@@LostInTech3D still.... the implications for printing larger are that whole part cooling is something that will need to be developed, not just the topmost layer. I dont know of any current tech that cools down middle layers
I have to thank you for the optimal layer height part at the end. I have noticed inconsistent line width on my parts and couldn't nail it down. I used that calculation to find the proper line width and it completely fixed the problem 🙏
Great video. It's amazing how far you were able to push the limit, especially with the entry-level hot end. I recently designed a fast-printing 'vase mode' modular drawer system that works best with a 0.8mm line width. I found that printing with a 0.4mm nozzle produces much sharper detail in the corners (important for this model) than using a 0.8mm nozzle, even at the same line width. When I shared this on thingiverse, I was surprised by how many comments there were from people who didn't realise it was possible to print a line width much wider than the nozzle size.
Yeah - in the part 1 I did a survey and even when prompted most people didn't seem to think you can go much over the nozzle width, I guess the cat is out of the bag now, maybe more vase mode designs will appear (I'm interested to see yours btw!)
For some reason, I can't seem to post a link here, but the design is on the top of the thingiverse landing page at the moment, so easy to find: 'Fast-Printing Modular Drawer System (Vase Mode)'
On my drive home tonight, I was hoping one of my channels would have an ASMR video with charts and some thermochromic filament being laid down in lines fatter than my shoe laces. I wasn't disappointed. Well done sir.
This video is amazing and deserves way more views. There are actually a lot of applications where printing super wide lines like that could make sense. Especially if the filament extruding heats the layer below it well enough to have good adhesion, this has lots of fun written all over it.
If extruder slippage is causing the under extrusion that implies that hot end pressure is getting too high. Maybe increasing the temperature would make the filament runner and reduce the pressure on the extruder. That would have the effect of introducing even more heat into the part which might cause other problems. Good video. Thanks!
@@LostInTech3D Mhh maybe it has something to do with mass correction. If you calibrate your printer for 0815 printers and prints you can use some kind of flowrate ofset for the filament. However if the filament is off by 1-2% this gap gap will get bigger up the line width, but maybe just one handicap up the way. Anyway very interesting for me as a newby on printing. Kinda like those test because i will mostly print stuff to use at the end.
That color changing filament is amazing. When the heat can propagate down not just to the previous layer but 2 and 3 layers deep, do you think that it significantly helps with the layer adhesion? Perhaps a collaboration with CNC kitchen to test this.
The most important bit for reproducing this is really @16:00 here. Right in there Lost In Tech points out that a good starting point is Height = (2/5)*Width. When increasing extrusion width too much, I've always gotten "blotchy" prints like what you've shown. Thanks for burning through some plastic waste on our behalf :) I just gave up when I got those results. You pushed through!
This is an amazing video. I watch several different well known TH-camrs and yes have learned quite a bit over the last year I have been in this but this video really takes the cake. Not even joking this is like the Nobel Prize Video for 3d printing. I have watched the videos, had some vague level of understanding on the topic maybe but after watching this video twice I finally feel like I actually have some level of control, a level of confidence about how to make the printer print like I want it to. Thank you!
I don't know how many times I've watched this video. There are so many insights from the heat diffusion between layers to the H/W ratio to the critical nature of flow rate. It seems like a gimmick, but turns out to be one of the most important printing videos ever.
I’m pretty new so most of what you said I don’t get but I am at a loss for words about your expertise and how well you do explain things. Thank you for what you do.
My guess is the loss comes from the over-height you get above your final line level. basically, it's like having a stack of thick baloney/ham slices, and bending them upwards. it's not that the slices become less wide, but each slice's width moves upward over the rest of the lines. It's pretty much the issue you're solving when choosing the layer height, and you're having with too little of a layer height. Such a cool little study!
My takeaway from this video is that there is still a whole lot more to learn about the dynamics of filament extrusion and a lot of opportunities for individuals to make a significant contribution with their experiments. Well done on this, I've not seen anyone else look at extreme extrusion like this - I think I recall Stephan doing a video showing extrusion at twice nozzle diameter in one video and asking if its worth getting larger nozzles but I think this was just something he noticed while looking at something else and he didn't seem to follow it up (unless I missed that video).
I love this video! I figured this out myself a couple years ago and glad you have laid out the facts. There are two ways to reduce print time: reduce the printed volume and print that volume faster. I print fat lines slow even with a small nozzle and knock my print times way down. If you know the max useful print head speed of your printer, you can cut down print time by making thicker extrusions. The sweet spot seems to be a a layer that is 2 to 3 times as wide as it is high. I rarely print with lines below 1 mm wide anymore.
