A few issues...

[stevewm]

Rostock Max v2, with Rev 6 heated bed. Purchased August 2014

As I have tried to start printing out things with fit together parts I have noticed nothing is dimensionally accurate. Up until now I've just dealt with it. Que massive amounts of sanding and filing to make parts fit together...

I have tried a few things to fix this, based on searches on this forum:
Calibrated the bed to print flat multiple times, just to make sure. Have tried using the end-stop screws and the software offset method.
Calibrated the E-Steps (using the helpful doodaddoes video on YouTube)
Measured every filament roll's diameter 10+ times and use the average in Slic3r.
Made adjustments to the delta rod diagonal length. (using this page: http://minow.blogspot.com/ as a reference)
I've checked and rechecked belt and cheapskate tension.
Checked that the stepper pulleys where aligned and set screws properly tightened.

Objects are still coming out with the wrong dimensions. And oddly enough, everything on the left side of the print bed (looking at the front of the printer) is much worse than the right side.

I have been using a 100mm X 6mm X 3mm calibration object made in OpenSCAD. Printed at .2mm layer height, with the first layer at .3mm. Using Slic3r. Using RepetierHost 1.0.6 and current firmware from SeeMeCNC's website.

Top view. Excuse the crude MS paint notes.. As you can see, some of the dimensions are off. Length is correct enough (only after adjusting the diagonal rod length). Z/height is off, but consistent across the entire object. And the width is off, but it varies from one side to the other.

Bottom view. You can see right in the middle of the object where it starts to really change. Looks good on the right side, and gets progressively worse to the left. Even though the right side looks good, it is still .36mm bigger than it should be.

I've printed multiple objects now. The actual print quality is generally good. Infill meets the walls, outside walls look great, top infill looks good. Things just never fit together!

Any help is greatly appreciated!

[RocketMagnet]

A really quick check is to try a couple of different Slicers and see how they fare (outside of plethora calibration issues which does look like your issue but...)
In my experience so far no slicer is the best and different parts print out better in different slicers.
Yes there are numerous variables over the slicers that are implemented differently and tweaking in each individual slicer can improve results
but sometimes using another fixes all your issues.

Personally on thin walled items i'll use Cura, seems really good with the stock SeeMeCNC settings but for most other stuff I'll use slicer.. it's often slower (not always) and is really
sensitive to filament diameter (seems to express itself in blobbing and tiny gaps in the outer walls of items).. Cura doesn't suffer from this as much IME and is more tolerant yet it's poor
at bridging and I find it requires a higher density infill usually to get good results but it could be due to the fact I've less experience with Cura than Slicer...

A simple test I use is the Nut and Bolts test. I thought I had an issue with machine but it turned out to be Slicer, my bolt test stopped fitting but it was a setting I'd fiddled with as printing with stock settings in Cura produced a good fit so I essentially proved nothing was seriously out of whack calibration wise.

Sure you'll get plenty of calibration advice it's just worth a quick play with some other slicers to see how they compare.

[stevewm]

I finally got a chance to try different slicers, same issue. It is exactly the same.

It has got to be something mechanical, but I am not sure what.

All the belts are adjusted well, the cheapskates are adjusted well. All the u-joints move freely. etc..

The issue appears to be either at or between the X and Z towers. Any part near or between the X and Z towers will grow bigger in the X and Y dimensions.

Anyone have any other ideas?

[JFettig]

I have similar results and I'm trying to figure out what all it is, I have my tower rotations, rod length, horizontal radius all dailed in so my bed is within .001". What did your rod length come out as? Mine is 267.8mm. My tower rotations aren't all that far away from where they were this morning so I'll see where it comes out. I do know that 267.8mm won't give me a 100mm long part, it will still be .5-1% short but the short dimension will be close(a little large), and it will vary depending on what side of X0 I'm on just like yours does.

I just hit start so we'll see what it prints out, I'm going to slice the same part but at 90 degrees to see if my results differ.

[Polygonhell]

If you have differrence results for small and large objects, the odds are you still have some issue with the delta radius.
I personally would be very careful with messing with the rod length, I used to recommend this, and if you made your own arms, I still would, but with the cast arms SeeMe ship, there shouldn't be much variance, certainly way less than a mm.
Also if you mess with the rod length, you can cause other issues, it also introduces a small amount of doming, that coupled with the correcting it using delta radius can introduce a saddle like shape to the effector motion, but there are many other things that introduce the same issue.

To the OP the fact your height is of implies to me your over extruding. With a 0.3mm first layer and 0.2 mm layer height I'd expect 3.1mm height, but you'd usually see something slightly under that because most people squish the first layer somewhat.

You probably also want to print something at least 10mm high and preferably 20mm, there is a phenomenon where if the bed temperature is set too high, the first 5-10mm of a print can display an hourglass shape, often referred to as an elephants foot, and it's best to measure above that area to rule it out.

[JFettig]

Chatting with geneb, we decided that we should swap motors - but swapping motors would be no fun, I'm going to try swapping my connectors after I'm done with this print, switching X and Z axes so it prints sideways - if you do this don't forget to recalibrate.

Polygonhell - delta radius is set up to calibrate dishing right, AKA horizontal radius? I had it set to hold the area between the towers to .001" before I started changing rod length, I could imagine that adjusting that will give you some X-Y sizing/scaling but it'll also cause dishing or doming. Rod length adjustments do change dishing or doming a small amount but not much. I had to change my horizontal radius by .05

OP - did you change your steps/mm? never change this value, its not a value that you can adjust and expect to get good results, changing it will result in scaled Z. It could be simply setting your Z0 and first layer height combined with fill and measuring accuracy.

