Print time accuracy

When printing a solid cylinder, estimated time is around 15% under actual time.
Same cylinder but hollow, estimated time is spot on.
On more complex solid prints it takes 25% more time than estimated.

Before this the printhead would pause for hundreds of milliseconds multiple times per layer.
"Fixed" this by setting retraction at .1 which "seemed" to stop the pauses.

The printer is a SDoodle2 refit with mega2560, ramps 1.4, marlin, repetierhost, slic3r, win7 printing from SD card.
Thanks

Most print time estimates do not take acceleration or deceleration times into account, but simply accumulate speed/distance for all moves so they under-estimate.

Reducing retraction may result in blobbing & stringing. You could instead tick "Only retract when crossing perimeters (under "Print Settings" > "Infill")

Dave

Thanks Dave
Yep, blobbing and stringing is correct.
The options selected seem not to matter, longer retraction setting = longer pauses and fugly seams.
The pausing may not be a slic3r issue, possibly with marlin configuration.h.

The same part sliced with pronterface (in 2012) printed in 2h9m, slic3r estimates 1h59m but prints at 2h30m.
Using ReplicatorG on knockoff ctc also estimates and prints the part in around 2 hours.

Main issue is the extra time printing. Hope the cracks here can bust it.
Been using slic3r with repetierhost for years, I heart slic3r.
Stuart

Edited 1 time(s). Last edit at 11/10/2016 12:08PM by instmkr.

Where are you getting the estimate? I don't recall Slic3r having that functionality and the estimates Repetier Host gives me for my Rostock Max when I load the gcode in the GUI (just to get the estimate) have been very good. I agree, it's probably in jerk and acceleration (which are bigger factors when you have infill, and especially complex infill). If your machine will maintain rigidity and stability and not skip steps at higher jerk and acceleration values, doing that would make things closer.

If you absolutely must have a perfect time estimate, the Duet board and its RepRapFirmware has a simulation mode that runs the print "dry" without heaters or motors but with the entire motion planner logic going. It's not fast (on version 1.15e, my 0.8.5 seems to take about 1/6 the print time to simulate parts with my usual settings), but it's on the money.

You're not going to be able to get rid of retraction, but you can decrease the number of retractions with the option previously mentioned or longer minimum travel distances for retraction. The avoid crossing perimeters option would reduce the penalty for bad retraction settings but the sub-optimal path generation would probably take more time than just doing proper retractions. You might be able to tune in a Pressure Advance setting that minimizes the required retraction length.

So the time estimates must be coming from rHost after slic3r generates code.

Will study more about acceleration and jerk settings. These strings and blobs really bite.

How does one go about getting reasonably priced and fair professional help in these reprap build matters?
Just seems like my rookie mistakes could be detected and fixed in 10 minutes of dedicated support.
Thanks all, Stuart

There are a lot of people here who will usually try to help if they can.

If you have a local makerspace, a membership would get you access to a wide variety of equipment and possibly some tutoring.

Genuinely professional support that knows your machine is a decision you locked in when you bought from a vendor that offers it or one that doesn't. There's a lot of variation in the market. The principles of operation are similar, but people who haven't sold and supported vast numbers of your specific kit may not know its quirks.

Yep, kind peeps in the various reprap forums have solved almost every question I've posted, dumb or otherwise.
The pausing issue is still there but I let that one go.

Great tip, The Forge GSO, NC is loaded with machining and woodworking equipment, and a couple of printers.
Plan to go check it out, thanks again.