Glass nozzles for extruders

Those look awesome! Nice work! I can't wait to try them out.

I finally put nichrome wire and a thermocouple around one of the first glass nozzles. I also for the first time tried my original nozzle build, a 0.0203" mig nozzle mounted in a PTFE block. So I was able to compare the force needed for each one to push out filament at usable rates. The brass nozzle worked fairly well at 200C, and even better at 220C. Hard say exactly what force I used by hand to force a piece of plain ABS filament, but I think about 10 to 15 pounds for the brass nozzle. When I tried it in the glass nozzle, 200C was not enough to get anything out. I suspect that the thermocoupe being outside the glass had to be at a much high temperature to get the plastic up to 200C. At a reading of 240C for the thermocouple, probably closer to 200C inside, it pushed through quite easily, probably less than 5 pounds.


Here is the output from the mig brass nozzle tip. It came out quite quickly at higher temp, and generated a filament about 0.85mm thick.


Here is the one of the first, practice glass nozzles. I was able to hold the other end of the glass through out the entire test, which means that the glass is easily capable of replacing the PTFE as the thermal break. However, the filament that came out was twice as thick, so the force comparisons may not be wholly correct.


This compares the two filaments. The thinner more coiled up one came from the brass nozzle. Thicker, straighter bunch on the right from the glass. Now I need to make some new nozzles with smaller openings. And wrap the wire closer to the tip. I noticed some real time delay between the thermocouple reading and the plastic flowing easily. Not only did the glass need to heat up, but the plastic right at the tip was not as hot as in the middle of the wire rap. And putting the thermocouple in with the plastic would help in getting more accurate readings, too.

Mike

That is very intriguing work. Do you have any thoughts on whether or not PLA would work a glass nozzle?

I suspect that PLA would work as well as ABS, just at a lower temp. However, I have no PLA to test with. That will have to wait for someone else to experiment with. Looks like my next steps are to try inserting the thermocouple directly into the plastic, and to get the nichrome windings closer to the tip.

Mike

I'll be testing PLA when the nozzles arrive, so we soon shall see!

Hmm inductive heating might be an option for you.

Put a short piece of thin brass tube inside the business end of the glass tube and wind an induction coil either direct or on a bobbin around the outside.

The brass tube acts as a single turn, short circuited secondary. ie it becomes a heating element inside the glass tube.

The outside coil induces very high currents at very low voltage into the single turn. Which gets hot.

Measuring the temp inside the tube still presents the same problem. Although as the temp inside will always be hotter measuring it outside will give you the opposite effect.

I wouldn't suggest operating it without some way to limit the temperature. Induction heating is capable of getting v hot quite quickly.

Of and make sure that you have enough space around the piece of brass tube for it to expand or it will crack the glass tube on expansion.

---

Necessity hopefully becomes the absentee parent of successfully invented children.

I continued experimenting today. Did a lot of pulled thin nozzle tips in the thick walled. I think I have some of them getting done to around 0.5mm. When the glass is pulled, the walls get much thinner. If the heating from the nichrome wire does not shatter it, it will heat much more quickly.

I also started working with inserting the tip of the thermocouple into the inside of the glass tip. It was tough to punch through a hole with glowing hot metal rod. Using a dremel with abrasive disk did not work well either. The set of cheap, diamond coated bits I got from Harbor Freight actually worked fairly well. But I could never close the opening again without destroying the sample wire I was using instead of expensive thermocouple wire. Eventually, I pulled out the roll of fiberglass insulated thermocouple wire and tried it a few times. Still no luck. Then I came up with the brilliant idea of using the tube pulling technique to make a glass sheath around the wire, and lower melting point fiberglass. I also discovered that the easiest way to make a hole in the side of the glass was to heat up one spot and blow into the cool end. It eventually causes a blowout, just right for sticking a glass covered thermocouple in.


This gives you an idea how think the glass walls get, and how narrow the point. I have not yet ground off the very tip to open up a hole. Some still had a narrow opening all the way across the two ends of the pull zone, so they have a minimum opening diameter, but is should generate smooth even flow. The others either pulled so far that the glass in between turned to thread, or had such large opening that I used the old technique of heating the tip and letting surface tension pull it closed. Then I can re-open it on the grinder, and get some choice as to size. I also had trouble putting a lip on the other side of the best looking pulled tips. Apparently, there is a lot of thermal stress left, and without and annealing oven it is hard to relieve. So when I start heating the other end and cause more differential heating stress, then stick the tip of the plumb bob in to widen the hole, and the combined stress is too much for it and the flange end fractures and I have to throw the whole nozzle out.



