Alright, folks. With your help I’ve been able to increase my printing quality a lot. Thank you very much for that!
I hope that you will continue to be so helpful because there’s another thing that I’d like to solve. I’ve started to increase the printing speed of my Velleman K8200. I know it’s not designed to print at higher speeds. But I’d like to know its limits. Actually, it’s printing pretty much without issues at higher speeds.
Well, aside from x- and y-axis overshooting and ghosting due to ripple at sharp corners increasing with the set speed. Is there anything I can do about it beyond reducing printing speed again? Playing with acceleration values maybe? Belt tension? I know it’s caused by the far too high inertia of the print bed. But I still hope that it can be solved by tuning and tweaking some parameters (aside from speed which is to be optimized to it’s physical upper limit).
Any advice would be greatly appreciated.
Cheers,
Hendrik
@Wayne_Friedt does that help? I have always had people tell me to decrease it. But I am not certain what jerk does at the firmware level. There has been a lot of people define what it means… But not what it changes
This Info I haved gleened from the Solidoodle forum and specificaly from Ian Johnson. Upping the jerk which is some strainge word for instantainous acceleration will make squarer courners. I would just suggest, you try it so you will know first hand if it really works.
Increase it? Wouldn’t that worsen the problem? Umh… let me wonder out loud about it. Jerk is the derivation of acceleration. That is, it’s how much acceleration changes, respectively can change, over time. So, I guess, increased jerk might actually indeed improve printing quality because the printer can change directions more quickly with a higher jerk. Am I right?
But at what cost might that happen? Increased potential of skipping steps? Anyway, I’ll give it a try. Due to the fact that I’m using a RasPi as printing server for network printing (with OctoPrint) it might however be a bit difficult to change the value.
Ah, I’ve just seen that I can change it through g-code… great. That’s easy. 
Jerk is the rate of change in the rate of acceleration… http://en.wikipedia.org/wiki/Jerk_(physics)
It’s not a perfect analogy, but I like to think of acceleration as the throttle position on my car and jerk as how fast I change the throttle position. So imagine you are stopped at a signal, there is no acceleration taking place. If you want to accelerate at 50% throttle you can move to that throttle position quickly (higher jerk) or slowly (lower jerk).
In terms of motor control, jerk settings are very import in preventing vibration and stress on the structure and maintaining smoothness and accuracy of movements.
@Matthew_Satterlee i’m sure that’s the technically correct way of defining jerk, but Marlin uses its own definition. The jerk value you enter will be the speed difference that the axis will cover without any acceleration, relying on the axis’ elasticity to take up the sudden movement. Yes, it’s not how it should be done properly, but it’s pretty much the only way that our printers can print non-straight shapes nicely without having to interpret splines via gcode.
To consolidate the increased speed and reduce step skipping (which I encountered several times in a test print last night) I’d like to raise the stepper driver reference voltage a little bit more. I’m at .55 volts and would like to know where I could go, which is the upper limit which the drivers and motors still survive and do work with without any cooling (!) and without falling into overheat mode. Is .6 volts still okay? .65 volts? I don’t want to destroy anything. That’s why I ask before I try it myself.
No prob with 6.5v. I have ran them over 1v Before.
Really over 1.0v? But you’ve got heatsinks on the drivers, don’t you? IIRC my drivers were set to 0.8v at delivery. After plugging everything in and forgetting to adjust them appropriately at first hand they fell into overheat mode almost immediately. So to me 0.8v already seem too high. But with at least some heatsink I can imagine that even 1.0v might work.
Yes heatsinks and a fan. I don’t generally run that high but when i was having some trouble i had turned them up to narrow down some problems.
Now on a reprap i have is a different story seems i need to run the Z steppers over 1V or one stepper will skip steps. Maybe be cause i run those in series, not really sure why tho.
For general use i would say in the .4 is a typical value to run at.
Keep in mind that the reference voltage can mean different current settings depending on the exact driver you’re using. This is because different manufacturers use different value for the sense resistors - Pololu uses 50mOhm, Stepstick uses 200mOhm, so on a stepstick you’ll have to set the reference voltage four times as high as on a genuine Pololu to get the same current output. Some vendors also use 100mOhm, just to make things even more complicated.
Good point, that could explain my voltages.
Thanks for the correction @Thomas_Sanladerer ! I work with various industrial robots and had always seen jerk used in the manner I described.
Ah, I see… well, it’s a stock Velleman K8200, thus, I have Pololu stepper drivers. According to the reprap wiki the current is calculated by:
I = Vref / (8 * Rs)
So with .55v my current current (I’m no native English speaker, how do you say “momentary charge flow” - current current? 
Anyway…) is about 1.375A. According to the type label of the Wantai NEMA 17 stepper it can stand 2.5A (which I doubt, it’s a Chinese product after all). So using the above mentioned equation I could run the steppers with a ref voltage of 1v but would apparently need some active cooling for the drivers.
Am I right?
//edit: I forgot about the driver itself. Apparently, it can only take about 2 amps when perfectly cooled. But 2 amps would probably help to get rid of skipping steps…