Issues with Parallel walls

[Aussie Thumper]

Hi Guys,

I've had a bit of a search through the forum topics and haven't been able to find anything similar so I apologise if this has been covered already

I've had an Orion for almost a year now and for the most part I've been more than happy with the printer. Unfortunately I've recently uncovered a bit of an issue with printing parallel walls which has gone un-noticed until now when I've been trying to fit two parts together.

What I'm seeing is when I print a test block (100x3x5mm with the 100 long printed perpendicular to the z tower aka horizontal) is a deviation from one end to the other.
One end measures 3mm and the other end measures 3.3mm.

Now one 100mm long side (the rear most side) prints perfectly straight and the other 100mm long side (closest to the printer control screen) prints the first 30mm or so at 3.3mm, the last 30mm or so at 3mm and the mid section seems to taper between the two.
As the print head comes around the right hand side 3mm wide wall and turns 90 degrees to start the front 100mm long wall the print head seems to flare out forming a hump for the first 30mm or so and then back to normal.

Close up of print head as it rounds the corner. Also shows "hump" which is creating the issue.

Now interestingly when I print the same part at 90 degrees it prints near perfect (0.04mm deviation)...

I've been in touch with Seeme CNC support over the past few weeks and the help has been absolutely great however its proving a bit difficult to track down the issue.
We narrowed it down to a mechanical problem with a carriage and unfortunately after upgrading the arms and carriages to the ball cup injection molded components the problem still exists.

Thus far I've:
-Replaced the stock melamine carriages with the upgraded injection molded set and also upgraded the arms to the upgraded ball cup versions. This helped, it took about 0.2mm of error out of the prints (it was printing 3.0mm to 3.5mm over the 100mm) however it didn't solve all the error.
-Tried to maintain an even amount of belt tension on all three belts as this seems to account for slight errors if they are too tight or loose.
-Last weekend I individually swapped out each stepper motor (x y and z) with the extruder stepper motor to see if I had a dodgy stepper but they all seemed to print about the same.
-I asked Seeme CNC help to print my GCode which they did and got a slight variation (0.17mm)

I'm currently using:
-Orion with 0.91 firmware
-Upgraded cheapskates (now running the injection molded set)
-Upgraded Ball joints and arms
-Solidworks to create 3D data
-Mattercontrol and matterslice as the slicer
-Printing PLA at 210deg C.

So I guess the questions I'm asking are:
1. What sort of tolerances should I be expecting on vertical wall thickness? (my assumption is it should be better than 0.3mm, more like 0.1 or less, correct me if I'm wrong)
2. Has anyone else experienced anything like this and if so what was the problem?
3. Any suggestions on what I should be checking?

Anyway, any help to solve the issue (or potentially multiple issues) would be appreciated

Thanks guys

[bot]

I've experienced the same thing, and so have others. I have posted in a thread about it before, if you're willing to dig through my post history to find it you could see what didn't resolve it. My max still has this problem. I built an entirely new machine, because this problem seemed inherent to my max. I think it's warped melamine frame.

[Aussie Thumper]

Thankyou Bot

I'll have a look and see what I can find!

[nebbian]

I've been through the dimension saga myself, it takes weeks of careful measurement, adjustment, hypothesising, noticing that the error is still there, trying something else, having a sleep, and general problem solving to get to the bottom of it.

Things I would look at:
* Take careful measurements of the distance between the towers, top and bottom, to ensure that they are all parallel.
* Carefully ensure that the top and bottom triangle are not twisted in relation to each other.
* Print my effector tilt meter: http://www.thingiverse.com/thing:1369444 and see if the effector is tilting in various places on the bed
* Rotate the electronics by 120 degrees (X motor connections and endstops move to Y, Y to Z, Z to X) and see if the problem also rotates 120 degrees
* Rotate the effector, rods and carriages by 120 degrees and see if the problem also moves around 120 degrees

Good luck, it's a tricky thing to find.

[Aussie Thumper]

Thanks Nebbian, its on the print bed as we speak
I've already been through the rotating the electronics saga without any success along with checking dimensions and adjusting towers.

I'll give the tilt meter a go and see if there is anything obvious there.
Thanks again

[Eaglezsoar]

Have you tried a different slicer?

[JFettig]

Its been a problem for me on the maxv2 as well, slicers don't matter, etc. Its always big on the left for some reason in the y dimension.

[Eaglezsoar]

If it is happening on two different printers then it would appear to be something in common with the printers. The odds of the same exact problem happening on two different printers
is quite small but what is in common?

[bot]

I'm pretty sure it happens on every single max v2, but not everybody notices or is willing to admit it.

[Eaglezsoar]

I'm pretty sure it happens on every single max v2, but not everybody notices or is willing to admit it.

Wow, that is quite a statement, one that I do not agree with at all but right now I cannot verify. What does the rest of our community think?

