Fix the problem at its origin. All you do is add more headache by bypassing the actual problem. Post some pictures of what your machine is doing here (as many examples as possible), and we will try to help you find a good solution.
Mark. EXCELLENT example. My tires take Nitrogen. If it is a few pounds, I would remove the air. My neighbor’s son has no problem bringing home a tank of Nitrogen at 130psi. Adding to both tires is free and easy. If I add air while on the road, the tire needs to be purged, filled, purged, filled until it is a VERY high percentage of Nitrogen. The dealership charges a BIG fee, then does all four tires, checks the battery, fluids and washes the car.
The person they hire to do all of these tasks has no certification in automotive repair. They simply do what someone tells them. I find my dealership breaks more things than repairs. when the mechanic does the work things get fixed.
Bottom line, I would do it myself or wait until I couldn’t hold it on the road.
As for the printer, it was awesome just 2 months ago. However, an endstop did not function and the nozzle pushed as hard as it could into [0,0] for a good time. The motor connector got much taller (replaced), the z rods don’t exactly spin at [0,0,210] but at an orbit of that point when Z is low. It’s not extreme, but enough to easily see on a printed wall.
Dual optical sensors watching a blinking led (from firmware to mark the point as stable) and the rising Z axis (with a mirror on the nozzle) at 25 X,Y points could do amazing repair… The accuracy could get very high by using optics to exaggerate movements of the nozzle.
I.E. I could get the mechanic to build a free nitrogen tire pump…
I’ll replace the Zaxis 8mm threaded rods… I’ll add anti backlash and replace the 8mm nuts. I’ll look at keeping the 8mm rods centered at the top. That will be a start to where I was before my auto bed leveling damages occurred…
@ThantiK Does the picture above give ideas? I have a HICtop Prusa i3pd08 plastic. The left bearing “may have” lost 3 balls when I built the unit (4 months ago). The bar didn’t wobble before. How does one measure perpendicularity of the axis while traveling?
What is the angular error allowable in each axis fto achieve .06 mm for a 200x280x310 bed? (I can measure .06, straight up Z plane but that’s 100% error? What is the XY error on a .06 print?)
So if I were to embark on compensating the Z-axis via software, I should calibrate the “error introducer” (8mm rods). I.E. Here is the (height, angle of inclination “movement of points on a bottom and top point of the two horizontal rods in X axis”) measurement at each delta angle for this rod with a mean, std dev (or some other correction mechanism).
Generate good statistics for each rod. Then go figure if the data is useful, consistent and improves accuracy over the cost of obtaining it.
(Printing is OK, but calibrating is much more fun. Not all people think this way I’m starting to think…)
If they’re 8mm rods, I’d definitely set a layer height of 0.0625mm as your layer height. This gives a good matchup with a 200 step/rotation motor. You could be having layer height banding due to mismatch of layer height and full step mismatch.
I used to argue with @Whosa_whatsis about this; saying that I never got this banding but I’ve noticed that the finer layers you go, the more pronounced it becomes. I was always printing pretty thick layers. Since you’re printing so thin, it’s possible that you’re seeing a resulting error in the microstep positioning of the threaded rod.
@Mark_Rehorst I am used to working with engineers, programmers, everyone with an advanced degree in a field and have worked in that field for 30 years.
When I get “It’s a waste of time” response I always listen because an engineer with 30 years experience is telling me something. I absolutely appreciate someone wrapping up a 300 page essay into 1 word.
I am not a mechanical engineer. I know electronics and computers. It seems obvious to me that understanding the bolt to a micro level, developing a predictive model and then using that information to change
the output would work.
These printers are SLOW. In the time it takes a screw to spin 1 turn, 4,000,000,000 x 8 instructions can be performed. In my world, do that once. Study the results.Determine if it is random or not. If not random, determine if the correction is useful, repeatable, of enough precision, of far too much accuracy and proceed or not.
You are telling me that it will not work. I can see many obstacles that could cause this to not work. Many at the physical level. Many at the operator level. (Like: To get this to work I need to run this before every print. It would work, but just takes too much time and frustration.)
I absolutely appreciate your responses. I will assume you are much smarter than me, have a lot more energy, have a lot more experience. To me a direct statement, “This is this” is an alpha dog statement. (An alpha dog is the best dog declares itself the leader. All other team dogs listen and follow or else.)