Extruder melt-back woes!

I've built a Mendel out of parts printed by Adrian and Makerbot electronics and running Adrian's Host software. Mechanically it all works very nicely - apart from the extruder.

It took a bit of playing to get the Makerbot boards all talking to one another and the PID heater control to work - but that's all sorted.

The issue I have is that it will extrude nicely for a few seconds (up to about 30s) but the filament is melting back along it's length. I'm using Adrian's geared extruder which can apply plenty of force to the filament (that may be part of the problem).

The filament is melting back along it's length and at the first discontinuity (where the heater barrel joins the PTFE insulator) it forms a plug. As the filament starts to expand, it also starts to cool so the force required to push more filament into the heater barrel increases, slowing it down which increases the back-melt which makes it worse - and so-on.

The pressure behind the plug just stripped the thread of the M6 barrel out of the PTFE which pushed the barrel out of the end and melted a nice hole in the bed!

Since then I've tried half a dozen different heater designs to try to improve matters - but eventually they all seem to suffer from the same problem.

Since there seem to be plenty of people who have got it working - I must have done something wrong!

In an ideal extruder, I presume that the bore of it needs to be smooth - but how far back from the barrel does it have to stay smooth?

Is it better for the bore of the heater barrel to be tapered or straight and at what bore diameter should it start and finish at? How close to the tip should the heater be - should the heat be localised to the very tip or spread along the length?

Assuming you cannot stop the filament from melting back along it's length, how far back from the barrel does it melt to the point that it can bulge out?

Is it better to make the entrance to the barrel bigger than the filament (so you have an air gap) or does it just melt into this, increasing the friction and thus force required to extrude?

Sorry for all the questions - but I'm close to the point of throwing the whole thing out the window and building something else easier instead (like a Fusion reactor ).

I suppose I need to know what the key points, do's & dont's, are in extruder design. I have a CNC shop - so can make most turned or milled things so I would prefer to make my own if possible.

Thanks in advance,

Si

I would suggest that you read nophead's blog from start to end for anything related to the extruder.

---

Bob Morrison
Wörth am Rhein, Germany
"Luke, use the source!"
BLOG - PHOTOS - Thingiverse

I have actually read it - it was that which inspired me to have a go myself rather than just buying an alternative.

It's still difficult to boil it all down into which design aspects are important. Sure I could just copy one of the designs shown - this for example:


But I would rather experiment a little as well. I suppose what I'm looking for is the wisdom other people have gained going down this route. As I said, I'm on my 6th non working heater / nozzle design. What I've not figured out is which of the variables are important. It seams that some extruder designs (Like Adrian's) work for some people - but not others. What are they doing differently? Unless I can figure that out - even if I copy someone else's design there may still be a missing part of the puzzle that will prevent it actually working.

Si

I think you miss-understand where the filament melts. All of the brass barrel is well above the melting point and so is the PTFE that joins it. It probably something like 5mm up the PTFE that the temperature drops below the melting point and higher still before the glass transition. All that region of the insulator must be smooth and slippery. It only works because PTFE is super slippery and it only works if there is a seamless join where the brass meets the PTFE. Any gap gets filled with rubbery plastic that will jam solid.

---

[www.hydraraptor.blogspot.com]

That is pretty much what I had assumed. How small a gap is too big?

I think the problem with the original version was the PTFE expanding as it heated, opening up a gap big enough for it to get caught.

Most of the subsequent versions were attempts at constraining the PTFE so it would not expand radially (and let go of the thread) - but perhaps it just forced it to expand longitudinally and open up a gap?


This is one of the versions incorporating a heat sink into the barrel, a ring to constrain the PTFE and tapered union between the PTFE & Brass.

Where can I buy PEEK from? We have an account with an Industrial Plastics supplier - but they can only supply 4.5m lengths (a bit more than I'll ever need I suspect!).

Si

Edited 1 time(s). Last edit at 07/07/2010 11:51AM by SimonRafferty.

Quote

Assuming you cannot stop the filament from melting back along it's length, how far back from the barrel does it melt to the point that it can bulge out?

Where ever the barrel surrounding the filament is warmer than the filament melting temperature, the filament will melt. Easy, isn't it? ;-)

The conclusion is, the entry end of the barrel has to be colder than the filament's melting temperature. That's why everybody is talking about "thermal barriers", why the above barrel is so thin walled 12 mm from the bottom and why many experimenters put heat sinks towards the top end of their extruder.

In your case, the heat obviously finds a way to the entry of the barrel. Can't help you on finding that way, though.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Simon,

another area to experiment with is moving the melt point closer to the extruder orifice. We currently use copper or brass for the the extruder tip to keep the temperature uniform. But with the nicrome wire wrapped less that a millimeter away from the ABS over the whole surface area of the exposed tip, it does not matter how good a thermal conductor you use, it will get hot there with anything. But the copper carries that heat uniformly from end to end, so that middle of the PTFE barrel is just as hot. If you used a longer tip, and drilled it out of stainless steel rod. If you make the threaded end longer, and use the poorer heat conducting stainless steel, you may be able to drop the temperature at the opening of the heater element below the plastic point of the ABS, so that ALL the melting only occurs inside the metal. This should reduce or eliminate the plugging problem. Wish I had a working extruder to test this on, myself!