Good to see someone discovering my own print philosophy about flow rate I can tell you exactly where the extra width is going. If you watch it extrude, especially wider than the outer diameter of the nozzle tip, you'll notice that the plastic begins to flow upwards around the nozzle. So the same amount of plastic is going out, it's just going up as well, not to the side like you would want it to.
This is an amazing video , with superb results. I will reuse it. You say that nobody talks about flowrate, then look for the voron community and the speedboat benchy challenge, where folks print benchys (regular size) in less than 5 minutes. They max out regular nozzles and volcano too, so they have alternatives. They address the flowrate issues, as well as cooling issues. Some made air jet cooling, from a compressor, others used large industrial fans with a shroud to send a blanket of compressed air on everything below the print head, effectively cooling everything fast. Obviously you'd run in some limitations in the heat transfer capacity of the plastic with thick beads, but I bet you could see an improvement. Also cht nozzles beats volcano at regular size. CNC kitchen tested it recently and even found a way to DIY an equivalent. Look for : -speedboat benchy challenge -cht nozzles -voron printers -hevort printer (this guy made some of the insane cooling solutions for insane flow rates, over 50mm cube per second). Cheers!
We talk about all this stuff all the time on discord! Even the air pump cooling :) I just don't have the time to untangle it into sensible videos because it's all a bit crazy 😂
@@LostInTech3D After writing it, i figured that considering what you did, there was actually no chances that you didn't hear about all of this :) Sure, crazy stuff, but really interesting, i sure would watch more videos about it :)
14:55 That was sick!!! Reminds me of the hours-long microscope-sessions I used to do (cancer cells), hoping to observe something new... Only the patient get rewarded!
Crazy! I never knew you could cleanly extrude anything significantly above the nozzle diameter, much less several times over. I think you could use this to greatly speed up 100% infill, printing the outer walls with the natural nozzle size but then doing infill with double or triple diameter. Cura can do infill layers thicker than wall layers, and can use a different extruder with larger nozzle for infill. So you could probably slice it like that, and then post-process the gcode to make it use a single extruder for everything. You'd also need to have relative extrusion enabled, since the commands generated for separate extruders would have different absolute positions.
Fun things. As it happens I printed a ~2.3mm TPU "string" the other day from a 0.35mm nozzle using custom G-code. I needed this as a crucial repair for our caravan before a weekend trip. I took a _lot_ of experiments, and extremely slow speeds to succeed. My plan was for a "normal" print first, but the TPU was too soft to be able to print such a slender print. But doing it in a single extrusion move, slowly, worked beatifully! Greater strength is a plus as well for a structural print like that.
What I took from this video is, to be able to print PET not PETG well and reliably is a shorter area for the filament to be melted in, basically making every bit of the filament that goes into the melting point need to be pushed out of that area faster because PET likes to crystallize the longer it is I contact with heat hot enough to melt it. Thank you so much for giving me this thought process! Just a theory now, but it does follow the science around PET and is the reason PETG is even a thing in 3D printing.
I'm trying to get consistently good prints out of PET (I have gotten exactly two prints, an AR lower and a grip, out of half a roll, with a lot of hard jams). If you figure out a flow rate range that works, especially with retraction, let me know!
@@paulstrealer5414 it varies some with the geometry of your part. As Lost in Tech points out, the thicker your your walls, the harder it is to dissipate heat. I routinely print at 5 mm^3/s and 265 C on my Prusa MK3. If I notice any crystallization I increase the part cooling. You absolutely have to dry your PET to get a good print. These days my hard jams with PET mostly from not gripping the filament tightly enough (it's slippery!). I use 2 mm retraction with a direct drive extruder, but I think that's more by chance than by tuning.
interesting, layer height of 0.16 has always been recommended for 0.4 by all slicers for "quality" prints. Since your last video about vase mode i changed my widths to 0.45mm and then had to bump my print height to 0.18mm to get clean walls I wasn't sure why but i was able to drop infill to 8% from 15% without losing rigidity using my calibrated hands, so I wasn't complaining.
@@LostInTech3D So... 0.25mm nozzle/width -> 0.1mm layer height? I usually go for 0.08mm but I guess those "rules/estimates" probably don't apply all that much to small nozzles/widths.
I'm not sure I'm willing to commit to the idea that the rule works for all nozzle widths, but hey you never know!! That layer height does look about right for a 0.25 nozzle!
Just throwing my two cents in, when I print in Vase mode I use a cheap chinese 0.8mm nozzle on a mostly stock prusa mini (other than a bondtech extruder upgrade). I print a 0.6mm layer height with 2.4mm extrusion width and it prints beautifully at 2.35-2.4 measured layer width. My one trick is limiting the flow rate to 12mm³/s and the speeds sort themselves out. Love your content!
