All kinds of calibrations

[Titus 170]

Dear Community,

I've been printing for a while know and things are going pretty well. Quite some nice things have come out of my UMO. That being said, I have the feeling I'm still working sub optimal. I can't really explain why, but I have the feeling things should be printed faster or neater. From the occasional first layer problems to over/under extrusion and various other problems. Being someone who likes to tweak and update things, both hardware and software, I thought it would be nice to compile a list of calibrate-able things and optimize my printer, compiling each step into a possible list/guide for other people.

Background info:

I usually print with 0.2 or 0.1 layer heights, 0.8 shell and 1.2 top/bot(to ensure it's really closed well) and 15-20% infill. The temperature is mainly 210C and for Doodle3D prints 230C

Speed vary from 50-100 mm/s.

Travel speed is 200 mm/s, though I could reach more if I oil every week.

Some things I was thinking about:

1) Zsteps: http://www.extrudable.me/2013/04/02/the-myth-of-z-speed/

2) Esteps: http://umforum.ultimaker.com/index.php?/topic/1647-extruder-calibration/

3) Temperature: http://umforum.ultimaker.com/index.php?/topic/6774-pid-autotune-kpkikd/

I should be able to do these things since there is enough/some documentation. Though I'll try to compile all the useful stuff in 1 document. Note that these are settings related tweaks, I already did a hardware tweak a while ago with aligning everything etc.

Remaining questions

-Is there also a way to calibrate X and Y?

-Are there other settings such as jerk that are calibrate-able?

Other things I want to improve but not sure what to get/print:

-More silent fans(while still efficient) (or duals)

-A different fan shroud(/dual fan shrouds)

-Get my dampers installed without having the backlash issues I once had, since I want my machine to be more silent.

-Start experimenting with larger nozzle diameters than 0.4mm

What do you think of this? Any tips, things I missed?

 

[gr5 2,172]

I would skip zsteps, esteps, and temperature. I would leave X and Y alone. I don't think there is anything to gain in those areas.

However for UMO you can make the Z axis much faster by increasing acceleration. I doubled mine (long forgot from what to what) and this greatlyl reduces any Z scaring because switching layers is more of a click than a move - very fast.

On the UM2 I think the Extruder can probably go much faster. Someone should play with extruder acceleration and max velocity. I have a feeling it can go maybe 10X faster. If retractions were more of a click than a zip, that would not just speed up prints but it should improve quality also.

 

[mnis 11]

I think ...

Many problems with the machine are very much dependent on the quality of the filament current used. Color additives and the overall composition have a very large influence on the resulting for property quality and even on the adhesion behavior. There are colors such as magenta, or yellow, which did not even want to adhere properly to a special print plate. Because, unfortunately, small tweaks are almost invariably required. I print mostly with inexpensive PLA, which regularly is cheaper by almost half compared to Ultimaker filament. New materials I test with identical settings, characterized itself quality differences of the various filaments can be identified very quickly.

A major problem also exists in the material feeding in a Bowden extruder system with only one drive wheel. Even apparently well-modified material feeder, fundamentally based on the ordinary principle with a single drive wheel, can often not keep the material flow over a longer period consistent enough, since the most accurate calculations can unfortunately do anything about.

Observe once the material transport to the naked eye, at any open material feeder, such as that of iRobert. Then it looks at the current machine maybe normal, but on closer inspection of various filament pieces that have been run by the same machine, a problem can be clearly seen. Unfortunately, my not yet published BEMF-Universal has this unicycle problem:

The distances of the bite-marks of the knurled wheel in the filaments are often differing from one another, and sometimes even sanding marks to heavy abrasion can be seen. A very small grip-loss can at the final object already left clearly visible traces. It helps a bit, if the material feeder allows a wide adjustment range for the force of the counter pressure roller, but then high print speeds are almost unthinkable.

Here two materials from a single supplier, printed with identical machine Settings:

Markus

 

[Titus 170]

@ gr5, I think I did indeed meant Zacceleration instead of steps.

If not these area's, where do you think improvements can be achieved?

 

[Titus 170]

So I did the PID autotune today:

First swing with original PID was 9.something up and 5 down.

< echo:PID settings:

< echo: M301 P22.20 I1.08 D114.00

It settled on 3.4 up and down with the following PID

M301 P13.77 I0.43 D111.21

 

[Titus 170]

@ Gr5,

You said you doubled you Zacceleration? Could you have a look what is in your settings?

Because the article says 1500 where it was 100 originally

It also talks about Zspeed, which was set to 10 originally and changed to 100.

In my printer it was set to 5 originally...