Filament diameter compensation

koreandi

I was wondering if there are any new thought or ideas about hardware and software to compensate the flowrate dynamic by measruing the filament.

There is a reason why I pull out this old topic.

I built a laborytory printer based on a DUET-board.
The material I print is non-commercial and metal-filled.

The parts need to be 100% filled to get a 100% dense part after debinding and sintering.
As the material I print is still in developement there are only small batches done and the diameter is quite inconsistant. And though the tolerances are mostly stable nowadays... On filaments with Ø 1,75 mm and a deviation of 0,01 mm the cross-section of the filament will vary 1%.

I hope to get some good advices or even some links or documentation to any built up setup.

Best wishes from Germany
koreandi

Phaedrux

Something like this?

https://www.youtube.com/watch?v=RYgdLPe_T0c&ab_channel=ThomasSanladerer

Though I'd say ovality is as bigger problem as I've never seen a filament that is truly round. So your sensor would need to measure in at least 3 orientations.

T3P3Tony

@koreandi I see you have asked for this before:
https://forum.duet3d.com/topic/9481/filament-diameter-tolerance-compensation/11?_=1678120319096

You may be interested in this implementation:
https://qualup.com/en/filamentcontrol.php#Home

koreandi

@T3P3Tony
Yes, now I remember this again.
This question was at the beginning for this pfoject and now the project changed from scratch to a finished printer.
And I have to commmit that I didn´t react to this question:

What is the preferred communication method between the filament width measuring system and the Duet: async serial, or something else?

I couldn´t tell tell answer back then and even not today.

I know this device from Qualup from earlier days and I am missing some Information about the implementation with DUET/RFF. I will contact them and ask for any succesfull projects.

CNCModeller

@T3P3Tony said in Filament diameter compensation:

You may be interested in this implementation:
https://qualup.com/en/filamentcontrol.php#Home

This is an interesting concept, so could I stream gcode updates of the filament dimensions to a duet over the panel due interface from an Arduino like the CNC pendant? I guess I'd probably need to read filament length too.

If so I feel a homebrew version coming on.

This is something that I've wanted to try for ages to make using lower quality (and hence cheeper) filament an option for large prints.

Very cool

Cheers
Barry M

jens55

@CNCModeller said in Filament diameter compensation:

This is something that I've wanted to try for ages to make using lower quality (and hence cheeper) filament an option for large prints.

These days you have to really search for filament that is outside the +- 0.05mm tolerance range and if by chance you do find it, I would think the material itself would be suspect. If the manufacturer is saving on the QC for diameter control he is most certainly not bothering with QC for the (cheapest) ingredients he can find.
Somebody is in for a world of hurt !

CNCModeller

@jens55 what you say may be true, however when creating large prints with 5kg industrial spools I'm still finding variations of thin wall layer widths resulting in surface layer lines and if a solution like this helps then it's worth giving it a go!

And if nothing else it'll be another interesting project to investigate.

jens55

@CNCModeller said in Filament diameter compensation:

@jens55 what you say may be true, however when creating large prints with 5kg industrial spools I'm still finding variations of thin wall layer widths resulting in surface layer lines and if a solution like this helps then it's worth giving it a go!

And if nothing else it'll be another interesting project to investigate.

That is a LOT of filament for one print .... have you thought of going to pellets instead? You talked about sintering and I have no clue if your material is available in pellet form but it's a lot cheaper to buy pellets in bulk and you don't even have to compromise on quality. There are several examples on youtube of people using pellets, especially for larger prints.
While I don't have a link for you, there is one chap who air-feeds his pellets to a small hopper on the extruder and that seems like a slick system.
On a related note, it seems that all the self built very large 3D printers have crappy surface finish. I am sure the size of the nozzle is part of that equation but overall system flex must get madening when you get to 100010001000 and larger.

CNCModeller

@jens55 I've looked into pellets and came to the conclusion that it must be really difficult to get tight control on extrusion rates so I've stuck with filament. In terms of rigidity, if you build in lots of triangulation it's not too bad. Resonance can be an issue. The key thing is to try to mitigate all of the sources of error and filament diameter is just one of those.

Shoki

Hi everyone,

Just wanted to let the Duet team knows that this is also something of interest to us.

But we also are at an early stage of our work regarding this matter.