Coming back to this video a couple years later. I think it would be very interesting to see you do a video on purposeful overhang drooping behavior that is common to see in 3d printed lamps and textured prints especially in the context of how different layer thicknesses and heights impact this behavior. And what about doubling up on wall stacks etc? It’s super aesthetic and I’d love to learn from the best 💪
super cool shots...i print a lot in vase mode with my sidewinder and the 1mm nozlzle... with silk on some geometrics i had issues, looking to me like heat creep with pla plus that didn´t happen.... thats explaining it really good... the flex amount in silk curls/blobs up again when it gets heated from the above layer and kind a wrinkles...think thats was happend to my prints, i goes...especially at small vases
@@LostInTech3D Haha great. The only problem I still got is stringing but after fumbling around with retraction wiping and alike I am pritty sure it is because my PETG is wet. I should probably store it outside, we got - 5°C or so. I will probably get a filament heater and or storage
I am running a 0.4 nozzle and compared my values H=0.3 W=0.7. It seems close to your H=2/5*W formula. Before I had bad layers, now they are quite good.
I believe that nozzle diameter is a constraint also, because force required to squeeze a viscous fluid through an aperture at a given flow rate is inversely proportional to aperture diameter. The effect becomes noticeable as early as 0.4mm nozzles, which don't seem to achieve the same results (proportionally) as a 0.8mm nozzle. I have found that I need to use a 0.8mm nozzle to achieve the full flow rate of the hot end. There is probably a crossover point somewhere between 0.6mm to 0.8mm nozzles where the limiting factor switches from being the nozzle diameter to the melt zone characteristics.
yes I agree - and my data did support this in the graphs, if I had more sensitive equipment (a micrometer screw gauge for example) I would probably have seen that more in the lower nozzle sizes. A 0.1mm resolution isn't really enough to make assertions about this, perhaps I will pencil that in for the future.
I wish more stuff was rated in flow. Pretty much every filament had recommended print speeds, but even those are kind of irrelevant depending on your slicer. The thing I finally learned in Cura was to show all the speeds on the settings so I could set them the same. I was having issues where it was printing inside walls way faster than my hot end could flow. Now I just set it all the same and target common flow rates. I have an old Flashforge Creator clone with a mk10 hot end, but it's hard finding data out there for flow rates. Almost everything online for testing is based on the E3D style hot ends. With my all metal upgrade though I can reliably print PETG at 12-15 mm3/s, so I've been reasonably happy.
I realize I'm getting to view this after awhile, but blame it on the YT algo if you ask me. What sort of data are you looking for from your "test slicer"? This ( Thermochroic filament )is fascinating to me, and one of the reasons I tried to argue ( with myself ) that in order to properly print with ABS, I need to immediately hit it with the fan to cause the latest filament printed to cool after attaching the previous layer. I have only managed to print with ABS using an ooze shield about 4mm from the model. I have tried building a box, but that doesn't work for me. I am using a Creality 10S, 300mm x 300mm x 400mm, direct drive, BLTouch, other than that pretty much stock. I have PETG, ABS, ASA, PLA, TPU, and NInjaFlex filaments.
Nice use of color changing filament! That looks great. I have some that I might break out for fan duct testing because that looks a lot better than a 100x100 pixel thermal camera image. I'll see about using your script to benchmark the 1.8mm CHT. Nice work, looking forward to whats next (mr shakey?)
Flow rate varies based on what plastic you're melting (PLA melts far easier than PETG, so it gets a higher flow rate). Also, temp affects it, higher temp, higher flow rate. So all printers have a range of flow rates (Prusa specifies approximate numbers, not sure about Crealities.)
I think the overextrusion you see for some prints is related as you mentioned to the fact, that the extruder is calibrated for the force of a 0.4mm nozzle. However, this just means that the extruder is bad in the first place, as it already slips at nominal speed. It slips less on low flow rates with large nozzles than calibrated for. I'd like to see if a quality extruder, e.g. Bondtech would exhibit the same effect. For the underextrusion at higher widths I can imagine two effects. First, the extruded material might bulge upwards when it's wider than the flat part of the nozzle reducing the cross section. Second, it might already have cooled down so the extruder cannot make the extrusion much wider (trying to deform rigid material). Also all hotends tend to underextrude a little when approaching their maximum flow rate. However with 10mm3/s on a volcano you should have some headroom. Another remark: you can print pretty fast with Prusa Slicer if you set all speeds to your travel speed and just limit the print speed via the volumetric flow of the filament. That way you never have to worry again if you overload your printer. Changes to layer heights and line widths will change your print speed automatically to arrive at your desired flow rate. To increase your flow rate significantly on standard hotends look into Bondtech CHT nozzles. They perform better than a volcano hotend with a non CHT nozzle.
