VOAC calculating offset wrong by the same amount every time
Just like the title says, my X offset between T0 and T1 is always wrong by 0.9mm in X (Y is perfect). Any idea on what would cause this? My frame square and both toolheads print really well. Skew calibration prints close to perfectly for each.
37 Replies
Best guess is that your pxprmm setting is off.
I fixed it temporarily by altering the idex_xoffset in the ratos-variable.cfg.
How does the VOAC calculate offset? I am guessing it is using the x endstops? Problem with that is there a ton of tolerances involved that can build up in the wrong way.
by checking how many px's you dragged the image to align T1 relative to T0.
It's not using the x endstops.
Those are only used for homing
Ah pixels/mm. But I am lining up the the nozzle diameter really closely. Hard to believe I am almost 1mm out. I do the calibration until it repeats perfectly. I do notice that sometimes when I drag the nozzle in the image, it doesn't move completely in terms of the amount I move it. It does at first but repeated moves without changing from t0 to t1 or vice versa result in ever smaller movements relative to how much I drag it. Almost like there's a buildup of error, perhaps in the software?
from what I've noticed in the past is that it works better if you do only one move per toolhead at a time. doesn't matter if it's not perfectly aligned after one move, just switch, do one move, switch back and continue like that - get's me aligned in 2-3 moves typically
Yes I started doing that as well. I just find it interesting that the software is doing that. It seems like a bug. But maybe not, you can fill a lot of libraries with stuff I don't know lol At any rate maybe it's related to my issue.
Almost like there's a buildup of error, perhaps in the software?No that's because your pxprmm is wrong. Make sure you've set the correct nozzle size in the configurator, you might be trying to match an improperly sized circle because you haven't told it what nozzle you're using. The inner ring is the important one the outer ring can be set via the diameter input in the settings dialog
Yes I have this exactly matching. It has to be something else.
How far off would I have to be to be creating a 0.9mm difference?
Can you show me a screenshot of the nozzle matching the crosshair with the settings panel open?
Depends on how much adjustment you need, and how many times you drag it because it doesn't move to the spot you dragged it to (because the pxprmm calibration is wrong)
If the nozzle doesn't end up at the crosshair, your pxprmm calibration is wrong.
Each time you drag the image, you add error.
Will do in a couple hours I am printing something until about 630 pm est
The last pic is an IDEX calibration I just printed. If you're not familiar with this model, it demonstrates physically where the two idex toolhead are in relation to each other with T0 as the reference (orange) and T1 as the measured offset (yellow). The first tick closest to x or y is the zero mark where the offset would be perfect. In my case, the Y axis is pretty much dead on. On the x axis you count in .1mm increments from left to right (starting at zero) until a yellow tick lines up with an orange one. In this case it's about -0.7mm to -0.8mm out and this after the ratos offset is applied, in my case -0.14565 for a total of what it should be is -0.84565 to -0.94565 or about -0.9mm
Double checked configurator I have it set for a .4mm nozzle which I am using on both toolheads.
What is the offset compared to? Obviously the IDEX_XOFFSET is not the actual distance between toolheads. The VAOC routine moves the toolheads a specified distance and then compares where the nozzles are relative to that? So for whatever reason I'm guessing that relative distance is about .75mm different than what the software calculates. If the T1 endstop was out of position by .75mm along the X axis, the software would have the toolhead in a different place in X than it actually is relative to an assumed X max. I am not sure how you keep the endstop location out of play here. Granted that doesn't mean you don't keep it out of play, I am just not sure how you do it.
VAOC can measure the offset only behind the bed at one single spot. if you have a gantry skew then you will have such printing issues. VAOC can take a skew calibration into account, you jsut need to activate a profile for it in ratos
Well I read the new commissioning guide and when I went through the gantry check it was out on the right side by 3mm! So I have a badly skewed gantry. Prints great like that lol. Anyway I corrected the issue by loosening up the screws on the x gantry plates to the Y carriages, and also loosened up both Y linear rails and re-aligned everything. Now I have it all back together and the gantry is perfect in the back and front. Going through IS now and when that's good I will do the initial setup for VAOC over again and then check if the offset is correct. Report back when done.
Ok I've squared up the gantry and redid the shapers, did VAOC, did skew correction, redid VAOC and the offset is now incorrect by -0.5mm in X. Y continues to be perfect. Where do I go from here? In the mean time I will alter manually the IDEX_XOFFSET value in ratos-variables.cfg by adding -0.5 to whatever value VAOC calculates.
what so you mean the xoffset is oncorrect?
