Print Time estimate  

Re: Print Time estimate

Hello,

I noticed that the estimate print time in Prusa Control was off. Sometimes by the factor of 1.5. so it took for example 3 hours instead of 2. I mostly got the same print time estimates using http://gcode.ws/ as i got of prusa control.

I just printed the stock gcode Adalinda that came on the sd card. so no changes at all.
http://gcode.ws/ gives me 3:52:7
Prusa control gives me 4:01
file on the sd card says 5h
running the print on my mk3 is running close to 7 hours. (See picture) EDIT: It took 7:05 in total.

So my question. is there any way to make a better guess? is it possible that firmware on my mk3 has like "wrong" or old setting throwing the estimate of?
I am not realy complaining about the speed of the printer. my whole desk is shaking violently becaus it is moving so fast (Stone plate and Enclosure incoming soon 😉 )
i just want to have some sort of guess how long a print will take. Somewhat within an hour would be good enough i guess 😀

Any ideas or experience with print time estimates that might help me?

thanks everybody.
Cheers
Oliver

Usually its model dependent. A big factor in print time estimation being wrong is on longer prints than that. That's weird that its off so much on a relatively short print. 24 hour prints are usually off by an hour or 2 for me. Its usually the acceleration/deceleration of the motors that add up over time where the .gcode estimates is full speed motor movements. Are you using the 2.0 drivers? Try slicing in Slic3r and see what that gcode produces on http://gcode.ws .

For me, PrusaControl has been far closer at print time esitmates than Astroprint/Octoprint have been.

Re: Print Time estimate

It seems as though software should be able to predict it down to the second, with only two sources for error:

For this to work, though, the slicer would have to know all firmware factors (acceleration, jerk, etc) and model the actual behavior of the firmware. I’m assuming current slicers either assume infinite acceleration/jerk, or estimate some conservative value.

Re: Print Time estimate

Firmware and Slic3r are on the newest Versions.
I just did a smal part in PETG. Gcode created with Slic3r Prusa edition 1.3.7.2
Gcode.ws said 01:26
Part ran 1:46
This is not a big differenz but fits into the 130% - 150% error i have observed.

Re: Print Time estimate

Run your gcodes through Repetier Host. It is by far the most accurate Gcode time estimator I've used. Also, if you use repetier server (better solution than octoprint, imo) it will do the same calculation. It's accurate to within a few minutes a majority of the time.

Here, I ran the default SD card gcodes through repetier host, and these are the results:

MK3_PLA_3DBenchy_0.15_2H - 1h:42m:41s
MK3_PLA_3DHubs_Marvin_100um_40M - 0h:54m:16s
MK3_PLA_Adalinda_200um_5H - 7h:5m:28s
MK3_PLA_Batman_200um_15M - 0h:15m:35s
MK3_PLA_BeerOpener_ColorPrint_200um_40M - 0h:35m:13s
MK3_PLA_Buddy_150um_1H - 1h:31m:28s
MK3_PLA_Castle_100um_9H - 11h:13m:39s
MK3_PLA_Gear_Bearing_150um_2H - 2h:54m:50s
MK3_PLA_Nefertiti_150um_5.30H - 3h:40m:36s
MK3_PLA_Prusa_200um_15M - 0h:14m:49s
MK3_PLA_Treefrog_50um_3H - 4h:47m:44s
MK3_PLA_Triceratops_Skull_150um_3H - 4h:35m:32s
MK3_PLA_Vase_200um_5H - 10h:30m:18s
MK3_PLA_Whistle_200um_20M - 0h:21m:38s

Re: Print Time estimate

I found the same thing, last night I printed a little clamp part consisting of three pieces. Prusa control estimated like 1:45 but the print took about 3 hrs. Cura which I have used in the past on an i3 variant seems to estimate time fairly accurately, certainly within a few minutes, I think there is some miscalculation going on.

Re: Print Time estimate

The latest (alpha) version of Prusa Slic3r has time estimation as part of its gcode generation. It appears in the lower right where it displays filament used etc.

Apparently this Slic3r time estimate is far more accurate.