and Y axis have been mentioned, I haven't really calibrated the X and Y mm/step in the firmware. I have steppers that are 0.9° per step instead of the standard 1.8° and just doubled the mm/step numbers originally. I've noticed that my 20mm calibration cubes come out to be about 20.5mm x 18.9mm.

I did go ahead and tighten down my X and Y axis belts and increased the torque yesterday and didnt really get much improvement. the steppers are a bit warm to the touch now. I dont really see any friction spots on the belt or along it either.

Anyone have suggestions on how to calibrate the X and Y mm/step or if this is even needed? I plan on just printing a 5" x 5" square and measuring with my calipers then adjusting the firmware.

Thanks,
Chad C

Unless your X and Y belts have different pitches it looks like you have backlash on your smaller side. I have never needed to calibrate my axes. I just multiply the belt pitch by the number of teeth.

---

[www.hydraraptor.blogspot.com]

Measuring with calipers will be inaccurate - small differences in amount of plastic deposited etc can (will) result in being a bit off. You'd have to measure from center of extrusion to center of extrusion with hand-written g-code (so offsets don't mess it up) and without other interfering plastic deposits. And even then backlash may cause problems.

I think it's generally agreed that calculating step/mm from step angle and mechanics is more reliable.

---

--
I'm building it with Baling Wire

I have the same problem, circles not being circles and straight lines not being straight

Here is the gcode seen on Skeinforge:


And just the first layer plotted with a pencil:


My X and Y gears were printed with highest quality of PP3DP UP Printer, with 0.2mm layer. Before I had other gears which were far from perfect/round and I still had this kind of problems.
My belts are tight to avoid backslash.

NOTE: I get the same results while printing. My pencil is tight against my extruder.
My calibration/config.txt file is the same as here: [reprap.org]

Edited 2 time(s). Last edit at 01/29/2011 09:03AM by casainho.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

You have backlash on your X axis. You should be able to see it just by turning the motor by hand and see the carriage pause when you change direction. Or hold the pulley in one hand and the carriage in the other and feel they are not firmly connected.

Causes:
Belt too loose.
Pulley loose on the shaft.
Wires in the the belt snapped.
Too much friction.
Filament dragging on the head.
Too little motor current.

Edited 1 time(s). Last edit at 01/29/2011 09:45AM by nophead.

---

[www.hydraraptor.blogspot.com]

Looking at the pencil-plot, my experience with machining is screaming "backlash!" The belts are tight, but something else could be loose. A sloppy bearing on one of the axes, perhaps? Does the X axis carriage have any wiggle room in the Y direction, and vice versa? Can the Z axis carriage be wiggled? Can the carriage be wiggled in any direction with the stepper holding position?

From old motorcycle troubleshooting, I learned to go back and check everything that I know is right. Because something isn't.

Edit: It's edifying to see that, while I"m a few minutes slower than nophead, my thoughts were running along the same lines.

Edited 1 time(s). Last edit at 01/29/2011 09:56AM by Dale Dunn.

nophead Wrote:
-------------------------------------------------------
> You have backlash on your X axis.

Ok, good! I see now that circles at Y movement are all perfect but on X part movement they are all bad!

Ok, monday I will see this and I will post the real problem/solution. Then I may also write this on Mendel Manual: [www.reprap.org]


You should be
> able to see it just by turning the motor by hand
> and see the carriage pause when you change
> direction. Or hold the pulley in one hand and the
> carriage in the other and feel they are not firmly
> connected.
>
> Causes:
> Belt too loose.
> Pulley loose on the shaft.
> Wires in the the belt snapped.
> Too much friction.
> Filament dragging on the head.

Very strange is that over the time I were changing the pulleys and tight the belt... over the time, the belt is the same!! I am betting now on a wrong type of belt that I bought or in to much friction...


> Too little motor current.

I guess not, because that would mean missing steps by motor, right?? -- if was missing steps on X and Y, then on layers after (Z increase) I would not get almost the same X and Y "plotting". Yeah, I didn't said this before.

