Bouncing Estimations in Print Progress
-
Can't explain these non-linear print time estimations:
The pattern is consistent: either the value for filament usage or the estimation of file progress is way too high - the other one is too low. As the screenshots show, that interplay happened 4 times in just two minutes. At least, this behaviour is consistent with all print jobs, and the values in DWC are identical to those on my PanelDue.Hardware: Duet 2 WiFi, firmware: 3.4b7
-
Is there anything notable about the print moves in those time periods?
-
@phaedrux No. All 1250 layers à 0.2 mm are identical, each takes about 1m15s, nothing special at all. With 2.05, that worked fine. Then I made the jump recently to 3.4b7, so I can't tell when the calculations might have been changed. However, I'm missing the incremental layer count in the state line of DWC, but I don't know if that is related to the phenomenon.
-
@infiniteloop I've seen similar and remarked upon it on numerous occasions. It just seems to me that DWC takes a small snapshot in time of recent moves, rather than a wider perspective of the actual print progress. None of the estimations make any sense and I've given up on them completely.
-
@deckingman said in Bouncing Estimations in Print Progress:
It just seems to me that DWC takes a small snapshot in time of recent moves
Maybe, but with my special print object these snapshots should be very precise, as all layers take the same time to print - which makes manual calculation easy: <time_per_layer> * <remaining_layer_count>. However, with 2.05, the estimations of the firmware were better in a sense of "more linear".
-
@infiniteloop I have a feeling that it all went haywire about the time that they ditched the estimation based on layer time. I was in the minority who found that one was marginally more useful than the other estimations, which are about as much use as an ashtray on a motor bike.
-
undefined Phaedrux moved this topic from Using Duet Controllers
-
@deckingman said in Bouncing Estimations in Print Progress:
which are about as much use as an ashtray on a motor bike.
Beautiful.@chrishamm would have to comment on the current estimates.
-
@phaedrux said in Bouncing Estimations in Print Progress:
@chrishamm would have to comment on the current estimates.
Hm, really? The estimation values on the Paneldue are identical, so I suppose they originate from the firmware. Or am I wrong (as so often
)?