It seems that most of those color change filaments change color somewhere roughly between room temperature and life threatening fever temperature. Would be nice to have one that changes color within a few degrees of turning solid or maybe heat deflection temperature. This would not only be really interesting, but help with tuning part cooling.
Here's an idea the thick layers seem to not be able to cool fast enough before the next layer extrudes which is why it doesn't look the best. so what if we cool down the layers with a fan? Maybe just one that stands away from the printer or a hotend mount that uses a computer fan?
Yes - and it works too, I used that trick to make the "wave" pattern more obvious when filming. It would be worth testing on the thicker filament, for sure!
I make hollow statues and looking to print a thicker layer for strength and speed of print. Would love to see any special options for something like that? Thanks again for the vid! 😀
After watching this video as well as CNC Kitchen's line width video, I am starting to wonder if measurement of extrusion force needs to become part of the discussion. The bed leveling springs have a finite stiffness, as does the bed and the machine frame. At these extreme widths, the pressure applied to the bed and in reverse to the extruder, even if the extruder does not skip, must be large enough to deform the bed springs at least. I wonder if the linearity in line width underdelivery is really just measuring the spring rate of the bed leveling and machine stiffness, and the wander is due to the variability of stiffness at different parts of the build plate.
I wonder if the line with is because of the surface tension of the plastic. When your printing wider than the the flat part of the nozzle it looks like the plastic is "springing" up and trying to reduce surface area, aka become more circular. You might be able to see this by checking if the layer heights.
If I understood CNC Kitchens verdict properly then using a Bondtech CHT Nozzle is a viable solution to increasing your stock Hotends Melting capability to Volcano like levels. Maybe worth a consideration? You could technically even combine that concept with the Volcano adapter CNC Kitchen sells in their store allowing the use of a CHT Nozzle in a Volcano Hotend.
The bottleneck becomes the ability to cool the printed part. Thickening the extrusion increases the volume much faster than it increases the surface area where cooling can happen. So at some point the material's ability to conduct heat to the cool surface, combined with the total size of the model, will be a limit on the useful extrusion rate.
Any chance you could run this test using the new Sunlu Meta PLA, it's capable of much higher flow rates Vs other filaments and has a huge temperature range, starting much lower than usual filaments.
I've always tought line width wasn't a parameter meant to produce such an output, but the feedback you give to the slicer regarding how much material is squeezed under the noozle and spreads around and sideways accordibgly to the material fluidity, the flow rate, and the layer height, so the slicer knows how much to spread appart parallel lines as a baseline on top of wich to match the actual parameters like, overlaping, for there's no other way for the slicer to know the outcome. At least I made some tests (although not so thoroughly as yours) with Simplify 3D a couple of years ago and they led to my conclussion, but never tested it in other slicers, and now I wonder how it does Cura to know the actual outcome for a given parameters set. I'm puzzled now with your video.
What slicer were you using initially before you wrote your own? Can this be accomplished simply by increasing the line width setting in Cura? Or are there other settings that need to be changed too? I don't want to create ridiculously thick lines like demonstrated here, I just want to make a vase with 1mm wall thickness using a 0.4mm nozzle, for instance.
Enjoyable, but how did non-vase mode turn out? I print a lot of stuff with 2mm walls, and while I have used .8 and 1.0mm nozzles, swapping around is a pain. if I could make 3 passes with a .4 or 2 passes with a .4 and get a reasonable wall, I'd be a happy camper. Most of the stuff is structural and goes inside things, so the lines/edging aren't a real big deal.
Have you given thought to machining a custom nozzle with a large flat width? The edges would need to be chambered with a sharp edge, I guess just like a normal nozzle.
So refreshing to watch someone caring about details and testing. Reminds me about myself :-) Your Video should be made available in German language. The community would love it. Would be glad to do the narrating. My voice and speech is widely commented as being very appreciated :-) So far I strictly film my very own ideas and still have plenty ideas to care about. Would be fun to do a translation. I have no monetary intrests, so you would publish the German for fun on your channel. The extraordinary content is worth it to spread out. Mike from Berlin Germany
@@LostInTech3D Hope You don't mind featuring you in my todays video. The first 1000 viewers within 3 houres expressed great pleasure learning about your experiments. So don't get surprised to get some German viewers added to your upcoming statistics :-)
With this, and high flow nozzles like the volcano, are people tuning their PID values with filament flowing through the nozzle at speed? It seems like that would make a big difference.
