Could you bypass the patent by having the nozzle two part? lets call it a heat zone instead, seperate it from the nozzle itself and bingo. Patent avoided, this is how patents lead to innovation. Lets do this. Having a heatzone with internal heating "fins" or wires is here by free for all. Refer to this post as prior art in relation to any potential lawsuits!
hey stefan, i think the melting performance of the über-volcano would be even better if you wouldn't have milled the sides down. the milled down surfaces reduce the contact area with the heater block thus reducing heat transmission.
I agree. And to create a flat spot to start the drill, screw the nozzle in a M6 nut. Then drill through the nut and the nozzle, and remove the nut afterwards.
True although it does help start the drill. My suggestion though would be to plunge the end mill only in the locations you plan on drilling the holes and use an end mill only a little bit over the size that you plan to use for the drill bit.that way the solder completely fills the hole and you get full contact. I also wonder if you can get a steel Volcano nozzle. That way you could install steel wires and braze them or weld them in rather than using solder that has a melting point so close to the plastic. This would also allow you to increase the temperature and or use higher temperature plastic.
Interrupted threads don't appear to be a problem, but might be better served with just a 1mm slot on each side compared to a full reduction, yes. Which incidentally... might allow affixing the wires in a different way: don't drill through but stick it into the other side, then bend up the wire into the channel before soldering it in.
Depends. As long as the heat transfer into the nozzle is >= the heat going into the filament then it doesn't matter. Think of the nozzle as a bucket with a hole in it. And the heat flowing into the nozzle as a hose filling the bucket with water. You can restrict the hose as much as you want as long as it flows at least as much water as the hole in the bucket then the bucket will never empty. As long as you can fill the bucket as fast as the hole can drain it then the bucket will stay full. He only needs as much heat transfer into the nozzle as the nozzle transfers into the filament. If this happens then the temperature of the nozzle won't drop.
If you want to keep the bit from wandering, make an M6 nut with 3 sides, and tighten the nozzle inside. Then you can use the flats of the nut as your drill point without having to sacrifice the threads on the side of the nozzle. It's essentially just making a drill bush, but for a radial drilling direction instead of an axial direction.
@@samthenerf Good idea. Better still, a square nut (maybe a long one, made from a bit of brass bar) with a longitudinal split right through one side, and the guide hole drilled opposite the split. Then when you tighten the vice it'll flex and clamp the nozzle threads. Jig and fixture all in one! Bonus marks if one side is rounded so that the flexure of the jig doesn't affect the grip of the vice jaws.
Could also just spotface each hole with a small center cutting end mill and a slow feedrate. Or even just use a chunky center drill barely enough to get past the threads.
Hey Stefan! I'm unsure if it's already been commented, but one idea I have for increasing turbulence in the flow is to offset the copper wires from the center of the flow path by a little. This would cause uneven resistances down each side path, leading to slight travel time differences of the split fillament, and when compounded with the 120° rotations might work well enough to cause some sorts of mixing. Varying the amount of offset for each wire would likely work even better as it'll introduce some chaos to the paths the filament needs to flow down. Just a thought! Good luck with the CMYK mixer idea!
I wouldn't want to braze brass, ie, it's a pain if the filler and parent metal melt at the same temp because the whole nozzle can melt. Sliver solder would be better, or even 99% high tin solder. I personally would just peen the wire ends to seal them, or centre punch the ends so they expand.
@@SidneyCritic good point. I just looked it up and brass melts at around double the temperature of brazing solder. Brazing copper tubing and brass fittings is common in HVAC applications
@@nrdesign1991 It's more a thermology thing. Like some people classify silver soldering brazing because some SS is brass coloured. Where as I classify brazing as welding with brass, ie, the same material. When you weld cast iron with high grade brass rods, the weld can be so hard that when it cools/shrinks it can crack the base metal. You are perfectly correct in saying there is fillers that will melt at a higher temp than solder, but not as high as the parent metal.
