User avatar
Posts: 9
Joined: Mon Dec 15, 2014 8:37 am
Location: Belgia
Contact: Website

Re: Print times fix

As already mentioned colleague print time can not be calculated exactly, why?
Simplify 3D program would have to have the support of the firmware and retrieve data such as default acceleration, maximum acceleration, maximum FEEDRATE and XYZJERK for each axis
These are the main parameters of the travel speed and acceleration.
For machines with low speed limits of error may be small 10/15%, but with bigger machines is a different matter. My machine normally prints at speeds of 250 to 300 mm / sec but only for large parts because I have set acceleration so that the printer does not jump me across the floor making small prints.
For example, the range of error when printing large parts margin of error is also about 10% but at small with lots of changes of direction and retraction even 115% of the time error print.
To program was able to correctly calculate the time printout must have options for downloading parameters from the motherboard, but it only works if the printer is connected. There is one option better or manually enter dese parameters such as speed, type Z, XY and extrusion, just ask the printer manufacturer, or delivered in the printer parameter file and uploaded via the Configuration Assistant. But whether this feature will once one knows, but I hope for it.
Posts: 43
Joined: Thu Apr 24, 2014 11:30 pm

Print time estimate adjustment

My print time estimate is always off and not by a small amount. If I see an estimate of 3 hours I know it's 4:30 to 5 hours.

Slic3r had a simple % adjustment for the time estimater and after a few adjustments and test prints you could get pretty close. A 10 minute estimate might be 1 minute off and a 5 hour estimate might be 10 minutes off instead of hours off.

Basically the adjustment would be setup like extrusion multiplier where "1" = 100% and ".9" = 90% and "1.5" = 150% and so on.

I would think this would be fairly easy to do.

Posts: 10
Joined: Sat Jun 13, 2015 8:43 pm

Re: Print time estimate adjustment

Yes, a way to more accurately estimate the print time would be nice. If you're going to implement this you might as well go ahead and build in some intelligence. Something like a running history of the previous ten successfully completed builds. Take the actual print time divided by estimated print time for each of the last ten builds and average those ratios together to come up with the print time modifier. That should be a pretty robust method, it would be even more robust if you used the median value rather than the mean.
Posts: 159
Joined: Tue Sep 30, 2014 5:37 pm

Re: Print time estimate adjustment

The error isn't likely to be linear though. In my case it's pretty accurate up to about an hour, but beyond that the acceleration and jerk settings of the printer come into play, and the complexity of the print path can make the estimate high or low. For example, see this graph of 91 prints I've done. The X axis is the estimate from S3D and the Y axis is the actual print time. The values are all minutes:
Screen Shot 2016-01-10 at 5.09.49 PM.png
Having a single adjustment factor might get you closer, but it still won't be accurate except for small/fast prints.
Posts: 43
Joined: Thu Apr 24, 2014 11:30 pm

Re: Print time estimate adjustment

Having a single adjustment factor might get you closer, but it still won't be accurate except for small/fast prints.
I would gladly accept "closer". Like I said, Slic3r has it and it gets much closer than SF3d at least for me.

It could actually learn a little. Say the estimate is 3 hours and it knows it actually took 4:30 then maybe it could learn what % it's off by and make minor adjustments each time until it's average is +/- 10%. It could build up a curve as well making short and long prints closer by a lot.'

For now though I would be happy with a simple % that I could change.
Posts: 2
Joined: Mon Feb 08, 2016 11:23 am

Re: Print time estimate adjustment

Something else I have noticed is it seems that pauses added through the Scripts tab are ignored in estimation. I have input scripts to pause 10+ seconds between layers, checked the estimation, then removed the script and the estimation has remained the same.
Posts: 7
Joined: Thu Aug 17, 2017 3:54 am

Better Build Time calculation

Hello again!

I've already talked with support about this feature. This software has serious troubles to give us an actual "build time" after generating the g-code. I've seen several prints that expends much more than twice the time calculated by this software. This issue is a big one for any professional 3D printing service like ours.

Your support team point me the G-Code Analyzer website so I can get better build time information, but it force us to waste more time per each g-code as we need to check it everytime. The last versions of Cura Software already have the acceleration and jerk settings in there so its timing calculations are much more accurate than Simplify3D.

Please consider seriously to add this feature along with the "adaptative infill" option. Those two features from Cura are turning a lot of people to abandon your software and migrate to this one. It is a pitty because your software is awesome, but it seems insane that a licensed software like this lack several features that another open software already has.

Posts: 16
Joined: Wed Oct 04, 2017 11:45 am

Re: Better Build Time calculation

Sorry to interfere as a normal user with a feature request, however, what you're requesting is near impossible to code. The problem with this is that every kind of controller board has a different algorithm to cope with things like acceleration and jerk, amongst other. They _could_ estimate slightly better by incorporating this, but still it would be an estimate and it would still be possible, for certain kinds of prints, to get estimates which are way off.

The problem is that the acceleration/jerk settings are used on EVERY line in a print. Therefor, if a calculation is only 5% off, cumulative this will constitute a major estimation-fault.
Posts: 463
Joined: Thu Oct 06, 2016 5:13 pm

Re: Better Build Time calculation

FlyingDutch wrote:what you're requesting is near impossible to code. The problem with this is that every kind of controller board has a different algorithm to cope with things like acceleration and jerk
"amongst other" - the only other thing is time for heaters to arrive to set temps

"impossible" no, not even hard

"every kind of controller..." - actually there's less then 5 firmware "types" supported by s3d and they use 3 different type of accel/jerk settings
- nothing (most old firmwares)
- accel/jerk (marlin and co)
- junction deviation (smoothie and co)

and even just handling marlin accel/jerk solves over 90% of the market and gets junction dev. ones to "close" values .. just like any of the websites that calculate this for you (for free, on-line) do..
gcodestat integrates with Simplify3D and allow you to
Calculate print time accurately (acceleration, max speed, junction deviation all taken into consideration)
Embed M117 codes into G-Code
Upload your G-Code directly to Octoprint
open source and unlicence
Posts: 6
Joined: Wed Nov 15, 2017 9:01 am

Re: Better Build Time calculation

This is very important. I am working with a large format printer (1mx1mx1m) and the difference between knowing a 50hr build will be more like a 70hr build is very important. This affects scheduling of personnel for finishing, print delivery, and scheduling the next job. It is not acceptable that free software has these features when Simplify does not. I am on the verge of moving to Cura for this and the adaptive infill settings.

Return to “Feature Requests”