[Xenocrates]

I had similar weirdness happen. I believe that it may be caused by an error in the kinematics somewhere, or how the slicer handles things crossing the origin. Perhaps the best solution to figure stuff out is to play with it. try printing three or four at once, one on the origin, one along the X axis, one along the Y, and one in quadrant 1 if possible to fit them all on. Once you've done that, perhaps pursue a different firmware to test. Perhaps Mike is in the best position to chase down a bug like this, thanks to the sheer variety of hardware/firmware he can use, and the large number of slicers he routinely uses. It's a very curious issue, and logically, there's no reason for it to come up.

It may also pay to check the G code for parts like this.

[Aussie Thumper]

Thanks guys for the responses.

To answer some questions:

1.

* Rotate the electronics by 120 degrees (X motor connections and endstops move to Y, Y to Z, Z to X) and see if the problem also rotates 120 degrees

I decided to change the stepper motors (and corresponding end stop switches) along one axis prior to physically dismantling anything.
Basically I pulled all the stepper motor plugs out of the RAMBo board and put the X axis stepper in the Y port, Y axis stepper in the Z port and Z axis stepper in the X port then re-printed the same 100x3x5mm print.
I repeated the swapping of stepper motors, end stops and re-printing until I was back at the beginning.

Results are below, I measures each end of the 100mm long X 3mm wide print.

When:
X axis stepper motor plugged into X axis on RAMBo - variation of 3.14mm to 2.81mm (0.33mm) - standard Orion configuration (I'll call this the baseline print)
X axis stepper motor plugged into Y axis on RAMBo - variation of 3.14mm to 2.89mm (0.25mm) - x axis rotated counter clockwise 120 degrees
X axis stepper motor plugged into Z axis on RAMBo- variation of 3.07mm to 2.95mm (0.12mm) - x axis rotated counter clock wise 240 degrees

So basically the problem exists in all axis however the horizontal print (perpindicular to Z tower) is by far the worst.
Now interestingly I can print near parallel walls (deviation of 0.07mm) when I print vertically on the print bed (parallel to Z tower). And I can replicate the results when I rotate the print to be parallel with each tower.

So thinking it could be a dodgy stepper (and running out of ideas), I bit the bullet and changed the Z axis stepper with the extruder stepper motor which proved to make very little difference
3.13 to 2.84 (0.29mm)

I then changed the Z axis stepper back to the original stepper motor and replaced the Y axis stepper with the extruder stepper motor.
This also made little difference, slightly improved on baseline.
3.13 to 2.85 (0.28mm)

Becoming increasingly frustrated I changed the X axis stepper with the one removed from the Y axis and once again, it still printed the same as baseline.
3.17 to 2.84 (0.33mm).

So based on the above I think I can exclude the stepper motors from being the problem. As for the remaining electronics, I don't know.

2.

* Print my effector tilt meter: http://www.thingiverse.com/thing:1369444" onclick="window.open(this.href);return false; and see if the effector is tilting in various places on the bed

Effector seems to be running flat and true based on the tilt meter

3.

Have you tried a different slicer?

Yes, Being relatively new I've only tried matterslice and slic3r and unfortunately as JFettig said, it makes no difference.
Is there a slicer that the Delta printers seem to prefer? Or is it personal preference?

4.

It may also pay to check the G code for parts like this.

I've been in touch with Seeme CNC support and so far when they print my Gcode they get a deviation also (2.92mm - 2.75mm = 0.17mm).
When they print my STL (and slice it themselves) they get a deviation (5.40mm to 5.07mm = 0.33mm).
When they create an STL, slice it and print it they get a deviation (5.20mm - 5.06mm = 0.14mm).
When I print their Gcode, I get a devation (5.35mm - 4.84mm = 0.51mm)

5.

* Take careful measurements of the distance between the towers, top and bottom, to ensure that they are all parallel.
* Carefully ensure that the top and bottom triangle are not twisted in relation to each other.

I'm thinking based on your feedback that there could be an assembly issue in my Orion.
My next checks are going to be dimensionally, I'm waiting to catch up with a friend to borrow his set of 300mm vernier calipers and do some measurements on the towers to see if they are twisted or out of square. I have already checked for a dimensional error however not with the verniers.
If that fails I'm thinking I'll have to tear it all apart and inspect all the components.... not looking forward to that job

Thanks again for your help guys.

[nebbian]

I feel your pain. Hang in there, you'll find it eventually if you stick at it.

I really think at this stage that it's something that's wrong in the physical geometry of your printer. You need to get all towers absolutely parallel to each other (within 0.05mm), and absolutely not twisted.

Good luck finding the issue.

[bot]

I've never heard of someone figuring out this issue and resolving it without building a new printer. Also, getting the frame components positioned with a tolerance of 0.05mm is literally impossible unless you keep the thing sealed in a temperature controlled environment so that there are no effects of thermal expansion on the tower extrusions. The extrusions will grow in length by about .2mm when going from 20 degrees C to 30 degrees C.