FYI US patent laws place no restrictions on personal experimentation with patented ideas. Only if you started selling something would it matter. The whole purpose of the patent system is to encourage inventors to disclose their ideas in exchange for patent protections so that other people can understand and improve on the idea. I’ve seen a lot of misconception about this in the 3D printing community (that said, I know nothing about German law). Excellent job as always and very informative video!
Just seperate nozzle and heatzone. make special heatzone, then use bog standard nozzles in the end, boom patent avoided, and hereby declared public domain!
When we are trying to get good mixing in our pellet extruders we choose our dual 20mm extruder with about 10 stages of mixing sections. These are very extreme mixing sections, they look like aggressive toothed sections, and we introduce our fillers at different stages.
@@infernaldaedra I think it would help with turbulence at least since the rounded wire acts more as a speed bump and just slows down the flow a bit instead of causing a lot of mixing/turbulence. Something more chaotic, a bladed shaped obstruction could be used to push the flow from side to side thus causing more turbulence instead of just a speed bump effect of the rounded wire.
Your Über Volcano nozzle looks sooooooooo tempting to try out! Wonder what the flow rate would be at 0.6 and 0.8 nozzle diameter, getting close to the supervolcano but much less mass? Exciting!
Wow, that Uber Volcano looked insane. You could probably make a solderless wire insert by drilling out the holes to exactly the same size or slightly less than the wire being added then freezing the wire (to shrink it) so it fits. As the wire heats back up to room temperature it will create a very good metal-metal contact. Essentially the same thing you would do with fitting bearings or bushings. The biggest issue I could see with that is that brass (ie the nozzle) has a slightly higher thermal expansion than copper, so might loosen the wire at printing temperatures. Aluminium has a slightly higher coefficient than brass, but of course has lower thermal conductivity than copper. Just food for thought regarding the manufacture of home made CHT-esque nozzles
Thanks for the interesting video! To drill holes in the uber volcano nozzle, you can take an aluminum hex jig, inside which you make a threaded hole for the nozzle, and drill holes perpendicular to the centerline that will guide you when drilling the nozzle.
Loving these videos about not only the dual color filament but also these home made nozzles! Keep it up! Really interesting thought experiments and results!
working with microfluidic devices for a while as well as trying to make my own microfluidics mixers, I truly feel home with your video. Thank you like always for your quality content, friendly delivery and top notch video quality
Have you thought about using your router to mill slots in the threaded section of the nozzle? you could insert copper "blades" with sharpened edges to reduce resistance and increase surface area for heat transfer.
You present the topic as a mature experimental physicist. You have attracted outstanding comments from your viewers which are a treat - no crappy comments. Well done, Stephan.
I'm sure I am not the first to think of this. But I have not seen it mentioned, and I have watched 8,000 videos about 3d printing. The idea is that as you print layers (arbitrary for example) 9, 10 and 11, you leave holes or slots in strategic places. Then on layer 12, you pump those holes full of hot filament and you have effectively created a rivet that securely connects layers 9-12 and improves layer adhesion. You have run across this idea before? Willing to test it????? Thanks for another great episode.
@@andreaspoulsen8017 I'm hoping Stefan spots it and goes on to test it. He has plenty of expertise to do it. If it works I would approach Cura at that time.
Truly, you are doing some fascinating work and education materials for everyone to watch, which I am truly glad that you are here with us, stay safe and be healthy, my dear friend. ❤
it has been done many years ago. It's called the Diamond Hotend, and there's 3 and 5 color ones. See my post further up for my experiences with it! ;-)
Hi Stefan. Great video. I did my thesis on the topic of static mixing nozzles in college and thought I’d share my findings. We tested the three-way Diamond hotend extruder you showed(10:12), the Zmorph dual filament head, the M3D Quad Crane nozzle, and finally a custom metal 3D printed nozzle with a static mixing design similar to epoxy mixing nozzles you showed(4:25). Although the Dimond hotend, Zmorph and M3D Quad Crane fed several colored filaments into a single chamber, they didn’t add any turbulence or mixing agent to mitigate the “toothpaste effect” where colors were only visible at certain angles. The 3D printed metal nozzle with epoxy mixing design by far had the best results. Would be happy to share images but can’t post pics on RU-vid comment. The ultimate pitfall that kept from continuous printing was leakage due to significant back pressure from from forcing the two filament through the mixing chamber. We never tested a dual-colored filament like the one you tested. Feeding this single filament could reduce back pressure and allow for continuous printing, just an idea. Best, Noah P.S. I’ve been a big fan of your channel for years, thank you for your contributions to the 3D community.