[JFettig]

I did a quick test, printed at 0, 90, and 45. The 90 degree print is the same on both ends, and the 45 degree print is large at the opposite end! X positive, Y negative from zero instead of X negative and Y0. Kind of strange in my mind.

Z arrow is pointing towards the z tower.

[bot]

I've recently noticed this issue as well, and am racking my brain trying to come up with a solution. Any more thoughts on this?

[Eaglezsoar]

Are you sure that you calibrated your cold extruder so that when it is supposed to output a known amount it actually does?

[bot]

Yes.

I think this is clearly a mechanical issue. Over many different prints, this issue is consistent and also consistent across machines miles away from each other. I have the exact same symptoms as JFettig. If it were extrusion settings, the problem would be different for every print, no?

I almost think it has to do with the Y tower motor axis being mirrored. How is the motor reversed? Perhaps the floating point math for the calculation of that one axis is wildly different than that of the other axes?

[stevewm]

Are you sure that you calibrated your cold extruder so that when it is supposed to output a known amount it actually does?

Same here... Have done it multiple times.

I recently started over entirely from scratch; cleared the EPROM settings, reflashed the current SeeMeCNC Repitier build. Re-checked tower squareness, re-did radius and extruder calibrations. Problem still remains just as before.

[bot]

I've checked almost everything and come full circle. I think I'm going to order new arms, as the ones I have the holes are slightly too large on some of them.

[stevewm]

Sorry to resurrect a old thread, but I think I figured out my problem! And boy do I feel stupid...

Can you spot the problem here? (this is on the Y tower)

Yeah.... Built the printer almost a year ago now. Have went over it countless times, and never once did I notice one of the arms was upside down. Everything opposite of the Y tower prints distorted the further you get away from the center of the printer. This has to be the cause. Going to correct this and re-calibrate.

[geneb]

Did flipping the arm fix the problem?

g.

[stevewm]

Haven't yet got a chance to do so. Hoping to get to it later tonight.

[stevewm]

Well crap, that didn't fix it. Still doing the same exact thing.


Printed the same object as in my first post. Same results. The right side (towards the Y tower) measures almost spot on, 6.08mm wide while the left side towards the X tower is 6.6mm wide. The height and length of the object is spot on as well.

I have also noticed that on the left side, the object gets wider as it gets taller! At the bottom it measures 6.09 and increases in width as it goes up, measuring 6.6mm at the top. The right side of the object is uniform!

Below is a top down view of another object I printed. A small AAA battery holder from Thingiverse.. Notice the gaps between the walls in the purple boxes? When the object starts these walls meet, but as it gets taller, the gap/distortion also gets bigger. On tall enough objects, some distortion also starts to appear in the bottom right... It is not a slicer issue, I get the same results no matter the slicer used.

Any one have any more ideas? I am really at a loss here. I can print trinkets and vases all day long, but anything that needs to be anywhere close to dimensionally accurate is out.

[geneb]

This is going to sound nuts, but bear with me...

Swap the Y arms with the X arms and see if the issue moves. If it doesn't, swap them back and try swapping X with Z. There MIGHT be some issue with the arms that's causing this. I'm wildly grasping at straws at this point.

You should also check the Acetal u-joints to make sure they're not loose on the axles, nor on the arms.

Also, check all six axles to see if any have been bent a little - especially on the X and Z axes.

g.

[stevewm]

Thanks.. I'll give that a try later tonight.

[bot]

It's not the arms. I've swapped arms every which way, gotten two new sets, put them upside down, backwards, inside out, you name it I've tried it. The problem is not with the orientation of the arms.

I still think it has to do with the Y axis motor being reversed.

[geneb]

bot, if the motor was reversed, it wouldn't operate properly at all.

g.

[bot]

Well, what I mean is that the y axis has the motor on the opposite side of the tower relative to the other two. I should clarify that I think it has to do with a combination of the astrosyn dampers with the motor being on the opposite side.

[geneb]

It's not physically possible to do that on the v2. The X and Y motors "face" each other across the front of the machine. In order to mount one backwards, you'd have to install the whole mount backwards and that's not possible due to the tabbed design of the motor mounts.

g.

[bot]

You are missing my point. The design of the printer has the motors on opposite sides of the tower... this isn't a mistake, this is the design. By doing so, and with having the astrosyn dampers, any flex imparted on the y axis belt motion due to the dampers will be opposite to that of the other towers. This is where I think the error comes from. This is just a wild guess, though.

[stevewm]

I don't have any dampers installed. So its not that, at least on mine anyways.

[bvandiepenbos]

You are missing my point. The design of the printer has the motors on opposite sides of the tower... this isn't a mistake, this is the design. By doing so, and with having the astrosyn dampers, any flex imparted on the y axis belt motion due to the dampers will be opposite to that of the other towers. This is where I think the error comes from. This is just a wild guess, though.

The pulleys turn the same direction regardless of which side motor is on. So I don't see how that could make any difference.
The Astrosyn dampers do negatively affect print quality from my expirence. They let motors wiggle.
Take them off and see if print quality improves.

2mm thick Cork or rubber washer shaped like stepper does quiet machine but does not let motors wiggle.
Wonder if nijaflex printed spacer would work?
You will need longer screws, m3 x 14 mm works with 2mm spacer.
Remember to move pulleys out on motor shaft.