If you need 2mm of retraction distance then...

If you need 2mm of retraction distance then you need 2mm of retraction distance. If the hot end does not work with the retraction distance you need then the hot end has failed.

I think evidence is building for PLA filament requiring it's own hot end design. One hot end design extrudes all could be achievable if PLA was not included; like the all metal hot ends.

If the hot end is designed to work with PLA filament then PLA will extruded regardless of what retraction distance you have set. Tweaking g-code to solve jamming issues is just not the way. The hot end is either not designed to extrude PLA or there is an assembly quality issue.

If the hot end is designed to work with PLA and the hot end is assembled correctly, then PLA will extrude.

I guess the point is don't waste hours or days tweaking g-code to solve jamming issues.

All I use is PLA filament and have never got jams with the current hot end but I'm after buying two new ones. The latest post on my blog is about getting data from hot ends that can be used to improve bowden extruder g-code configurations for better quality 3d prints.
Airtripper Extruder Filament Force Sensor – Introduction

---

Blog - [airtripper.com]
Extrude Filament Force Sensor - [airtripper.com]

All I can say is WOW! This is great work and a great contribution to the community. Thanks so much for your efforts.

KDog

Great blog post, and I have actually been looking to utilize load cells in a different area of the 3d printer.

Anyways, have you gained any insight from this about specific qualities that a 'PLA specific' extruder might have (geometry, etc)?

After having a crap hot end, i can say that spending the extra money is worth it. the hot end i have can use any retraction setting, pla, or abs. its the Europa, and we are working on a new design that will be able to handle any type of plastic at temperatures above 400c, all with zero jamming, and no failure due to the well thought out design and materials used in its construction.

I appreciate the comments guys, thanks.

Basically, the extruder filament force sensor idea is just about a month old and have had very little time to do a proper range of tests with being busy with other stuff. Before I do any more testing I'm going to get the build guide wrote-up to allow other people to get involved and do their own testing.

I built the extruder filament force sensor because it seemed like a good idea after reading the following snippet from another post in this forum. (Full thread, message May 02, 2013 03:55PM )

Quote
thejollygrimreaper
the big problem i think all of us have now is that there is no feedback from the extruder as to how much back pressure/resistance there is until it's too late, so at the moment we cannot measure and test for upper and lower limits of this speed band,

I already had a test circuit with load cell ready to go from another project I was working on from about a year ago; so I had the extruder filament force sensor prototype up and running within a day.

It is only now that I am fully realising how important this piece of kit can be. It is not just a filament force sensor, it is a 3d printer health monitoring system in it's simplest form.

I believe the filament force sensor would be essential to those who use a lot of filament and want a faster problem solving route. I think a filament force sensor kit will be especially essential to those who will be using their own home made filament, whether it's from granules or recycled plastic.

With wifi enabled printers, performance data logging and and 3d printer health monitoring from a far would be possible.

I've got an easy breadboard circuit non soldering solution guide for this project coming soon which would allow most people to get up and running quickly.

---

Blog - [airtripper.com]
Extrude Filament Force Sensor - [airtripper.com]

I think this is a great piece of work and would be a massive benefit to printer tuning. Can't wait to get the plans!

With my printer the extruder and hotend degraded slowly over time, going from working nice to barely making a print. Coincidentally I had changed filaments so went on a wild goose chase with that. After doing a refurb of extruder and hotend I am back to working nicely - even with the filament I thought must be "faulty".

In the process I bought various new hotends, but evaluating relative performance is difficult. I suspected some require more extruder force but without proper measurement it becomes guesswork.

Am I right in saying that the filament load is equivalent to pressure in the hotend? Could be interesting if the firmware could use the feedback in realtime to adjust the extrusion.

Having the data from the force, temperature, the E position, power from resistor (possibly: measured E position) simultaniously in machine-readable form would be nice. I mean the theory that it cools down sounds possible on its face, but we might be in reach of actually having a more specific model of that, and testing it.(Do note that there do seem to be waves in the prints)

One guess at a more specific model:

F = k(x-E) = (dx/dt)/f(T) (edited: + → =)

With E the E-position and x the length actually consumed in the hot end. k is the'spring constant' of the filament between the drive and hot end, and f(T) is the friction dependent on T, f(T)=f₀ + f₁⋅(T-T₀)) could be a good shot.