1. Use silver plumbing solder, goes a lil hotter and is stronger. It also conducts more heat. 2. Use a torch to get the whole thing hot, the just let the solder wick in. 3. This is great. I'll post some of my custom nozzles once I get my printer working again. 4. Patent laws don't apply until you sell it. Experiment away!
1. If you want to drill holes in surfaces that are not perpendicular to the drill, consider using a center drill. 2. You could tap the cross holes and use threaded stock instead of round wire. This would increase contact with nozzle and filament. 3. In the US, as long as you don't sell (in this case) the nozzle you can build and use whatever you want. Finally, your videos are uniformly excellent. Thank you.
First time commenting on your channel. The Uber Volcano looked fantastic. Thank you for all of your efforts and videos. They are a great inspiration for 3D printing and building in general. Keep up the good work.
Hi Stefan! I have some advice for your frozen bread rolls from some tests that I conducted a few months ago. Based on my findings, going for a lower temperature and a longer time will usually always improve bread quality, however, the temp-time ratio often is dependent to the volume of the bread since a bigger volume will need a longer time for the heat to reach the middle. That's all. great vid as always hope my advice helps!
Servus Stef, have you considered silver solder for fixing the wires? I forget the soldering temp (might be too high for copper), but it's much higher than led and way stronger too... Also using a spotting drill (6-10mm short stiff drill bit) to start the holes prevents the walking you've experienced without having to mill the threads flat. Just need to make sure your work holding is rigid. Another option is to thread a brass nut on the nozzle and use the flat sides to start your drill and just drill straight through. Great videos and lots of fun watching the progress of your diy nozzles! Vielen Dank!
I think it's great that you show the mistakes as well. Milling down the sides of the über volcano nozzle probably reduced the heat conduction into the nozzle somewhat. If you can figure out how to drill holes without milling the sides down, maybe your nozzle would perform even better.
@@infernaldaedra I'm not clear what "shal" means, but I think if the wire was made to have an elliptical or even rectangular cross section, even more heat could be transferred into the filament by the wires, due to the increased surface area.
Gotta love engineers: the only type of crazy people to film bread heating with a thermal camera. Proud to be of that species. Also that censored print at the start is very suggestive, love it.
It would seem that removing the material from the nozzle would reduce contact surface area thereby reducing the amount of heat transfer. I wonder how much heat transfer is lost on the flat surface areas and if the flow would be much better with less material being removed? I Love the experiments. You can learn so much from doing them.
Great video. Very interesting. Just a few observations/suggestions... 1. If you didn't already, I'd consider using HMP Solder (high melting point) / Silver Solder / Brazing Solder to ensure that it will not melt or soften at normal 3DP temperatures. 2. Do not mill the entire length of threads flat. This decreases the the surface contact area and reduces heat conductivity which could result in a cooler than expected melt-zone. Instead, just use the edge of a file (or a small milling bit) to flatten the actual place where the hole will be drilled. 3. Make the holes smaller so that wire is a better fit and minimize the risk of solder creeping and blobbing inside the nozzle.
You touched on it briefly, when you mentioned the 'smooth' copper wire, but one of the features of the static melt mixers that you show at 4:38 are sharp corners. The corners provide places for eddies to form from.
I wonder about if you were to put slots in the walls of the nozzle threading and put a DE safety razor as what would split the filament rather than the soldered copper wire @CNC Kitchen? I also think putting flat sides on the Uber volcano would have a negative effect on "even" heat transfer.
