What causes a y offset to occur in the middle of a print?

[akkrolnik]

Radio Shack (where they are still around) has a digital caliper for under $30 US. Another source, if you have one in your area, is a Harbor Freight store. I have the RS caliper and it works great and seems accurate. I got a digital micrometer at HF specifically for measuring filament, and it was probably about $25, and it's certainly good enough. Both are selectable for inches or mm.

Dale

Dale, I got a nice caliper from Amazon for around $30. I was able to accurately measure the wall thickness of a calibration cube and adjust my extrusion multiplier. So I am set there.

Regarding my original issue with skipping in the y direction, I think the combination of a large part, fast printing, and high default acceleration makes a lot of sense. Having never worked with stepper motors, I hadn't realized that there was an accuracy and torque trade-off. My conclusion is that the y-offset was caused by a high torque which caused the motor to skip a couple steps. I always thought that the ideal set-up would be a bedplate that only moves up and down. I figured the accuracy would be higher due to less potential for swaying of the part. I hadn't considered this possibility. Once again, thanks to all those who replied. I now understand my printer better.

[Slipshine]

I am just now getting back to this. I believe my settings are a bit high.

ACCEL.JPG (48.76 KiB) Viewed 10754 times

If I am reading this correctly the xy acceleration can be as high as 9000
but the m204 sets it to 3000. And since the z is set to 30 max it will still be 30

[jimc]

send M204 S1000 to the printer. this will change the acceleration. if you like it at 1000 then send M500 to make it permanent or you can put that m204 command in the start script of your process so you can change it with your print.

[Slipshine]

Thanks JIm I tried that Last night. worked ok but it was a small print.

Ill print one I have been having problems with. Just putting off going back to 1/8 steps. I always get edgy updating firmware.

[jimc]

if your using s3d the xy movement speed is set to something crazy like 18000. bring that down some. i have mine at 9000.

[Slipshine]

I had already slowed it to 12000 after the first shift. Only time will tell. I will be out of town for about a week so I wont be able to do any printing.

But if the slower accel doesn't work I will go back to 1/8 steps. I never had a shifted print prior to the switch to 1/16 . And I had done some stupid stuff.
I have been trying to find a formula for the difference in torque between the two. I may just borrow a digital torque gauge from work and test it.

But seriously thanks for all the ideas.

[Tim]

This is a thing, more steps is lower torque, but it's also smoother motion. You can change the microstepping in the firmware if you really want to, it's in configuration_adv.h

You can change the microstepping in configuration_adv.h, but you won't get back 1/8 microstepping. Due to the way the stepper driver chip (the Allegro A4982, which replaces the A4984 on older RAMBos) is designed, the Allegro A4982 is actually missing a 1/8 step mode. It's modes are 1, 1/2, 1/4, and 1/16 stepping. This is done for backwards compatibility due to the fact that that the step selection is a 2-pin input, so only four modes of operation can be supported by the chip. It seems really awkward, to me, but that's the way they designed it. And it means that if you have the newer RAMBo, you lose 1/8 stepping completely.

[insta]

This is a thing, more steps is lower torque, but it's also smoother motion. You can change the microstepping in the firmware if you really want to, it's in configuration_adv.h

You can change the microstepping in configuration_adv.h, but you won't get back 1/8 microstepping. Due to the way the stepper driver chip (the Allegro A4982, which replaces the A4984 on older RAMBos) is designed, the Allegro A4982 is actually missing a 1/8 step mode. It's modes are 1, 1/2, 1/4, and 1/16 stepping. This is done for backwards compatibility due to the fact that that the step selection is a 2-pin input, so only four modes of operation can be supported by the chip. It seems really awkward, to me, but that's the way they designed it. And it means that if you have the newer RAMBo, you lose 1/8 stepping completely.

Well that's the stupidest of stupid things.

[ednisley]

the stupidest of stupid things.

It probably doesn't matter:
http://softsolder.com/2011/09/11/steppe ... vs-torque/



Each dot shows the pull-out torque at a given speed, with the same color for various microstep modes and the dot-and-caret for full-step mode. There's little difference between the points in each color group; any defensible error bars would swallow up most of the variation.

For printing speeds around 50 mm/s = 3000 mm/min, the motor turns at 83 rpm in the dot cloud in the upper left corner. Any microstepping mode produces more torque than full stepping at those speeds, but there's not much difference between 1/2 and 1/16 mode.

The data came from a homebrew dyamometer:
http://softsolder.com/2011/08/25/steppe ... mechanics/



The methodology definitely left something to be desired... [grin]

[jsc]

1, 2, 3, Fight!