Friday, October 17, 2008


PIC32 Printing

As promised a couple of pictures of the last parts I have printed.
The complete screw holder is straight off the board without any trimming, the thin section is more a test to see if the text could be improved.
Material: PCL, 210DegC, 7.5mm/s, 27.5mm/s fast move.
Extrusion rate 9mm/s using 0.5 nozzle

Printed onto MDF.

Dimensional accuracy of the part varies from zero to +0.3mm

Wow that's very good. 210C is very hot for PCL, why so high? Do you think that improves the definition?
First try? Ha! Brilliant work! :-D
Thanks for the feedback Nophead, I seem to have migrated upwards with the temperature as testing with this material continued I started off with about 180DegC using a 0.8 Brass acorn nut, this in retrospect had a relatively good thermal characteristic in that heat readily flowed to the point of extrusion and this meant I could run the fan continuously. Printing from the Brass nozzle was reasonable but I wanted to get a finer resolution so switched to one of Ian's 0.5 Aluminum nozzles. This change presented several problems, first the extrusion pressure went sky high and I sprang a major leak at the PTFE junction. I fixed the leak and reinforced the join to the heater barrel but the pressure was still very high to the point it now slipped on the screw thread. I put the temperature up to reduce the material viscosity and we were in business; that was until the fan went on and we were back to square one. The fan just chills the material in the tip. Next step was loads of insulation right up to the tip. A plumbers solder mat bound up with natural Hemp (Garden Twine – On the old bikes we used to bind exhaust pipes with this stuff so you did not burn your leg. Exhaust headers run at >500DegC) The motor runs very light duty at the high temperature, I can have the fan on full and it seems to work. I was going to ask you if you had any better luck printing the text detail on this part. Another thing I should have mentioned is that the part is made with the extruder running continuously.

My material has now run out so printing has stopped for a while, hence the recent blogging.
By the way, many thanks for posting all your work on this and Hydraptor I have used it as a reference from day one, reading it reassured me that we could get some nice parts off the machine.
Can I ask what control you are putting on your Darwin machine, are you running the Ethernet solution?

Forrest - not first try, just the best so far....
Thanks for your intro into Reprap politics, I dont want to go there!
"Forrest - not first try, just the best so far....
Thanks for your intro into Reprap politics, I dont want to go there!"

Best is just to do what you want to do and ignore dogma in all its forms. Sounds like you are headed in some really interesting directions. :-)
I haven't done PCL recently, but when I did my thermistor was mounted on the nozzle, so running with a fan was no problem. I was extruding at 120C-150C. IIRC Ian's extruder design has it in the fire clay with the heater. I think that may be why people are having problems.

I will probably use a Coldfire with on board Ethernet PHY for my next controller as I have one lying around. I can develop and debug with Codewarrior but release with gcc, so it might be acceptable.

A quick Google tells me gcc can do PIC32, so you might be able to do the same. I.e. develop with MPLAB but release source code that builds with gcc.
I have posted some pictures of Reprapped text for comparison. About the same I think.

RepRapped Text

I think it needs a 0.3mm nozzle to do it justice.

I am running one of Ian's extruder's and it is fitted with the thermistor in the fire clay just above the tip. For calibration I had a 3mm K-Type thermocouple down through the heater with its tip resting on the back of the nozzle. Temps measured this way were about 10Deg lower than at the site of the thermistor. I remapped the thermistor to the thermocouple. Not the most accurate method but readings should be in the right ball park. I will confess that I have not done any measurements at the tip with the extruder working, having said that, I would be amazed if the true readings were down as low as your settings, maybe the truth lies somewhere in between. I rechecked my notes on the items pictured on the blog, the complete screw holder was done at 205DegC and the test at 210DegC to see if it made any difference. The results are far from conclusive but I think the hotter one has the edge on detail. I will complete these tests as soon as I get resupplied with plastic and let you know. I will also see if I can recheck calibration when running to make sure we are comparing correct temps.

Regarding the compiler, I had also seen some reference to GCC even on the Microchip web site, I think that C32 compiler was developed in GCC. I must have a look to see how compatible they actually are.

