RE: Gantry resonant mode and VFA artifacts
But today there are good optical systems (think computer mouse). Additionally, an accelerometer on the head, plus some sensor fusion algorithm.
Good point; measuring with the required resolution should certainly be achievable. But for 3D printing, would you really want a positioning system where you accept random slippage of the drive and compensate it via closed-loop control? Not all components in the system respond quickly, especially the filament flow will lag behind somewhat when the XY speed fluctuates, I'd assume.
It's clearly technologically more advanced but I wouldn't say it can't be done and perform better than conventional open-loop steppers.
After all, the loop needs to correct only for belt slippage, otherwise it's the same as a conventional system. The high-frequency information comes from the accelerometer, is combined with low-frequency / static information from the optical sensor. Algorithms in that area are well understood e.g. used in GPS and friends, "Wiener filter" comes to mind.
Or put a 2nd actuator level e.g. solenoids or piezo on the print head to act directly on the accelerometer feedback minus known intended movement. Looks simple on paper 🙂
RE: Gantry resonant mode and VFA artifacts
Do you think the "waves" on the measured vibration match the VFA? Probably hard to say..
But on the Z axis they do seem to be torsional since they are very aligned on the alternating pattern. Even o X they alternate a bit.
Could it have to do with the bearings on the rails? I don't think so, since they should be smooth unless there were defective balls in it.
PS: Please don't take this in the wrong way, but it would be nice to keep this thread on topic! It is nice to read on the ideas of an alternative to the CoreXY system and on software mitigation strategies, but maybe in a separate thread 😀
Lets not derail it from the great work that @huggypanda is doing!
RE: Gantry resonant mode and VFA artifacts
You're absolutely right, I didn't mean to belittle his work. He does a great job. @huggypanda: At what print speed was the graph measured?
RE: Gantry resonant mode and VFA artifacts
It's only good work if it fixes the issue and so far I have not. 🙂 Already had one false start and no good answers. I wish it was as simple as belt tension for my setup but it just hasn't fixed it or really even improved it.
The print speed was 90mm/s and horizontal motion (left to right). That's just my most problematic speed. I'm sure everyone's a bit different.
You're absolutely right, I didn't mean to belittle his work. He does a great job. @huggypanda: At what print speed was the graph measured?
RE: Gantry resonant mode and VFA artifacts
When I look at those pictures on page 1, the VFAs seem to change amplitude suddenly at discrete heights. Most VFAs on my printer run the entire height of the part, save for about 3 mm near the build plate that's glossier. Does that vertical spacing correspond to anything?
RE: Gantry resonant mode and VFA artifacts
As for controlling the movement and position of the print head, just look at how it is solved in ordinary computer printers. Old types of dot matrix printers were actually controlled by stepper motors in both directions (print head position (X axis) and paper movement (Y axis)) without feedback. Modern printers (inkjet) are controlled by ordinary DC motors. In the path of movement of the print head (X axis) you will find a transparent strip with a whole series of dark lines and on the head carriage there is an optical sensor that reads these lines and its signal forms a feedback loop for controlling the movement motor and thus the position of the carriage. It is similar with controlling the paper position (Y axis) only it is arranged in a circle. Why this arrangement is chosen is absolutely clear: Far greater position accuracy in both axes even at the cost of greater circuit complexity. But since these printers are produced in millions of series, there is no problem integrating this control into a specialized control processor. This accuracy is unattainable with stepper motors. It would certainly be interesting to try to use this principle of controlling the position of the print head in 3D printing, but the question is whether it would be worthwhile at all.