Kossel Mini xy calibration issue

Hi everyone
I'm an engineering student so I now something about technology. Normally I can solve the issues myself or via Google. This time, however, I'm sitting for hours in front of my newly build kossel mini (even uses rails). This is my problem:

I get a difference of about 2.5-3% in x and y dimensions of the object, z-height is perfectly calibrated... I can only calibrate x or y and the other one is off by about 3%. I already have a MendelMax 1.5+ updated (so I'm not a beginner) and I used it to print the Kossel parts and tried to put them together as best as I could. I also tried to calibrate the kossel using different guides but the error still exists. Then using Rich's Marlin with "G30 A" command the difference gets even worse (up to 5%).
I can only see/measure very very little build errors and can't image that they would have such a huge impact on the build quality as others surely haven't build their delta as carefully as I have...
Also all steps per unit are the same, the current per motor is exactly 1A as advised (more doesn't change anything regarding this error). I'm using FSR currently but am changing to a proximty sensor soon. I'm using Ramps 1.4

If you could please help my that'd great. I've spent hours trying to fix this...

Seems you need to calibrate your "DELTA_DIAGONAL_ROD" on Marlin-RC you can do it on EEPROM command M666, just cant remember what letter it is (think is "D"). See "4- Calibrate the firmware to print to correct dimensions." in the link below it can help you

[minow.blogspot.mx]

Regards

It's not your DELTA_DIAGONAL_ROD, as I'm sure you're aware.

You've got a mechanical issue in your build somewhere. Good luck finding it. Sadly, Marlin doesn't have separate X and Y scaling for deltas, unlike RepRapFirmware. But for RRF you need to spend hundreds of dollars on a 32 bit board and associated peripherals. So you're left with getting your build perfect.

I've recently had the same tedious, frustrating issue, which you can read about here:
[forums.reprap.org]

I'd recommend trying to print the object I posted in there, which might help you to get closer to finding the issue.

As your dimensions are a long way off each other (3%), you should be able to find the issue with your build without too much difficulty, it should be measurable with normal instruments.

Here's the list of things I did to find the issue on my build:

• Rebuilt diagonal arms, all are closer than I can measure (<0.1mm different)
• Spaced out diagonal arm mounting points so that all are mounted within 0.1mm (within tolerance of what I can measure, measured when mounted, I find these difficult to measure properly)
• Adjusted frame so that linear bearing tracks are parallel (< 0.05mm difference between top and bottom of tracks, measured between towers)
• Adjusted frame so that linear bearing tracks are equidistant (< 0.05mm difference between between towers)
• Ensured that all towers are at exactly 90 degrees to the bed
• Adjusted steps per mm in the firmware so that all carriages move exactly 100 mm when I move the effector down 100 mm using Pronterface. All are between 80.1 and 80.2 steps/mm.
• Rewired motors and endstops so that X became Y, Y became Z etc to see if it was a firmware issue. The problem stayed with the hardware, it didn't rotate around.
• Adjusted steps per mm in the firmware so that all dimensions were spot on as printed (Z was at 79.6 or so), but this introduced other errors such as a crinkled print surface which meant that I couldn't use this as a solution. I put the steps per mm back to how it was, so the carriages moved 100mm when commanded.
• Made up a trapezoidal piece that clamps to the effector in various positions so that I can see effector tilts of less than 0.5 degrees
• Printed a new effector, and carefully filed the ball joint mounts so that there was less than 0.1mm difference between the carriage and the effector for each diagonal rod pair
• Put a crush piece under each carriage, and adjusted screw tension so that the ball joints were at exactly parallel to the other two rails (they were a fair bit out, quite a few degrees off parallel)
• Tightened up the mounting of the hotend where it mounts to the effector
• Rotated the effector + diagonal rods 120 degrees to see if the problem was with the effector, or the geometry of the machine.

It's a long road, but worth it. I'm now within 0.1% of each other, and 0.2% of absolute scaling, which is acceptable to me.