Thanks for posting the pictures, I think you still have the edge on detail, I have yet to try other plastics on my machine, I thought I would stick with one type for now and try to get some experience under my belt before I move on.

I think that now a few people are starting to print, is it worth trying to put up a Gallery of printed parts? Maybe split into material type each with brief setup notes. It would be a valuable resource to gauge progress as individuals tune their machines and also to catalogue development of the machine.
I think the fire cement has poor thermal conductivity. When you cool the nozzle with a fan the temperature drop across it will increase as more heat is flowing. If the thermistor is in the fire cement it will see some of that temperature difference, so the 10C offset you calibrated with will increase. If you look here I had a 25C difference between the heater and the nozzle with a thin layer of JB-Weld which I would think is a much better conductor of heat than fire cement. And that was without a fan.

PCL melts at 60C and is quite fluid at 120C, much less viscous than HDPE at 240C, so I would be really surprised if you could not extrude it at 120C, which is why I suspect your readings.

There is a gallery of things I have made here.

Thanks for the comments, the heater and thermistor assembly on my machine has given problems with heat loss and lack of power. The current heater has a thick layer of fire cement over the core, the element is wound onto this, the thermistor is positioned just on top of the heater with what I suspect is a very thin layer of cement. I did not assemble the heater so can not be sure. The more I think about it the less confidence I have in the readings I am getting.

I will try and track down some JB Weld and do a new extrusion head with the thermistor on the nozzle to see if I can reproduce your results.

I wouldn't recommend JB Weld as it only lasts a few tens of hours. If the thermistor falls off you get thermal runaway unless you have some logic in your heater control to detect the situation. Something like if the heater has been on for 5 seconds and the temperature has not risen then error.
Interesting comment about the JB Weld. The heater has never given a problem with material failure of the fire cement, all my changes have been with insulation and nozzel type. The heater has done many hours and shows no sign of degradation. Maybe for the new heater I should go for the thinnest possible layer of fire cement and see how tight we can keep the element to the core.
Sounds like the next experiment is shaping up!

I think the important thing is not to share the fire cement between the heater and the thermistor. I.e. you can stick the thermistor to the barrel with fire cement as long as the heat from the heater is not flowing through the same bit of fire cement to get to the barrel. Think of it like an electrical circuit.
I agree with your analogy its all about keeping the “resistance” to heat flow very low down to the nozzle and high back up through the feed mechanism. The plan for the next heater barrel is to get the element close to the core, maybe machining semi-cured cement back to give a uniform thin layer. I think you have done the same with the JB Weld. The thermistor will be either on the nozzle or at the interface between nozzle and the barrel.

I did a quick test last night to get an external reading on the tip. Temps without the fan running were 13DegC Lower than at the thermistor. Conduction down through the K-Type probe must have been significant during calibration. With the fan running it would be much worse.
The Aluminum nozzle has a very sharp reduction in section approaching the tip, the internal drill dia coming very close to the external taper. This will give a high resistance to heat flow or in dynamic terms, the temperature gradient will be considerable from the barrel measurement to the conditions at the tip. This is what I was getting at with my initial comments about the dome nut being relatively good having much more cross section to efficiently flow heat to the tip.

Opening the debate up a little, I am not convinced the use of an extended tip of any kind is a step forward. I have seen mentioned use of Pentel pencil for the nozzle, if anyone does go down this route I bet they finish up cutting off most if not all the extended tip, also I think the CSA is small in terms of heat flow. I think it was one of your designs that is the best “off the shelf” version I have seen ie using a dome nut with a very small nipple machined at the point of extrusion. For a more academic approach, I am sure if we put it out there, one of our fellow Rappers could work up the Thermodynamics to give us the optimum profile of the tip.
I just use 0.025 inch copper shim to make my extruder orifice. I dimple a strip of it and then braise it onto the 5/32 inch copper tube that forms the body of the extruder barrel. Works like a champ and has a minimal path distance between the inside of the extruder and the outside environment as Adrian has suggested on numerous occasions. :-)
Post a Comment

Links to this post:

Create a Link

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to
Posts [Atom]