Here is a slightly larger view showing the glass covered thermocouple embedded in the side of the glass. Many other test pieces, too little time to micrograph them all and post. Maybe later.

Everything I tried in terms of winding nichrome around the glass and embedding it or covering it with another layer of glass did not work. Any flame hot enough to fuse the boro-silicate burns the nichrome. I tired a hot air gun, not hot enough, even when dripping molten glass onto it. Looks like the heater wire will just have to be taped on the old fashioned way.

This one should be good for the next full up extruder to see if I can get a better temperature from the embedded thermocouple, if the plastic will flow around it or come to an abrupt stop, and if the tip will shatter from the temperature or the pressure. Hopefully results by the end of the week.

Mike

Edited 1 time(s). Last edit at 08/09/2010 04:19PM by rocket_scientist.

These are more pictures from Sunday's work. I have gotten the orifice size down fairly small with the pulled thick wall tubing, but I think the walls have become too thin, also. The tips are strong enough. I can press against the tips sideways with more force than they should run into hitting the work piece. But I would be more comfortable something thick enough to be long lasting and trouble free. Sorry that the scale is in 64ths of an inch, I could not find a finely divided mm scale.












And this is another attempt at protecting the thermocouple with a pulled thin point before welding it in a hole in the side of the extruder tip.

Mike

Those tips look great!

Although the flame temperature tables I found indicate that MAPP gas/air is only slightly hotter flame than propane/air, it makes a real difference when trying to work borosilicate glass. The propane torch had no effect, the MAPP gas torch, without supplemental oxygen, was able to melt the glass and make it easy to work. So if anyone else wants to experiment with glass nozzles, you can buy the glass tubes fairly cheap, and just get the biggest, highest rated torch for a MAPP gas can that you can find locally. I will see if I can make some youtube videos of the nozzle making process later. However, it did seem that I could get better nozzles with the smaller flame from the oxy-acetylene torch.

Mike

I just finished drawing, annealing, and grinding the third batch of glass nozzles. Despite breaks, blobs, and other failures, 11 good ones were produced. 5 are spoken for, but I will put the other 6 up on ebay as soon as I can.

This batch came out much more long and slender, and the tips have VERY small openings. I have no way to measure anything this tiny, other than comparing lines in the micrographs. It looks like a few of the tips came all the way down to 0.2mm! It would take forever to print a single part, but it would be very smooth and accurate!

Here are the pictures:





















And here is the link the hi-res picture of all of them, to see how long and tapered the tips are.

Mike

I would imagine the best way of grading the apertures is to use a sequence of wires with a known diameter a bit like using feeler guages.

---

Necessity hopefully becomes the absentee parent of successfully invented children.

I have seen sets of 'pin gauges', but the price is typically rather a bit much for just making a few glass nozzles. But I have ordered some drill bit blanks in the smallest number sizes, which will help measure around 0.3mm to 0.5mm. I may try a Harbor Freight cheap set of numbered drill bits, if it goes all the way up to 80, not just 60, to measure the 0.5mm to 1.0mm sizes.

On a side note, and just barely avoiding triple posting in my own thread , last night I discovered that one of the blunt nosed glass nozzles from the first batch got left in my pocket when it went through the wash and dryer. The dryer was making a horrible clanking sound, but when I opened it to search for the coin or bolt that got left in, I could not find it. The borosilicate glass is apparently VERY strong, and the blunt nosed nozzles should be very sturdy, long lasting, and forgiving of Z axes crashes, or dragging the tip across a work piece or even bed. I doubt that the new, drawn tips will be as rugged.

I am thinking about trying to make some more nozzles this weekend, enough to put a range of types and orifice sizes on ebay to see if anyone else in the group is interested in trying them. And I also want to try wrapping nichrome wire and a thermocouple around a tapered nozzle tip to see if it will work even better than the blunt tip.