Secondly you'd want to model how the temperature changes; something like

CdT/dt= P(t) - βdx/dt - γ(T-T_outside)

CdT/dt is the heat energy changing P(t) power from the resistor, and the gamma term is heat conducting out. However i think you need to model it at least as two 'nodes' or perhaps three; model heat conduction between the power resistor, thermistor and filament.

I dont know how easily to get the data. Imho it should be easy to pull real-time data and use whatever tool you want to use to control the printer at the same time. I think mpd is a good example.

Note: this post is very incomplete. See if i can make a model and a way to actually fit it to data. Why do you think the force climbs slowly?

Edited 1 time(s). Last edit at 06/02/2013 07:17PM by Jasper1984.

Quote
Jasper1984
Note: this post is very incomplete. See if i can make a model and a way to actually fit it to data. Why do you think the force climbs slowly?

I think the slowly climbing extrusion force is a characteristic of my particular nozzle design due to the nozzle orifice extending quit far from the heater block. My theory is that as well as heat being lost from the nozzle tip, heat is also lost during extrusion.

I believe the nozzle tip cools during extrusion causing the plastic melt temperature to drop. As the melt temperature drops, the filament becomes harder to extrude causing pressure to rise. The graphs are about 40 seconds across and with the nozzle being made of brass I think there is enough room for a heating and cooling rate to ramp pressure between retractions.



In graph 11 you see a lot of problems, but if you are looking at the graph being plotted during the whole print you would see a small climb in nozzle pressure. I believe this is due to the heated bed temperature having less influence on the nozzle; making the nozzle cooler. This could explain why the second half of the print looks suddenly worse than the first half (like the image above). It could also explain why nozzles jam after an hour, for instance.



The filament slippage and extruder stepper motor stalls shown in the graph above became more frequent towards the second half of printing as the pressure increased.

Edited 1 time(s). Last edit at 06/03/2013 10:18AM by Airtripper.

---

Blog - [airtripper.com]
Extrude Filament Force Sensor - [airtripper.com]

great work Airtripper!


I'm still reading the details.


back pressure does change during print. I think what that some of it means though is that flow rate if even slightly too much can eventually lead to too much back pressure. dialing it in is critical.

Any updates on this since your June 27th blog post? I'm getting close to wanting to try this on my machine, and was curious if you were any closer to providing a write-up before I go the road alone.

This is very useful information.. and even somewhat contradictory to what I have read about PLA elsewhere. Some people say that the viscosity of PLA shouldn't decrease with higher temperatures.. which led me to believe that everything in the 185-220 range should be similar. Your plots clearly show that increasing the temperature can significantly decrease the force required to extrude the material.

If temperature oscillations at the nozzle are indeed important then there are some things that *should* be done to remedy it -- for instance, providing a clamping force to the heater for better thermal coupling and using a material with a higher thermal conductivity. Brass is not a terrible thermal conductor, but it's not that great either as far as metals go..

What ive noticed is that if you run it faster, it wont all be fully melted, so thats why you run the temp higher, in order to melt it more quickly. for instance when your running a 0.6mm nozzle with 0.45 layer height, the plastic is moving very quickly through the nozzle, if I run it at 30mm/s its fine, when i run it at 150 or 200mm/s the plastic coming out is not fully melted and becomes stringy and clumpy.

With abs running at a high temp is an option because it really never melts (amorphous) just becomes more liquid like, but with pla it eventually does 1 of two things, softens to glass transition above the hot end which is a death sentence for a hot end. or crystallization and break down of material which leaves solid chucks that require disassembly and torching and cleaning to get nozzle back to normal performance.

Don’t go much over 220deg at the hot end unless you have a constant high flow and active cooling of cold end, or a larger nozzle that can work thru it (0.6mm or greater).

Edited 1 time(s). Last edit at 09/05/2013 01:14AM by jamesdanielv.