Pretty cool that I made it into the video. I actually found out on my own when I bought some green/white quantum for Christmas prints. If you have a large flat surface and use a octagram spiral for a top or bottom pattern it looks amazing.
Milling flats on the thread is overkill plus it reduces heat transfer because you removed metal in a crucial area. The best way to do this is to start your holes with a stubby endmill the same size or a little larger than your desired finished diameter. The start hole created by the endmill only needs to be a few mm deep. Once you have your start hole created by the endmill you may now finish your through hole with a drill bit. For those who aren't familiar with machining metal, the reason you use an endmill rather than a drill bit is because and end mill is designed to cut on the outside and bottom therefor it will not deflect when it encounter side forces such as cutting into a threaded screw, as seen in this video. Using something called a Center Drill, otherwise known as a Combination Drill & Countersink, instead of an endmill would also work if you don't have a small endmill. Endmills are designed to work in milling machines so if you only have a drill press a center drill bit would be a better choice for starting your hole in the thread..
You can use a spot drill to start the hole, and/or mill a tiny flat spot with a small end mill and drill into that. You have the cnc capability to that would be easy to do!
Out of curiosity, if you have access to a lathe, of course, have you considered making your own nozels? That way, could you implement your copper obstructions with more control whilst also maximising thermal conduction? The second question is regarding your soldering. Would the use of a more precisely applied flux with a torch allow for a better thermal interface?
Your machining is pretty good. Id like to see you try skipping the soldering an use press fit pins, that wouldn’t need to go all the way through. Copper might be too soft though.
I think that making at least the first* copper wire "sharp" or at least shaping it so it has a triangular´ish´ cross-section should help split the polymer with less resistance , increasing flow, and of course finding a way to make centred holes without having to remove half the contact area (the nut trick might work) would turn it in pretty much the best nozzle ever, Great Video. *(the one that comes in contact with the molten polymer first)
What if you used a Tesla-valve in reverse? It'd obviously result in a lot of resistance and thus slow flowrate, but I would expect it to cause turbulance.
I wonder if a mixing nozzle would allow 3d printing of thermoset polymers? Some sort of 2-part system where both can be extruded alone as a thermoplastic but when mixed chemically react to produce a thermally stable end product. Kind of like the 2-part epoxy mixing nozzle but instead of liquid epoxy the two parts are in a plastic matrix. Cool video! Would be interesting to see how the uber volcano performs in the internal stress tests.
This is an interesting idea! I guess the nozzle would get clogged though unless there was some way of flushing it after use so that no material could set inside it.
I tried making a diy CHT nozzle by just cutting a piece of copper sheet metal the width of the nozzle inlet and sticking it in there, theoretically splitting the flow in two. It seemed to work, I was printing pretty fast and thick in petg with no issues. That was actually the reason I went with this method - petg temps are too close to solder melting temps for my comfort.
Hey stefan, Your phrasing on mesh gave me an idea. Could you try adding a literal metal mesh between the nozzel and heat break? This would only require you to seal the mesh so there isn't any leaking in that area and no drilling but would also increase surface area
I have the Sovol SV02 which has a 2-into-1 hotend, but it's not a mixing hotend. You have to extract one filament before you can insert the other. I picked up a 3-into-1 hotend for a custom build but that works the same way. It seems this is the way most of them work however Sovol does have a 2-into-1 hotend for the SV02 that they specifically market as a "mixing hotend" (although, I haven't tried it), so might be worth reaching out ot them. :)
Any thoughts on a nozzle system which would allow end users to create their own dual color filament. Rather than feeding in one filament that is already dual color the nozzle would take two filaments and line them up next to each other.
would be probably easier to solder the wires to the nozzle using a mini torch to heat the nozzle up to the solder flow temps. Similar to how the copper pipes are (were) soldered. Great video as usual, thanks for sharing your knowledge!