Mike

Drill Bits are a great idea why didn't I think of that. They are usually precision ground.

---

Necessity hopefully becomes the absentee parent of successfully invented children.

I have placed the remaining 5 dawn to a fine point nozzles on ebay, and listed them in the for sale section.

glass nozzles


Let me know if anyone would like a specific size, shape, orifice size added to the list.

Mike

Just to let everyone know, the glass nozzles I ordered from Mike arrived today. =) They look great! I'll post more about this later because right now I'm in the middle of a major poster competition and am very busy.

I would like to see you attempt to make a blunt nozzle using a 0.5mm drill bit as a mandrel (i.e. put the drill bit in the tube and shrink the glass onto it, then remove it after it hardens a bit). I dont like the idea of the drawn nozzles. They look a bit flimsy to me. But using a mandrel may allow you to get a small hole where it stays small for some depth allowing you to grind the end to a nice rounded shape without enlarging the orifice. Any ideas as to what may make a good release agent to allow the drill bit to be removed more easily?

Greg,
I agree completely. MY main goal here is to increase reliability. From what I have read so far (I still don't even have a repSTARP of my own working yet), there are two main areas of difficulty. The first is software/firmware/tool chain when first getting started. Once you have all the problems ironed out and several good parts made, then next major headache is the extruder failing, shifting, plugging up, stopping, etc. I hope to develop a nozzle design, and extruder stepper, gears, temperature controls to go with it that work for years between failures, not hours.

I was experimenting with how far I can go getting very small nozzle openings by drawing the glass. But as you have pointed out, that makes the tips much more fragile. They are still much stronger than the look, but would not survive the first Z axis crash. The blunt nosed tubes will survive pretty much anything the rest of the extruder can survive.

Greg, I recently purchased a starter kit to make green frogs on a blue ball. The kit included a tiny bottle of release. I will try working with that and make some blunt nosed nozzles with a longer 0.5mm section. However, in my brief attempts at hand extruding, even the funnel to nothing shape at the tip with no cylindrical section at all and it still produced some very straight, smooth filament. Better then the brass nozzle with a 0.0203 inch drilled in opening. But I will make a few for you anyway.

As for nozzles made without a mandrel in the opening, so far I have nothing better to measure the diameter of the opening then the micrographs I have been posting. But I just ordered some drill blanks in the smallest, numbered sizes, and some even finer diameter copper wire to use a pin gauges. And I also purchased a 50X, built in light and reticle that measure down to 0.02mm, so I should be able to better grade the nozzles I make. I can also use the #76 drill blank as a mandrel to make the 0.5mm opening uniform for a millimeter or so in.

Thanks for the feedback!

Mike

Hi Mike,

... have you tried with embedding metallic wires before tapering the tube and etching the metal away, so you receive a perfect hole with the diamter of the wire?

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Once you have a way to make a strong tip with a reliable nozzle size (using either a mandrel or as Viktor suggests, just etching a write), my next concern with these nozzles would be getting the heat from the nichrome (or whatever heating block is used) into the plastic. The experiment that you did suggested that the thermocouple temperature needed to be higher than required for a brass nozzle. That may mean that it is difficult to control the heat in the melt chamber. It will fluctuate with extrude rate because there will be a big temperature difference between the outside and inside temperatures when extruding at a steady rate. When you stop extruding, the inside temperature may rise?

My final concern is about how well PLA sitcks to glass. Will it be possible to extrude it or will it just jam?

Glass seems to be the ideal material for the thermal break, but perhaps not so ideal for the nozzle. I wonder how difficult it would be to make a hybrid nozzle that had a metallic heater block with a tapered glass thermal break? Ever done glass blowing? See my attached back of an envelope scratching (literally).

Another thought I had: Would it be possible to create an internal thread in the glass using a threaded rod as a mandrel?

Id really like to see some experiments conducted where you try to shrink the glass onto something. That may make fabrication much more repeatable. And open up possibilities for how to attach the glass to the extruder/nozzle.

Edited 2 time(s). Last edit at 08/20/2010 07:47PM by Greg Frost.

Victor,
The etching will likely not be needed, because all metals have a larger coefficient of thermal expansion than borosilicate once the whole thing cools down, it should split off all by itself. Hopefully testing this weekend, assuming the drill bit blanks arrive in time.

