…after I’ve qgl and performed a bed mesh I always find that the left rear of the build plate to be always slightly lower than the rest of the plate.
After watching the printer perform the QGL I’ve noticed that when it’s adjusting the heights it seems to incorrectly adjust that particular corner for some reason!
Why would you assume it is wrong? Please show us your calculations based on the figures from your QGL probing?
Why did you start looking at this, is it printing wrong? Does something look wrong when you are printing? Or is it just the figures? If it prints ok, no need to fix anything IMHO.
Sorry to come with these questions but a micron is really small. Those numbers coming back from the probing need to be taken relative to the real world accuracy and deviation that mechanical systems can achieve.
For instance my klicky probe will step up in range by an order of 10 microns simply from being used every cycle. This is the effect of it being electrically used.
It is however nothing of concern and nothing to worry about. It is within its normal standard deviation being a mechanical switch.
What I’m having trouble with is the rear left part of the bed as it’s always too low even after performing a heat soak, QGL and mesh. It never seems to ‘adapt’ to that corner!
How does it print? Do you see a noticable difference when you perform a bed level print of just one layer in that corner compared to the others?
I understand you are mentioning the numbers and that something “feels” wrong about the adjustment it makes, but you need to put a little bit more meat on the bone so we can evaluate if your machine is doing something it is not supposed to.
My machine does the same thing from time to time it also throws a repeated out of tolerance at me in the same corner, always rear left. But for the moment I cannot find why, it is probably something about the build of that gantry that is not as it is supposed to be, and after a retry of the QCL it is ok again and the prints are never showing anything amiss with the gantry level or bed mesh. So I leave it for now and when I am rebuilding the machine with new belts and metal parts instead of the printed parts, I will pay attention to this to see if I can find what was going on.
The machine prints very well. It’s just in that left rear area which when it puts down the first layer there’s noticeable gapping between the lines.
I’m maybe just been picky!
So you said above that the bed in the other axes (I think you meant corners) are higher. So the gantry should also be higher in those places to be level to the bed. This is why it adjusts the corner of 1.948** down by -0.001**.
This is exactly what I would expect from the QGL.
Now you say that in that corner it shows gaps in the lines of your first layer. This means to me that in that corner the gantry should go down even further. It is not doing it wrong, according to your info you just gave us it needs to move even more.
I would look at if your bedmesh is working correctly and in particular if the values are saved correctly and the system is not using an old version.
At the bottom of the printer.cnf file there are some old bed_mesh saved from previous setups! But when printing it does state that’s it’s printing from “adaptive mesh”
Here is the printer.cfg printer.cfg (36.9 KB)
Ok hands up I cannot find why it would print in that corner with a higher nozzle height then elsewhere.
I can only suggest to go through all the points here, Bed Mesh - Klipper documentation , with a fine toothed comb, and make sure all boxes are ticked correctly.
The other thing I would try is to remove all old values of bed meshes, reset the cfg to the default and do a bed mesh to see whether it is actually storing it correctly and in the right place. There is maybe a slight chance that you think it is using the fresh values everytime but in fact it has found another saved set of figures due to some cfg error somewhere. Adaptive mesh should be ran every time for every new print, so if you do not print so far in the corner it would not even be a problem, then when you do go to the corner on the next print, and it is not using the fresh new set of values, it could be it is printing there without any compensation at all, or plain with the wrong compensation.
I do not know but maybe some other person with more experience can have a look as well.