I would love to see how profiling the copper wire would do for the flowability and heat performance. Something like a tear drop or hydrofoil since the CHT nozzle has sharp ingress dividers.
Using flux before soldering will improve your joints. I will also suggest that you use a small end mill to create small flats where you plan to drill the nozzle, grinding less material from the nozzle should improve heat transfer from the block to the nozzle. NS Tool makes end mills as small as 0.1mm.
I'm actually really surprised by the lack of mixing despite the round nature of the copper wire obstructions in the flow path. At what level of viscosity would it start to have turbulent flow and begin mixing? Very excellent video, Stefan. Thank you.
Thanks alot for this...saved me alot of work trying to get mixing from my Geeetech a10m...I guess the higher viscosity of filaments makes it pretty much undoable...
Not sure what the machining setup would be like, if even feasible, but what about having the copper wires come through at angles? They’d still go through the center of the cylinder, but angle up (or down). That way there’s more surface area covered and heat the “inside” of the filament for longer.
Would a standard nut pre-drilled on 2 opposite sides be an easy guide to trial the holes for the wires? It already has 6 faces, so rotating it the right number of turns would make it easy to achieve the angles, the nut face to drill is flat so it should be easy. And the result would have even more heat transfer since you would not need to made flat faces on the nozzle. Just an idea. Great video. Thanks
tip for drilling, use a centre drill, it does exactly what its name tells you, drill a hole in the centre, with a angle so the reggular drills can fall into the exact centre
have you considered cutting the tip off a cht nozzle and using it as an extension to create a cht volcano? you could connect the two via key and slot so they would srew in/out together without the risk of getting stuck in the hotend. it might melt the filament really well without the weird flow from a regular cht nozzle.
The dual color filament combined with 3d printing layer lines is actually a pretty convincing illusion of an anisotropic material. Not specifically anything like iridiscence or pearlescence or brushed metal but something vaguely in that realm.
This is awesome, I look forward to messing around with this. So glad your always pushing the industry further and testing with great accuracy. Could you talk about or look into wether or not keeping consistent volumetric flow rate close on all areas or none of them (ex: I find it is good to keep the volumetric flow rate close to equal from infill to walls, which even though sensors show not much change in temp I feel the nozzle it self would still loose some temp if only a micro second, but that micro second can be the difference between good and exquisite walls and start lines) learned some information about this from the channel Pro3D.
Another cool video. Can I ask where the video is on the prusament pc blend carbon is? Very curious on this filament, especially after your great standard pc blend video.
I have several GEEEEEEEtech A10T and you can use cura to create virtual nozzles that vary the mix of each of the 3 colors...which is pretty cool...and it does actually mix them... but CMYK includes black and assumes printing on a white surface. RGB is for mixing light using intensity. You can mix yellow and blue and make green, white and red and yellow to make orange, etc, but it's not light rays so you can't mix red green and blue and get white. I'm not sure if CMYK+white on a 5 color nozzle similar to a direct to garment printer would work or not...
There's a thought; most printers I'm familiar with use an aluminium heater block; good thermal transfer, low heat retention, low cost. The silicon thermal sock helps with the second. But I wonder if a better thermal mass material for the thermal block ie cast iron, steel, copper or brass would be a worthwhile investment & might positively affect feed rate. It's probably already a thing, but I haven't run into anything about it. As others have mentioned, thread contact between the block & the nozzle will affect transfer, I wonder if chasing efficiency may see the industry move to different flat geometry clamp fit nozzles or possibly precision cut threads designed to maximise surface contact - something most thread designs optimised against for ease of use reasons.
I wonder if we will hit a speed limit with 1.75mm filament and go back to 2.85 if these cht type nozzles become really fast (if melt rate is no longer the bottle neck, the extruder motor may be the next?)
Another guy on RU-vid made a 6-input hot end with the express goal of doing fully color prints by color-mixing. He hasn't found a passive mixing system that blends the colors well. His best efforts did show improved mixing, but also took forever to purge and had bad stringing. My intuition is that color-mixing FDM printing will require an extra motor to spin a mixing blade/screw in the hot end. Just have to figure out how to seal the spinning shaft...