Greg, I do not have any PLA as I was not intending to use any, so I can not directly test that myself. I will try to experiment with blowing the end larger, then inserting the mig welding tip (or any threaded bolt just for testing) to see if they can be combined. But it will likely suffer the same problem as provides the fix for getting the metal rod out. When the assembly cools, the brass will shrink more then the glass, and it will break or separate. Even if it doesn't, it will be under lots of stress and prone to failure. I think the threaded fitting to the end would be best, as then it could be connected cold.

I have ideas on how to embed the wire in the glass, but so far all have failed. When the flame is hot enough to soften the glass, the nichrome wire just burns away. I just received some larger diameter glass tubing that should fit over the outside of the 8mm stuff. That may be the way to get the thermocouple and nichrome part way in and also be quite protected. I am also experimenting with slotting the tube and glass welding the tip of the thermocouple parallel to the tubing so that it is very close to the plastic, but not quite in it. All attempts so far to actually extend the thermocouple tip into the melt zone created such a blockage that the nozzle might not work. Or made such huge holes that most of the plastic would be coming out the sides! My ideal at this point is a blunt nosed, ground down a little back of the tip to wrap the nichrome around, the thermocouple partially embedded in the glass, and a second layer of glass holding everything inside. Also, I want to try to transition from nichrome to copper wire under the outer glass to have the wires that stick out be strong and robust.

Mike

... i'm thinking loud - what's with brazing the glass to the nozzle?

I know there are some companies soldering or brazing glass ... my idea is to have the glass on the outside, the brazing/soldering paste between the glass and the nozzle ... and melt the paste through the glass with a diodelaser - i have 5Watt-modules or can combine some fibers to [n]x5Watts on the same spot ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Greg,
I tried 3 nozzles with steel wire inside to make a constant width section. The two best ones both show a lot of bubbles around the wire, probably some form of out gassing when the wires were red hot.



I will try etching away the wire is each one and see what happens, but so far I bet that the naturally formed ones will work much better. No pun intended, but they transition from 3.4mm to near 0mm smooth as glass.

Victor,
I am about to try blowing some glass tubing to get it wide enough to go around the threads of a bolt. For what you are talking about it might be easier to use a high temperature adhesive.

Getting #10 stainless steel bolts to stay in the tip was not too difficult. But the slightly large diameter mig welding tips went badly.


Surprisingly enough, both stainless steel bolts can be unscrewed from the glass once it cooled down.


The copper mig tips are harder to turn, perhaps in the extreme heat they have become out-of-round.

Looking at all the glass/metal interfacing I have done so far, pure glass still seems to be the way to go.


I have also started about 30 blunt nosed glass nozzles to put up on ebay, for those who prefer something more solid!

Mike

Interesting about the bubbles. I wonder if you need to heat treat the wire first. I take it you cant just pull the wire out once it has cooled.

Right. The larger 3mm steel rod pulls out somewhat easily, but the thin ones have much more surface area for the amount of shrinkage. I have tried my usual pcb etching solution, a mix of muriatic acid (hydrochloric acid) and hydrogen peroxide. I got some bubbles, but no significant reduction. Then I put it in ferric chloride, still had to see any change. I will leave it in there overnight. But I now wonder if I have the right etchant.

Mike

... i have some glass-powder from Schott, which is melting to solid glass at around 400 centigrades the first time, then need above 700 centigrades for remelting again ... maybe a good material for testing?

I think for the wires you need something more heat resistant (tungsten, platinum) or a glass melting at much lower temp ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

More glass nozzles for sale for those interested in trying them out.



$7 each blunt nosed, flanged glass nozzles



$5 each without flange and index key glass nozzles

More details in For Sale section.

Mike

Can you ship to Australia?

Looks like I can do USPS First Class International Package to Australia for the same rate as Europe. As far as I can tell, there should be no problems with the customs forms at my end. I am not as sure what happens at your end, but any duties taxes would be small. If you want to make a special request and can pay through paypal, I would be glad to put a package together for you.

Mike

Unless the situation has changed. Declare your parcel as samples, no commercial value and there should be no duties at the other end.

---

Necessity hopefully becomes the absentee parent of successfully invented children.