Thanks everyone

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

> > Too little motor current.
>
> I guess not, because that would mean missing steps
> by motor, right?? -- if was missing steps on X and
> Y, then on layers after (Z increase) I would not
> get almost the same X and Y "plotting". Yeah, I
> didn't said this before.
>

No not missing steps as you don't have any cumulative error. Unless you have zero static friction, or a motor with infinite torque, the motor will stop in a slightly different position when going in each direction.

---

[www.hydraraptor.blogspot.com]

nophead Wrote:
-------------------------------------------------------
> > > Too little motor current.
> >
> > I guess not, because that would mean missing
> steps
> > by motor, right?? -- if was missing steps on X
> and
> > Y, then on layers after (Z increase) I would
> not
> > get almost the same X and Y "plotting". Yeah, I
> > didn't said this before.
> >
>
> No not missing steps as you don't have any
> cumulative error.

Right.


Unless you have zero static
> friction, or a motor with infinite torque, the
> motor will stop in a slightly different position
> when going in each direction.

Well, but that "slightly different position" would be something less or equal to one step size, right? -- my step size is very small, I would not be able to see one step ---> STEPS_PER_MM_X = 80

Ok, thanks.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

casainho Wrote:
-------------------------------------------------------
> nophead Wrote:
> --------------------------------------------------
> -----
> > > > Too little motor current.
> > >
> > > I guess not, because that would mean missing
> > steps
> > > by motor, right?? -- if was missing steps on
> X
> > and
> > > Y, then on layers after (Z increase) I would
> > not
> > > get almost the same X and Y "plotting". Yeah,
> I
> > > didn't said this before.
> > >
> >
> > No not missing steps as you don't have any
> > cumulative error.
>
> Right.
>
>
> Unless you have zero static
> > friction, or a motor with infinite torque, the
> > motor will stop in a slightly different
> position
> > when going in each direction.
>
> Well, but that "slightly different position" would
> be something less or equal to one step size,
> right? -- my step size is very small, I would not
> be able to see one step ---> STEPS_PER_MM_X = 80
>
> Ok, thanks.

No it could be any value up to two full steps given enough friction, any more and it would skip. That would be 0.4mm. I.e. microstepping is only accurate when there is no torque on the motor.

---

[www.hydraraptor.blogspot.com]

I have a problem with small circles, like 3mm for M3.
the print head moves around small circles too fast and changes direction too often, pulling the filament off the surface before it can bond and harden. this severely distorts small circles. I'm sure i could prevent this from happening if i reduced my overall speed, but i think the speed (flow/feedrate) should change proportionally to the circle diameter. is there a setting i'm missing in skeinforge?

I can't remember if you use skeinforge or not, Nophead, but maybe someone else knows.

Buback wrote:
>I have a problem with small circles, like 3mm for M3.
>the print head moves around small circles too fast and changes direction too often, pulling the filament off the surface before >it can bond and harden. this severely distorts small circles.
---------------------------

There's a makerbot.org tutorial regarding small circles out there somewhere Buback. Apologies I can't find it for you. Essentially, creating a 'circle' (many sided polygon) in Sketchup (the program you use?) or something similar, and shrinking it down to 3mm, creates a relatively detailed shape with many short sides. The resulting Skein forge tool path involves lots of changes in a short interval. It overwhelms your printer's controller chip. If you make your 3mm circles out of fewer sided polys, like a hexagon, you will hopefully see something acceptable coming out of the printer. The hexagon would naturally need reaming.

The rule of thumb in the tutorial was to keep the edges of the polygon around 1mm minimum.

Woodsmoke
r:-3

I use SF, but not to control my flow/feed rate. I don't slow down on small circles, but it is probably a good idea. I don't have problems with them sticking but they do shrink a lot.