But I also agree with the others. If the hole in the PTFE is the same diameter as the hole drilled in the tip, and they line up well, there should be nothing for a plug to hang onto, and the gear driver should easily push it through.


Mike

Edited 1 time(s). Last edit at 07/07/2010 03:29PM by rocket_scientist.

Quote

The conclusion is, the entry end of the barrel has to be colder than the filament's melting temperature.

That isn't possible with the brass barrel / PTFE combination. That has a relatively long thermal transition zone in the PTFE and to be successful you need a seamless join with the brass. See [hydraraptor.blogspot.com]

If you don't have a bottoming tap, you need to remove the last few mm of thread on the barrel as I did here: [hydraraptor.blogspot.com]

You don't have to worry about the PTFE expanding and causing a gap. It does expand about 0.5mm over its whole length if un-constrained, but over one or two thread pitches it is only a small fraction of a millimetre. That wont be enough to jam it.

Quote

If you used a longer tip, and drilled it out of stainless steel rod. If you make the threaded end longer, and use the poorer heat conducting stainless steel, you may be able to drop the temperature at the opening of the heater element below the plastic point of the ABS, so that ALL the melting only occurs inside the metal.

That isn't going to work. The thermal conductivity of stainless steel, although poor for a metal is ~ 100 times more than PTFE. That means all the temperature drop would be in the the PTFE and virtually none in the SS that has one end heated and the other end wrapped in a very good insulator. To make the SS barrel work you need to have a heatsink and a fan to take enough heat away from the top end to keep it below the glass transition. Also, because SS is not very slippery as far as hot plastic is concerned, the transition has to be very sharp and preferable in a tapered section.

This design looks like it is easy to make apart from drilling a stainless steel bolt: [www.thingiverse.com]

I prefer a stainless steel version because it is bomb proof but I have a PTFE and brass version on my Mendel and it has been printing 24/7 for weeks now and never jammed.

Edited 2 time(s). Last edit at 07/07/2010 04:20PM by nophead.

---

[www.hydraraptor.blogspot.com]

Thanks - I think that answers my questions.

In all my versions, the bore of the heater barrel was different to the bore of the non threaded part of the PTFE (only by a few thou) - but it's clearly enough.

I'll have another go - and ream the PTFE and Brass / SS together.

Is there a handy supplier for PEEK rod?

Si

Si,

I've sent you a PM.

Craig

---

----------------------------------------------------------------------------------------------------
[blogger.kritzinger.net]

[www.directplasticsonline.co.uk]

---

[www.hydraraptor.blogspot.com]

One trick I've been using for the simple PTFE - Brass extruders is to use a much wider barrel than PTFE bore. IE, I use a 3 mm bore in the PTFE, and a 4 mm bore in the barrel. Seems to really reduce the friction, and also gives me more room for mis-alignment between the PTFE and barrel. This doesn't work well with an M6 brass barrel though; I usually make my barrels from 8 mm brass stock, and just build the nozzle directly into the barrel.

Thanks Chaps.

Would there be mileage in making the PTFE / PEEK screw into a female thread on the barrel - then the PTFE expanding will tighten the seal. Also, a larger diameter gives a deeper thread - less likely to strip?

Si

I've got a model that has the top of the heater barrel drilled out to match the O.D. of the piece of PTFE tubing that goes through the insulator. The PTFE tubing's I.D. is 1/8", the same as the I.D of the brass barrel. I haven't tested it yet.

---

--
I'm building it with Baling Wire

I've followed the advice of Nophead and the rest of you and made this:


And disassembled:


The barrel is made from PEEK (Thanks Craig!) with a M20 x 2.5mm pitch thread cut in the outside - good, deep thread for extra grip.
It has a 3mm bore hole through the middle.

The pointy end sticks into a conical recess in the next part to form a gas-tight seal when one is screwed into the other.

That part also has a 3mm bore and the outside is tapped M6.

The heater wire and thermistor will be wound on to the little bobbin and the hexagonal brass bit is drilled with a 0.4mm hole.

The intention of this is that the nozzle can be removed without dismantling the rest of the extruder. Likewise the heater can be removed without disturbing the wire or Kapton tape.

The PEEK is set into the extruder mechanism with Araldite Epoxy, like Adrians, but I sprayed it with grease before gluing it in place so once the glue had set, the PEEK could be unscrewed.

Hopefully, I'll have it up & running in the next couple of days & let you know if it works!

Edited 1 time(s). Last edit at 07/23/2010 06:12PM by SimonRafferty.

I have also just made a couple of experimental nozzles from glass tubing. This should eliminate need for an extra insulating plug of PEET or PTFE. I do not yet have an extruder to test it on, but will work on that soon. Orifice sizes seem to be adjustable at will down to very small!

Mike

Simon, that looks great, looking forward to seeing how well it works, as I need to make something a little tougher than just PTFE and brass.

