LePaul
Posts: 12
Joined: Tue Jan 26, 2016 9:57 am

Build time is way off

Sun Oct 15, 2017 9:52 pm

Since 4.0 hit my build times are just incredibly off.

A 21 hour print takes 44 and I could give countless examples of others that are 20% or more off.

Any idea when this problem will be fixed? I do a lot of large prints and being off so much is a real hassle.

Ayres3D
Posts: 85
Joined: Mon Aug 14, 2017 1:15 pm

Re: Build time is way off

Wed Oct 18, 2017 9:48 am

I'd be interested in the answer to this too. I'm seeing similar effects.

swatti
Posts: 17
Joined: Tue Jun 06, 2017 2:50 pm

Re: Build time is way off

Tue Oct 24, 2017 12:51 pm

If i remember correctly "travel moves" arent counted at all. There were others too.

Printing a pair of tall sticks on each corner of the plate are pretty much the same time as if they were right next to each other.

I also print big, very big and usually just add +50% to the time if i make more then one part at a time.

mathisyourfriend
Posts: 36
Joined: Tue Mar 12, 2013 10:50 pm

Re: Build time is way off

Wed Oct 25, 2017 7:05 pm

It has always underestimated the time. Weight it gets within 2% regularly. It's an easy length of filament measurement for the software. Time is much harder due to the number of movements/lines (prints of over 500,000 lines are common). Accelerations are also difficult to account for. There's a better explanation on the Makergear forum (I asked the same question).

gwhite
Posts: 95
Joined: Wed Jul 19, 2017 9:37 am

Re: Build time is way off

Wed Oct 25, 2017 9:00 pm

I would think it would be possible to do some rough calibration based on several prints. The time will depend on a variety of factors, and giving the software finished times for some prints should allow it to come a lot closer than it currently does.

Ander
Posts: 4
Joined: Thu Aug 04, 2016 3:20 pm

Re: Build time is way off

Fri Oct 27, 2017 6:17 am

I am as well very interested in solving this Problem because wrong calculation becomes a real problem when you need to print multiple pieces on demand like for a costumer.

I use S3D on 3 different Printers and found that each has an individual percentage of extra time that is needed. When printing on the Ultimaker 2+ i add around 20-25% of time, on the P3Steel 30-40%, and for the Kodama Trinus real printing time is far of anything that S3D calculates.
Right now i am printing something on the Trinus for which S3D calculated 8.5 Hours. It has taken more than 13 hours so far and by looking at it i expect it to print for another 2 hours.

Would really appreciate someone from S3D explaining and looking into this.

gearsawe
Posts: 141
Joined: Sun Sep 10, 2017 11:06 pm

Re: Build time is way off

Sat Nov 04, 2017 11:40 pm

It should be possible. the $100K printer can get it within a few minutes on a 20 hour job. There needs needs to be a good calibration system...
Like a "time calibration wizard"
it starts a pattern such as 100mm square 10 time. S3D could then see when it starts and ends over USB. Then do a 10mm square 10 time to get a feel for the acceleration of the printer. Again it should be possible

arhi
Posts: 377
Joined: Thu Oct 06, 2016 5:13 pm

Re: Build time is way off

Sun Nov 05, 2017 12:12 am

gearsawe wrote:It should be possible. the $100K printer can get it within a few minutes on a 20 hour job. ... Again it should be possible
no need for "calibration", it's rather simple, you just need to value in the firmware configuration values for acceleration and jerk/junction deviation and that's it.

here's an example, as you can see it's few seconds off only on 2.5h print .. if I grab some free time I might finish that app completely and make an octoprint plugin out of it, would be cool if for e.g. s3d supported extensions so we can create s3d plugin that shows the real print time but..
Attachments
Capture.JPG
the real stat from octoprint and the calculated print time by gcodestat
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

gearsawe
Posts: 141
Joined: Sun Sep 10, 2017 11:06 pm

Re: Build time is way off

Sun Nov 05, 2017 5:17 pm

And are your values for XY movements the same for Z? In my case Z retraction is slightly different accelerations and jerk.
The only reason I ask for a "Wizard" options is not every printer out there has these settings or has settings that are accessible.

arhi
Posts: 377
Joined: Thu Oct 06, 2016 5:13 pm

Re: Build time is way off

Mon Nov 06, 2017 5:46 am

gearsawe wrote:And are your values for XY movements the same for Z?
this was a "quick test" made while back to see if I can get proper print time at all so I have "hardcoded" values I use in smoothieware and I only support acceleration and junction deviation (smoothieware does not use jerk like marlin for e.g.)... but it's not a biggy to update the application to support jerk too... I have same values for all three axes in this code but it's tops 30min to have different X, Y and Z (attm I have XY one variable and Z other but they all have same value hardcoded) ... also there should be support for M codes that change acceleration and jerk and junction deviation settings on the fly - I don't support those here neither... as I said it was just an exercise to see if it's complicated to get proper time and well it's not..
gearsawe wrote: In my case Z retraction is slightly different accelerations and jerk.
in most cases Z uses different jerk/accel settings the XY it's normal and expected. anyhow even if you don't account for Z moves, the total Z moves in your print take few seconds, on a few hour print those few seconds total time are irrelevant, so having precise Z move measurement is not too important, but anyhow, it's not a big issue to add separate z settings too ... if I decide to put some more hours on that project I'll do that and put it on github.. especially as recently I added "z hop" option for retraction so Z is not that insignificant any more :D
gearsawe wrote: The only reason I ask for a "Wizard" options is not every printer out there has these settings or has settings that are accessible.
[/quote]
well I was planning to allow this code to "read the config file" for smoothieware and auto setup values from there as an option... most of other ones don't have "config file" so would not work for them but one should know what accel/jerk settings one set to the printer. Also most slicers allow you to configure accel/jerk direcly and will send those as G-code at the begining or you can generate that G-code yourself... if I put some time into this I do plan to parse those mcodes :D (M201, M203, M204, M205) .. so in theory your "start g-code" should have these and the parser will work properly without any wizard :) ... anyhow it's all in the realm of theory as "few hours" for me is something I can only imagine these days as I got a second kid few days ago and time became very relative again :D :D :D
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

Return to “General Discussion and Tips”