This isn't too uncommon. There's probably 2 big factors here that we really can't predict. The first is that many times the firmware actually steps in and limits the printing speed based on acceleration or max speed characteristics. For example, if you make a move at 8000 mm/min, but you have an extremely low max acceleration specified in your firmware, the average speed for that movement will be much less. The second big factor is all of the communication overhead. Sending millions of ASCII commands back and forth across a serial buffer has some significant latency. The SD card has about half the latency, but it's definitely still there. It takes time for the microcontrollers to process these commands and that adds up over longer and longer builds.
If things are really far off, I would probably suspect speed limiting by the firmware since that can have a much larger effect
You dont want to fix it. If your acceleration is set lower on your machine then prints will take longer. Higher acceleration means the printer will shake back and forth more causing lower quality prints. What i think s3d should have is an offset percentage in the prefrences so the user can tune the time estimate to come close to matching your printer
On a new machine I got, I found I get overestimates instead of underestimates. Last print said 35 hours and it only took 21. That's pretty far off.. but I was happy about it
Does S3D take any acceleration into the calculation? Cura, for example, does. It just has hard-coded numbers for it. If you find them in the code, you can tweak them according to your machine.
Maybe there could be an advanced setup dialog where you could enter the acceleration of the printer. But maybe this would lead to more misunderstandings (why don't have these numbers any effect on the shaking of my printer?) than benefit.
dennisjm wrote:On a new machine I got, I found I get overestimates instead of underestimates. Last print said 35 hours and it only took 21. That's pretty far off.. but I was happy about it
i found this to be the case with makerware slicer after i started using sailfish, overestimates are better than underestimates for those of us who quote based on print times
I am finding this as well. It consistently underestimates by as much as 25%. For example Simplify3D shows the estimated time at 2.51 and it took 3.30 to complete.