Sampling Filament Compression Force

This forum is for discussions specific to the Duet-equipped Rostock MAX v3.2 3D printer.
Post Reply
brianmichalk
Printmaster!
Posts: 39
Joined: Fri Aug 12, 2016 3:20 pm

Sampling Filament Compression Force

Post by brianmichalk » Mon Mar 11, 2019 3:03 pm

Problem:
My Cura projects have different settings based on the profile of the part I'm printing. If I turn on the fan at a low height, the plate cannot keep up with the loss of heat, and there is a much greater risk of part delaminating from the bed. So, for the lower heights, I restrict fan operations. I also print at 250C with a .7mm nozzle to keep print times down. The part fans are needed to keep the part cool for this, so I also use minimum layer times. If I'm manually monitoring, I'll lower my nozzle temp to 220C and keep up the print speed. If the nozzle is too cold, the extruder forces become too high, and the extruder skips steps. There's a lot of conflicting needs between print quality, print speed, and heat management.

What I want to do:
Measure filament force as a feedback to manage some of these parameters.
I like my first layers hot and gooey for good adhesion, but after that, regulate filament rate and hot end temperature to match a Cura value. Maybe max calories per cubic centimeter?

How to do it:
It seems like this is not a Cura problem, but the Duet or other controller issue.
There's a strain gauge already built into the HE300. It's very noisy though, and the compression force is not transmitted through the tabs. A modified attachment could be made that causes filament force to be carried through the tabs/arms. I think if the strain gauge were sine wave modulated, and by analyzing in the frequency domain, a very good signal could be obtained.
During printing, if a desired nozzle pressure is known, and if the actual pressure is too low, then the travel speed can be increased, the nozzle temp can be decreased, or the filament rate can be increased.

Benefits:
Elimination of curled up corners that are too hot.
Elimination of filament grinding
Elimination of skipped extruder steps (related to grinding above)
Another method of automatic plate leveling (non contact) by printing a layer 1 mesh and iteratively modifying head height based on force feedback.

Another observation not related to this is that I would like to be able to set a nozzle temperature curve based on layer time. This would benefit by having other parameters adjust. Another idea would be to calculate how much heat is going into the part in a specific area. If more heat is going in than is being released, then either increase the layer time, or decrease the nozzle temperature. For maximum production, I manually decrease nozzle temperature now.

So, my question is, is anyone working on anything like this? Some of the features I'm interested in:
1) Modify nozzle temperature based on layer time
2) Incorporate noise tolerant force measurement (electronics work required). However, there may be benefits anyway with little or no hardware modification by analyzing in the frequency domain (no modulation), allowing known sources of error to be removed from the signal of interest.
3) Incorporate filament force into Duet/Cura (this is a big feature)

dc42
Printmaster!
Posts: 434
Joined: Mon Mar 07, 2016 10:17 am

Re: Sampling Filament Compression Force

Post by dc42 » Tue Mar 12, 2019 6:03 pm

What you need is a load cell or similar sensor that measures the force that is stopping the Bowden tube popping out of the collets at each end. The strain gauge built into the HE300 isn't suitable because, like the the Duet3D Smart Effector is is based on, it is designed to detect changes in force rather than absolute force.

User avatar
geneb
ULTIMATE 3D JEDDI
Posts: 4868
Joined: Mon Oct 15, 2012 12:47 pm
Location: Graham, WA
Contact:

Re: Sampling Filament Compression Force

Post by geneb » Wed Mar 13, 2019 8:36 am

So...it measures delta for deltas. ;)

g.
Delta Power!
Defeat the Cartesian Agenda!
http://www.f15sim.com - 80-0007, The only one of its kind.
http://geneb.simpits.org - Technical and Simulator Projects

brianmichalk
Printmaster!
Posts: 39
Joined: Fri Aug 12, 2016 3:20 pm

Re: Sampling Filament Compression Force

Post by brianmichalk » Thu Mar 14, 2019 11:54 am

dc42 wrote:
Tue Mar 12, 2019 6:03 pm
What you need is a load cell or similar sensor that measures the force that is stopping the Bowden tube popping out of the collets at each end. The strain gauge built into the HE300 isn't suitable because, like the the Duet3D Smart Effector is is based on, it is designed to detect changes in force rather than absolute force.
Interesting. For inertial reasons, probably a compression strain gauge at the cold end of the bowden cable. This wouldn't work very well for the hot end mounted drivers.

It would seem to me that this feedback mechanism would work on a physics level. Is this a good or bad idea? Does it warrant the time and effort?

I see a lot of work that would have to be done in the Duet, and Cura to support it.

dc42
Printmaster!
Posts: 434
Joined: Mon Mar 07, 2016 10:17 am

Re: Sampling Filament Compression Force

Post by dc42 » Tue Mar 19, 2019 3:45 am

I don't think any work would need to be done in Cura. If you can come up with a suitable sensor, I'm sure we can support it in Duet 3, and maybe in Duet 2 also. Btw we already have a design for a load cell amplifier interface to the Duet.

Iamarobot
Plasticator
Posts: 12
Joined: Wed Mar 15, 2017 4:10 am

Re: Sampling Filament Compression Force

Post by Iamarobot » Thu Apr 11, 2019 9:49 pm

I'm curious - what materials/bed are you printing with that you're seeing poor adhesion to the bed? Nozzle temp to layer time would be an interesting and useful feature, as well as filament force.

Post Reply

Return to “Rostock MAX v3.2”