Mike, I'm interested in the glass nozzles - I've been thinking about a glass insulator for a while, but haven't found a good way to attach the nozzle. Any chance you could post some photos as well?

Perhaps a compression fitting on the cold end would help to lock the glass tubing to the extruder body.

Wade

The easiest option on a glass nozzle is just to heat up the wide end and stub it into a flat thing - which causes the end to flare a little. Pass the glass tube through a lump of PTFE or whatever such that the flare stops it pushing through.

Si

That looks fantastic. Keeping my fingers crossed that it works well.

Glad there was enough PEEK.

Craig

---

----------------------------------------------------------------------------------------------------
[blogger.kritzinger.net]

SimonRafferty Wrote:
> The heater wire and thermistor will be wound on to
> the little bobbin and the hexagonal brass bit is
> drilled with a 0.4mm hole.
>

Make sure you have a good thermal contact between the bobbin and the tube, then!

j> Make sure you have a good thermal contact between
> the bobbin and the tube, then!

Good point! I had figured on assembling it with a little heat sink compound.

The Thermistor is attached to the bobbin so it at least will not overheat. Even if the rest of the hot end lags behind by a few degrees, it's not the end of the world.

Has anyone tried using Diesel engine Glow-plugs (heater plugs) as part of the heater?

Si

I left my regular camera in Tucson. and am having trouble getting the new USB Microscope camera to make stills, not AVIs.

`Yes, the best way to mount it would be to flare the cold end with a lip, then us a rubber covered hose clamp. No need for ANY PTFE or PEET! When working with a 6'[' long piece, I could get one end glowing yellow hot while holding the other in my bare hand. It looks like any size opening can be easily made. Next I want to embed the nichrome wire and thermocouple. I will try to get photos soon.

I am hoping to make a much smaller extruder by eliminating the big PTFE plug, and maybe using a smaller motor with more gearing. But that is a long ways off.

I am also thinking of making simple glass nozzles to sell to people who do not have an oxy-acetylene torch and diamond glass grinder.

Mike

rocket_scientist Wrote:
-------------------------------------------------------
> I am also thinking of making simple glass nozzles
> to sell to people who do not have an oxy-acetylene
> torch and diamond glass grinder.

Actually, all you need is a Propane/Butane Blow Torch and a carborundum knife-sharpening block.

Si

Well, that was another failure!


This time, the plastic melted OK and the transition from solid to liquid was only over about 5mm - but it appeared that the friction between ABS and PEEK is too high and it jams.

After dismantling, there was no leakage between the PEEK and heater barrel - but the last 5mm of the PEEK bore was jammed solid to the extent that I had to put it on the Lathe & drill it out.

I suppose maybe a PEEK tube with a PTFE Liner might work?

The back pressure from the jam was enough to crack the Araldite holding the PEEK into the extruder!

There must be something else I have wrong - I have a great deal of perseverance, but I'm running out of ideas. Maybe my Thermistor is lying?

It's a shame there is nobody with a working machine close to here - it would be interesting to swap bits to see if the problem is with the extruder head, the control, the software, the firmware - or me! I don't suppose one of you would like to sell me a complete, previously working extruder?

Si

Hi

I put a PTFE liner in a peek extruder that was always jamming. It works very well on my extruder which is an original Bits from Bytes type.

---

Regards

Ian

Simon,
I am using boroscilicate glass to avoid thermal stress problems. It requires too high a temp to work with propane-air mixtures. Requires oxygen-fuel torch of one kind or another.

But you are right, the diamond grinder is for the lazy, carborundum or other glass shaping stones would work just as well, if a little slower!

Mike

What about Mapp Gas? It will go to about 2000 deg on it's own or nearly 3000 with Oxygen. We've almost given up using Acetylene except for cutting. Mapp is much lower cost for heating and forming applications.

Si

When extruding, there will be a steady stade of flowing material and flowing heat.

When stopping extrusion after printing a part, that steady state is interrupted and heat which was transported down through the nozzle outlet with the material onto the part is now free to move up and melt plastic in the extruder.

I suppose that this has to be taken into account somehow.

Even in glass, the filament could melt over the highest point of the heating mechanism when extrusion stops and -- maybe -- stick to the wall then. (Advantage: in glass you could observe what happens.)

Maybe a solution is to squeeze out molten material (after finishing printing) until the extruder is cold enough to block, thus shifting as much molten material down into the extruder part which heats up again the next time.
Maybe another solution would be to heat and simply wait long enough for the molten and re-cooled material over the hot part to re-melt by heat getting up through the molten-recooled filament itself.

Did anybody already develop the formulas needing to model/calculate the heat transfer in a multi-part extruder (e.g. with sections which have different lengths and heat transport properties) with changing input variables like "heating up/down" and "shifting slow and fast" over time? (Link?)
How good does ABS etc. transfer heat anyway? With respect to PTFE / PEEK? Does this change with different temperatures/states of the material?
It might be interesting to create a simulation software for this to visualize what is happening inside these opaque materials.