Partially Solved Need better accuracy

Specifically printing with petg. I know it has very little shrinkage which is one reason why i print it. On this particular template I'm about .020" small over 5.3". They want everything within +-.010". Does that sound like a shrinkage problem and I just need to scale my models to 1.0033ish to compensate? Should I be able to hit +-.010 consistently with the robo?

 

Yeah I realize that is a lot to ask of an $800 printer. I've been shopping around but even going up to $5k I haven't been able to find anything that suits my needs. Someone pointed me towards building a Herculean (maybe you) but I'm worried I'll even have problems tuning that. Any suggestions on a more commercial machine for $5k max? I would want a large build volume if I was spending that much too.

Sent from my LGLS990 using Tapatalk

 

Depends on how fast you are pushing it.

Generally speaking, if greater accuracy is needed, print in finer layers at slower speeds so that the inertia of the X & Y axes are less of a factor (Z-axis movement is rarely an issue so much as Z-axis rigidity while the X-axis is rapidly moving).

This all assumes you've calibrated all of the axes to move properly. You can do some minor scaling to compensate for material shrinkage, but no amount of scaling will compensate for a machine that isn't properly tuned.

EDIT: My $350 Monoprice Maker Select has been consistently printing at the exact sizes I design parts to, but I print things at 40mm/s for PLA and 15-20mm/s for PETG and just let things run overnight/etc. in an enclosed area so that no errant drafts cause any material shocks.

I attribute most of my success with that printer to it being a steel frame that is well assembled though.

Point is that a more expensive printer isn't necessarily going to be more accurate unless you get something known for good QC and packaging like the Taz6.

 

I have simplify set to run 60mm/s and adjust speed while it's printing from 60%-90% i also have turned down my acceleration and jerk settings so that has helped quite a bit. I may move acceleration down even further.

Sent from my LGLS990 using Tapatalk

 

Not really sure what else I can tune. I calibrated my extruder at 105 but I could tell some of my prints weren't coming out square so i turned that down to 101-102.

Sent from my LGLS990 using Tapatalk

 

So there's probably nothing for $5k that could run that accurate? I know im pretty much trying to compete with stratasy printers with an $800 printer.

Sent from my LGLS990 using Tapatalk

 

Maybe that's why the stratasys prints I have looked at look like they print with a .3mm or even smaller nozzle.

Sent from my LGLS990 using Tapatalk

 

Keep in mind the metric vs. imperial unit differences, too.

 

What is the biggest disadvantage to an FDM machine? Belts? Plastic shrinkage? Are there any diy ball screw driven machines? I think the physical limitations of melting plastic and trying to hold tight tolerances just don't work together. Even a stratasys 900 says accuracy of .0015" per inch. So my Robo holding .020" over 5.3" really isn't that bad.

I can see what you guys are saying tho. Getting one part to run consistent is oprobably possible but being that I print tooling there isn't much I print twice.

Sent from my LGLS990 using Tapatalk

 

I have a co-worker that has a taz 5 and he called them up because he was off about. 030" at 8". They told him to just adjust his steps per mm until he was good. I know you guys had said to not adjust that but he's confident he can get his prints within. 010". But like you guys said I think he'll have problems printing different parts. I gave him 4 decent sized check fixtures they want within. 010" because I didn't want to mess with them. Think he can do it by calibrating the way he did?

Sent from my LGLS990 using Tapatalk

 

Doing something along the lines of what @TRAVIS KEMP has been posting in the Shoutbox lately is probably the only way to approach it, but even then you need to switch from a belt drive system on X & Y to a CNC mill-style ballscrew/leadscrew system and accept the resulting slowdown in machine speeds.

EDIT: Something like the TRINUS concept on Kickstarter, for example, as far as the concept/mechanics go.

 

Any idea what kind of speeds are possible with a ball screw driven machine? Something along the lines of a Herculean with ball screws would be awesome.

Sent from my LGLS990 using Tapatalk

 

Depends entirely on the quality of the parts used and the motors driving them.

Keep in mind that the truly impressive CNC machines with ballscrews on all axes cost several thousand dollars before you even mount the controller on them.