After running VAOC it writes the ratos-variables.cfg. The value it wrote for idex_xoffset is 0.023689 and if I use this for a dual extruder print T1 is off by exactly .5mm (too close to x=0 by 0.5mm0). If I add -0.5mm to the idex_xoffset varable in ratos-variables.cfg and reboot, if I run a dual color print it is perfect.
then you still have some skew in your system
Before I corrected my skewed gantry it was out by -.9mm
I did the check per the new commision guide and the gantry is perfectly square front and back. The frame is very good also. Bed without magnet and flex was 0.14mm high to low, and skew is minimal XY .06 degrees, XZ .02 degrees and YZ .1 degrees
so its not the xoffset frmo vaoc thats wrong, VAOC can only measure it at the position where it is lcoated. every tny skew in your system or belt tension issues will be reflected in the accuracy of it onver the build palte
I've got really good shapers too. All MZV and some are reccommended ZV (but I use MZV anyway).
shapers have nothnig to do with it
its not related
Just referring to belt tensions indirectly.
well, good shaper graphs on a idex can be a indicator for too loose belts
IDEX shaper grpahs can be super misleading
especially if the y belts are too loose
They are not losing position during prints fwiw
Also the negative .5mm error is not varying
Anyway I've exhausted everything I can do with it. I will simply modify the xoffset until maybe I can figure out why it's out. I can't build a printer any better than I built this one. Not that it can't be done, just I can't do it.
@Helge Keck I forgot to mention something else. When I ran the zoffsets calibration, you have to run the XY calibration first and then run the zoffsets. Well here's the interesting part, both extruders hit the exact dead center of the Z endstop button. If one of the toolheads was off by .5mm it would be very easy to see.
Also no matter where I print on the bed, a dual extruder print is out by the same amount
Now I saw your thread on the same issue 🙂
Did you check if placing the print in different locations regarding Y distance fron the VAOC changes the offset? It would confirm if it's scew in the gantry. Going further should increase the offset error. (Just theorizing, as I've not fixed mine, waiting for the morning to come).
I had a similar issue where my Y was good for the IDEX and my X was off. I did the IDEX calibration test at all 4 corners of my 500 as well as the center and it was noticeably tracking worse farther away from teh VAOC. In my case, it ended up being that I had my X-axis belt tensioners adjusted differently (trying to get the IS graphs the same). Once I adjusted the belt tensioners to the same length (and have exact same belt lengths as suggested)
- that made the difference.
Sorry for the stupid question, but which parts do you refer to as X-axis belt tensioners?
The lower ones
If you read the message I posted directly above yours I mention that the xoffset is out the same everywhere on the bed. But I have changed a lot since this thread was started so I will check for that again.
I did a test print with two colors next to each other (pictures in Idex thread) and the 350 mm squares printed are with equal diagonals when the skew value is 0. In my understanding, this should mean, that the frame is good. The issue is I have 0.3 mm offset difference between the two colors in the rear end of the bed, compared to the front end. There should be a software correction to that, but I can't find the correct code. I'm bad with coding, so trying to make ChatGPT generate something, unfortunately, it can't handle it.
I just ran this simple IDEX offset measuring print. I placed the model in all 4 quadrants (by hand approximately) of the bed and applied my normal -0.5mm adjustment to the VAOC result for idex_xoffset in the ratos-variables.cfg file. All four prints show perfect x and y alignment for the two extruders.
@Helge Keck Obviously y offset is compared to zero which is what the y distance should be between T0 and T1. Can you tell me what the x offset is compared to and how it is calculated?
Knowing this might help me narrow down where the -0.5mm error is coming from.
the x offset is basically jsut the difference between your T0 and T1 endstop position
Thanks
Is it possible that the VAOC xoffset misalignment is caused by an out of tolerance relative distance in X between the toolhead endstops and their nozzle tips? It is combining for 0.5mm larger than expected, so when the offset in X is applied the effect is the actual T1 toolhead nozzle tip location is 0.5mm closer to Xmin than expected.
Used GATS to tension my printer. Used 3.07KG on each belt pulley. Gantry square front and back. Made belts all the same length. Still -0.5mm out on the VAOC after redoing the commissioning guide.
It's not possible having the same offset everytime if it's was a tolerance issue, no. All of the offsets are relative to the endstop trigger positions. There are no hardcoded expectations here beyond what you put in your limits in printer.cfg.
Remind me, when you say "Still -0.5mm out on the VAOC after redoing the commissioning guide.". How and when is that measured exactly? Use as many details as possible.
I redid GATS again to 3.3kg and now the error is -0.3mm so it's getting closer. So it's obvious to me this is a problem with my printer, as you stated all along.
This is why I start with VAOC and use the same second test you have there. I even made a third test that is an articulated joint to fine tune.