If I print a hexagon with its sides on the circle, i.e. vertexes = 3/cos(180/6), then I get pretty close to a circle the right size. Polygons get rounded corners but their edges are straight lines, so they are in the right place. In general if I set the vertexes to twice the diameter in mm I get a perfect fit, any more and it starts to shrink.

---

[www.hydraraptor.blogspot.com]

Based on your blog post i made up this chart:


I use sketchup, and when drawing polygons, it measures the distance from center-to-vertex as the radius of the polygon. So if i put in 2 mm, the flat is at 1.3mm. So by using this chart, you can see that if you want a 3mm hole, you need to enter 1.7 as the radius, and use a hexagon.

After many hours trying put my Mendel printing correctly, I quited. I were also in chat with Leonard from Portugal, because he were also having the same problem and tried to figure it out. Seems there are no solution or it is very hard for a person with my knowledge/technical to put it working.

I just wrote a warning about this problem on Mendel 1st page: [www.reprap.org]

I wish I did know about this before decided to buy all the materials for assembling the Mendel...

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

Just because you are too stupid to make your Mendel print circles doesn't mean there is anything wrong with the design of it. Mine prints hundreds of circles a day and they look pretty round to me.

---

[www.hydraraptor.blogspot.com]

nophead Wrote:
-------------------------------------------------------
> Just because you are too stupid to make your
> Mendel print circles doesn't mean there is
> anything wrong with the design of it. Mine prints
> hundreds of circles a day and they look pretty
> round to me.

You mean me and all others (the ones that started this message for example, Leonardo, etc) are all stupids?

Edited 1 time(s). Last edit at 02/13/2011 12:25PM by casainho.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

Quote
casainho
I have the same problem, circles not being circles and straight lines not being straight sad smiley

How fast is your feed rate? Does it improve if you slow down?

brnrd Wrote:
-------------------------------------------------------
> > I have the same problem, circles not being
> circles and straight lines not being straight sad
> smiley
>
>
> How fast is your feed rate? Does it improve if you
> slow down?

Feed rate = 8mm/s. It is slow already so I didn't went lower, since I print other layers at 25mm/s.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

my problems were fixed by reducing the vertex count in my models, and i think a fix could be implemented in skeinforge if someone is willing to write a plugin.

I don't find anything mechanically wrong with the reprap design. If you want to print smaller features, then you can upgrade to a ballscrew based design, and pay much much more, and have to work out all the details yourself. Or you can buy a commercial machine.

What exactly do you think is the issue that is preventing you from printing successfully? If we can fix your issue it will help everyone else.

edit: i also had backlash that i fixed. I bet most peoples issues are backlash related. I am also working on some improvements that i think help reduce backlash in the first place. if you have some suggestions i can try to incorporate them.

Edited 1 time(s). Last edit at 02/13/2011 04:59PM by Buback.

Buback Wrote:
-------------------------------------------------------
> What exactly do you think is the issue that is
> preventing you from printing successfully? If we
> can fix your issue it will help everyone else.

I have a CNC from which I printed my Mendel parts. I also have an UP printer. Both of them do that same holes round.

I see others with the same problem as me and without solution.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

both your CNC machine and the UP printer use different software than a mendel. have you tried multiple software solutions?

Buback Wrote:
-------------------------------------------------------
> both your CNC machine and the UP printer use
> different software than a mendel. have you tried
> multiple software solutions?

I used the same software on my CNC -- Skeinforge. Also Skeinforge simulation and RepSnapper shows me round holes, so, the problem is for sure mechanical.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

casainho Wrote:

> I have a CNC from which I printed my Mendel parts.
> I also have an UP printer. Both of them do that
> same holes round.
>
> I see others with the same problem as me and
> without solution.

I'm just a beginner compared to most people in this forum. I started building in December and I was able to produce working parts in less than a month that I have been using for my research. Thanks to the efforts of people like you in paving the way and sharing your experience online. My software settings are still far from optimal, but I haven't been having the problems that you have. The fact that others can do the same would seem to indicate that there is nothing wrong with the design and you have a specific hardware problem that needs to be fixed.