Partially Solved Need better accuracy

@Nathanfish So you buddy with the TAZ 4 now has it dialed in for an 8" part, what happens if he needs to print a 10" part or a 4" part? What if he changes vendors of filaments, or goes from PET-G to PET-T? Changing STEPS or any other parameter for that matter will only get him accurate parts for that single run, if he were to print is a second time with anything changed I am sure he will need to calibrate the steps again. That was backward thinking on the part of TAZ Tech Support. Basically you are building in errors into your printer setup to account for unknown material properties. It is stupid and I have said so many times on this forum. Unfortunately, it is the way it was done from the very beginning and we humans hate change and need to be dragged kicking and screaming to better ways of doing things.

 

Just a note here, building in errors is not always a bad thing. If you know precisely how you raw materials work, building in errors can result in dimensional accurate final parts. Back to LEGO for a moment. They build in precise errors into their molds since they use their own formula for ABS and know it's properties completely. That way they can keep their final dimensions precise. We are not Chemical Engineers, nor are we creating the filaments we use so there is simply no we we can be 100% accurate with our settings over a range of parts.

 

As I mentioned, a stupid way to do things

And what if he changes layer size or number of perimeters? You see what a kludge this is?

But as they say, "even a blind squirrel finds a nut every once in a while."

 

Even what @TRAVIS KEMP has done won't necessarily get you any closer. Too much moving mass. You need accuracy and repeatability, you need to minimize the moving mass and/or use much more precise and stronger stepper motors that can overcome 100% of the inertia applied by the moving mass. Besides this he did nothing to change the belts to anything more accurate.

A question: Does the Trinus use lead screws or ball screws? I suspect at their price point, provided they can actually deliver, it is lead screws.

 

@Nathanfish This is one of the points I was trying to make. Change something simple like the supplier and you end up with different final measurements. Even though we want to think it's all the same, PETG formulas differ from one supplier to another, and hence will print differently (PETG ≠ PETG ≠ PETG) .

Since extrusion amounts are calculated from other parameters if you change any of the extrusion parameters it can have an effect on other dimensions, likely minute, but still measurable.

 

Then a move to ball screws on the Trinus will also help.

With Travis' current setup, I am sure he can get very good accuracy at slow print speeds.

One other consideration most people forget is that microstepping is based on current not actual stepper steps, so theoretically a 400 step stepper with 8 microsteps should be more accurate than a 200 step stepper with 16 microsteps. So you could conclude that 400 steps + 16 microsteps would be even better with regard to part accuracy. But we are nearing the limits of compute power with all these Arduino 2560 based printer boards and there needs to be a wholesale move to 32-bit boards to make this a real solution.