So a few thing to try on the Uber nozzle V2, first try using a small bit when machining down the nozzle, also try going side to side only where you want to put the wire, instead of front to back, this should leave more contact area to allow for better heating, second try brazing the copper wire instead of soldering, this will allow you to bond the wires with something like brass or even copper, third I would do this as a separate test but add a thermal compound to allow for better heat transmission, thermal compound fills in all the little voids
should grind down the centers of the rods to form an propeller or twist, so the rods are not round in the center. should only need to drill from one side for all rods, no crossing of rods. i'm sure turning the rods into blades will make it mix.
Do you think crimping the 0.8mm wire in the nozzle, with like a pair of tweezers, would do any different? You'd be making the wire into more of a tear drop shape, so it might cut down on resistance. (barring that the flowing plastic doesn't crush the tear drop into something worse.)
since the 6 wires in the über volcano are offset by 120°, i wonder if it leads to the filament turning in the nozzle. Idk if this question makes sense, but did you notice anything in the tests that the filament was twisted after the nozzle compared to before?
Have you considered an asymmetrical placement of the wires to force the fluid to "slosh" around inside? seems like currently it divides and recombines pretty symmetrically.
Over the holiday break, I printed some roses in Matterhackers Quantum dual-color filament. With the silk effect they look stunning. So much so, that I had to order another color so I could have half a dozen in a vase with different color variations (raspberry/gold and raspberry/silver). Nearly every color was out of stock until Feb, maybe indicating how popular this filament type is right now. It's a good thing that these high flow nozzles don't mix the colors. I'd like to find a way to print with bands at different height having alternating perimeter print direction, so that an color stripe effect can be achieved, but I can't find a slicer that provides control over this. Anyone got any ideas?
If you get a some what linear but not laminar flow in the hot end you should be able to pre calculate the collor of a mixing nossil, having it pretty much making it a CMYK Nossil. Anyway, if you want to use it as just collor changing, having the change over be in the infill should solve the problem
Instead of soldering copper wires through the brass nozzle, could a metal insert be created that does the same job of increasing surface area for the material to melt? It would have to be fractionally larger than the space inside to make sure it stays put, meaning pressed into place and not removable.
Hey Stefan I really wonder what is going to happen if you use something like M1.6 screws rather than this copper wire and soldering technique. I know 1.6mm is may be too wide for 1.75mm filament but who is stopping you to drill wider the main hole ? Also i know M1.6mm screws but maybe there is much thinner screws that i dont know. This way this proccess may be more accessible in my opinion. Have a nice day. :)
For your DIY CHT - had you considered filing the copper wire to a teardrop shape before soldering it in? I know the two-wire version resulted in reduced max flow rate - but I'm thinking that's more of a shape thing than anything else.
Also: You can avoid having to mill off the threads by screwing a pre-drilled hex nut of the right threading onto the nozzle, then using the pre-drilled hole as a guide. To prep the nut, center punch the center of a face, drill through, then use a tap to clear the chaff from the internal threads.
i like how you said " *might* have even melted the solder" while showing us a clip of a bunch of spaghetti with solder _spheres_ (seriously your automoderation got rid of my previous comment using a common synonym for that word starting with B... might want to fix your "bad"word filter) also, i think the main reason your modified nozzles still had laminar flow is because the wires were all neatly centered, leaving all the fluid paths the same length. those mixing elements you've shown earlier all seem to rely on a concept of "bending it around a lot to slow down part of the fluid while letting another part through faster" which is what introduces way more vortices than just a centered obstruction (especially a round one, because coanda effect and all that).
You could solder with silver-solder. There are different solders with higher melting temperatures that usually are melted with a torch. You could also try to make the wires sharper to make the flow better.
I’m curious what would happen if you tapped the interior of the nozzle. If it would jam or if the increased surface area would help flow, or if the threads would create the turbulence needed for mixing?