Edited 1 time(s). Last edit at 02/24/2016 09:07PM by nebbian.

Thank you very much, especially @nebbian

I do have some quite exact instruments so I'll try to measure everything again. Yes DELTA_DIAGONAL_ROD doesn't have an impact on this issue which is logical. At Friday/Saturday I'll give it another try.

Today I had the following idea where the problem could come from: Even though the towers have the same measured (diagonally) distance from each other the corner brackets might still have build errors (the printer which printed them wasn't maybe calibrated perfectly?). However, why do I get no errors in angles (printed the 120° calibration object several times). Long question short: Does a not perfectly but good calibrated printer when using it to print all the kossel parts have such an enormous impact?

As I'm quite into programming I had the idea to rewrite some Marlin code regarding the movement control. Why couldn't I just calibrate x and z and then set an appropriate multiplier directly to the y distance-coordinate? I need to test this but the question remains whether this has an impact on the other coordinates? I do think this opens up a new set of problems as this is 3D-trigometry and due to higher moves of some parts to the outer end of the print bed the multiplier wouldn't be const, wouldn't it?

tbh I don't think that inaccuracies in the printer that printed the corners will have much effect, because you adjust the various lengths to get a perfect equilateral triangle top and bottom. Any inaccuracies will be taken care of when you do that adjustment.

However inaccuracies in the dimensions of the effector may well cause lots of issues.

First thing I'd do is check for effector tilt, using something like this:



Bad effector tilt can really throw off your printed measurements, especially if your hotend hangs down a long way from your effector.
Let me know if you want the stl for this piece.

I can't see any reason why you couldn't apply a mapping function to the desired X and Y position of the effector, before it goes through the delta kinematics. I think that this is something that should be present in Marlin, and being a programmer myself I was going to add it as well... until I sorted out my mechanics. tbh Marlin is pretty bad from a software engineering point of view, I'm in the process of migrating to Repetier to see if it has better delta support. You've probably already noticed that the effector slows down towards the edges of the bed -- it shouldn't do this. Apparently Repetier has proper delta kinematics that work properly.

Edited 1 time(s). Last edit at 02/25/2016 03:18AM by nebbian.

@nebbian yes I'd like the templates. I've already tried Repetier on a Delta. It definitely has different kinematics and better resource management and I'd prefer it on any catesian printer. On a delta however there are disadvantages as well: The auto calibration isn't a good as the marlin derivates people use these days. And it needs some adjustments for circular printbeds sometimes.

I'm going to take measurements on the carriages/the effector as they might not have the same hight in all corresponding points.

Good stuff Blueman, thanks for the info on Repetier.

I've published the tilt meter here:

[www.thingiverse.com]

Good luck finding your issue.

I have the auto calibration on my printer but when I print the test box or test wall, prints (kind of fast) irregular squares and starts from small to wider on top, see picture.

I got my previous problem fixed and one of the problems was the teflon tube was getting stuck on one of the arms when auto calibrating, now I have other issue and for some reason the tip of the extruder smashes the plate when printing and doesn't matter if I change the numbers, I had 300.20mm then I kept reducing to 298mm and still doing the same, does the auto level thing but when starts printing the tip smashes the glass so much the filament will not come out.

Try adjusting the Z probe offset. You can even do this from the front panel (at least on Marlin). Make the number bigger for more squish, smaller for less squish. In your case make the number closer to zero.

That worked, thanks!!! now I need to get a better print out, when I print a thin square 0.4 mm is not straight at all, I took a picture and the one from the left was the first one and the one on the right was the last one I printed out after changed the offset, still a lot of calibration to do.

You're making progress at least

Try increasing nozzle temperature, and also increasing flow rate. It looks like you're underextruding. Have you calibrated the extruder steps per mm?

I used the firmware that came with my kit, also I decided to have a fan blowing the part that's been printed because when was printing the part looked too soft and worked

Edited 1 time(s). Last edit at 07/21/2016 08:19AM by slanwar.