Page 1 of 1

V 4.1 Still with bad print time prediction

Posted: Wed Nov 07, 2018 7:39 am
by brunofporto
Ma latests prints:

S3D 4.1 Predicted Time: 99 mins
Duet DWC Simulation Predicted time: 125 mins
Real printing time: 128 mins
S3D 4.1 Error: 29,29%
DWC Error: 2,4%


S3D 4.1 Predicted Time: 183 mins
Duet DWC Simulation Predicted time: 254 mins
Real printing time: 257 mins
S3D 4.1 Error: 40,44%
DWC Error: 1,18%

Also GCode Analyser website (http://gcodeanalyser.com/ is always below 5% of error...

Re: V 4.1 Still with bad print time prediction

Posted: Wed Nov 07, 2018 9:00 am
by Raise3D Pro Plus 2
Interesting. I have also problem with overall print time with Idea Maker software. This problem is common.

Re: V 4.1 Still with bad print time prediction

Posted: Wed Nov 07, 2018 9:42 am
by andrewk72
I think it would help if you also attached the gcode files that you printed for those 2 test cases. That would help when comparing between versions.

What did 4.0.1 estimate?

Re: V 4.1 Still with bad print time prediction

Posted: Wed Nov 07, 2018 11:02 am
by brunofporto
Here is a graphic of my latest 60 prints comparing real time with Duet Web Control / Gcode Analyzer prediction with S3D 4.0.1.
chart.png
Maybe it is very precise with the default printers you can select but not for generic RepRap printers like mine. 4.0.1 times looks the same. Would be nice if we could add our own printers data like acceleration, jerk, retraction (I use firmware retraction...) like we do with Gcode Analyzer website and this will lead to good predictions.

Re: V 4.1 Still with bad print time prediction

Posted: Wed Nov 07, 2018 11:47 am
by Raise3D Pro Plus 2
Thank you mate

Re: V 4.1 Still with bad print time prediction

Posted: Wed Nov 07, 2018 9:46 pm
by tonyno
I print a lot of small parts and noticed this discrepancy with the new algorithm.

S3D 4.0.1 Build Time: 48 minutes
S3D 4.1.0 Build Time: 46 minutes
Actual Time: 35 minutes

I have an xyzprinting da Vinci 1.0A.

Re: V 4.1 Still with bad print time prediction

Posted: Thu Nov 08, 2018 4:48 am
by andrewk72
No one is going to be able to help if you just post a random time. If you want to give them more actual data that they can use, you probably need to include at a minimum:
  • What printer you are using
  • The time estimate from V4.1.0
  • The actual print time
  • ATTACH THE GCODE FILE

Just posting that X print took Y hours isn't going to help much without this other information

Re: V 4.1 Still with bad print time prediction

Posted: Thu Nov 08, 2018 7:26 am
by brunofporto
andrewk72 wrote:
Thu Nov 08, 2018 4:48 am
No one is going to be able to help if you just post a random time. If you want to give them more actual data that they can use, you probably need to include at a minimum:
  • What printer you are using
  • The time estimate from V4.1.0
  • The actual print time
  • ATTACH THE GCODE FILE

Just posting that X print took Y hours isn't going to help much without this other information
I do work with classified files per NDA contracts. Also, an example is not needed when it is consistently bad. Sum with the fact that I am not allowed to publish any gcode generated by S3D due to their own therms.

Re: V 4.1 Still with bad print time prediction

Posted: Thu Nov 08, 2018 7:31 am
by brunofporto
tonyno wrote:
Wed Nov 07, 2018 9:46 pm
I print a lot of small parts and noticed this discrepancy with the new algorithm.

S3D 4.0.1 Build Time: 48 minutes
S3D 4.1.0 Build Time: 46 minutes
Actual Time: 35 minutes

I have an xyzprinting da Vinci 1.0A.
Short time parts are always the worsts even with the more precise prediction tools like GCode analyzer and DWC simulation. At least you are getting the real part faster :D

Re: V 4.1 Still with bad print time prediction

Posted: Sat Nov 10, 2018 10:31 am
by tonyno
At least it's consistent! :D This is a different file from before (attached this time).

ET 4.1.0: 46 mins
ET 4.0.1: 47 mins
Actual Time: 35 mins
Da Vinci 1.0A