It's an interesting idea but I don't know if it would have much effect, based on octoprint thermistor graphs. Of course what's going on in the heater block itself is unknown, but the graphs are the same data PID uses. I think (and have heard this from many sources) that the constraining factor is the filament's thermal conductivity.
Have you tried Orca Slicer? It has a flow rate calibration built in. It has different flow rates built in to one model. It has a ton of other calibrations too. If you don't like it and think I should switch, why? I just got my first 3D printer a month ago (BL P1S) and have only used Orca and Bambu slicers. I would love to know if you think there's something better out there!
This makes me wonder how big of a nozzle bore you can go. Can I take an old nozzle and ream it out to 1.75? Do I then have to step up to thicker filament? Or can you just pass the stuff straight through and have it heat up properly. There'd be theoretically no pressure in the hot end which in my limited print experience, seems like you do kind of need some. Maybe you don't idk. Anyways I love these experimental videos pushing boundaries that people get stuck on. I don't have time, money, or enough knowledge to test out a lot of my questions. This was one of those, I see people saying you can't do this or that just because that's the way it is. No one ever answers why or how, people just get stuck on the popular thought or something they heard someone else say. Question everything 😁
Hmm perhaps the use of a cooling duck that has 3 sharp point outlets(to increase pressure) ,1 for each side and 1 in the middle might cool the thicker layers more .🤔 it appears that the inside doesn't get contact with the cooling and is insulated keeping it hot for so long?
It's the thermal mass the longer hot end has. It's not that it's in it longer, it's because the thermal mass is bigger and won't cool as fast as small hot end nozzles.
The sound track of the thermochroic filament part made me feel like I'm watching some documentary about deep space on discovery channel. PS: Tell us some examples using that magic number you found at 16:25
I never liked printing in vase mode because things end up very flimsy. Now I know why. Duh! Flow is an interesting setting ni Cura. Most people thing 100% is the limit but it can go higher than 100%. That concept took a while to sink into my head.
Still gotta sort my printers out but... I have 3 variations of the Anet A8, a JG Aurora, a 24V one I can't remember the name of, a wooden Geetech pro B i3 and a Geetech E180 modified with a CR10 hotend. Would be happy to run tests for you once I'm set up again.
@@LostInTech3D Crazy thing is, took me a year to pay my mate of the first one (Patient Zero, it was the first printer and it's usually a bit sick, but keeps going) Bowden mod A8, printer 3 and 4 I bought but, 2, 5 6 & 7 were gifted to me. Looking to learn more about my machines so once they are all up and running again I'll happily run tests and send you data!
The thermacroic filament test was some of the most insightful I’ve seen in any 3d printing test
It's a shame only 34% of people watched to that point 🙄
@@LostInTech3D still.... the implications for printing larger are that whole part cooling is something that will need to be developed, not just the topmost layer. I dont know of any current tech that cools down middle layers
thermochromic (heat and colored in greek)
Yeahh it looks so nice, and perfect for education.
@@LostInTech3D Hey that's okay, I'm back for a rewatch anyways so eventually it'll balance out
I have to thank you for the optimal layer height part at the end. I have noticed inconsistent line width on my parts and couldn't nail it down. I used that calculation to find the proper line width and it completely fixed the problem 🙏
Glad it helped!
Great video. It's amazing how far you were able to push the limit, especially with the entry-level hot end. I recently designed a fast-printing 'vase mode' modular drawer system that works best with a 0.8mm line width. I found that printing with a 0.4mm nozzle produces much sharper detail in the corners (important for this model) than using a 0.8mm nozzle, even at the same line width. When I shared this on thingiverse, I was surprised by how many comments there were from people who didn't realise it was possible to print a line width much wider than the nozzle size.
Yeah - in the part 1 I did a survey and even when prompted most people didn't seem to think you can go much over the nozzle width, I guess the cat is out of the bag now, maybe more vase mode designs will appear (I'm interested to see yours btw!)
For some reason, I can't seem to post a link here, but the design is on the top of the thingiverse landing page at the moment, so easy to find: 'Fast-Printing Modular Drawer System (Vase Mode)'
yeah it does moderate links but I usually get them in my moderation queue. Thing #5210727?
That's the one
I'll be trying that, it looks exactly like what I need for my electronics "collection" - thanks! 👍
Such a great video! The heat propagating down to previous layers was an amazing demonstration.
I'm gonna release that as a short, to try to get it out to more people!
On my drive home tonight, I was hoping one of my channels would have an ASMR video with charts and some thermochromic filament being laid down in lines fatter than my shoe laces. I wasn't disappointed. Well done sir.
I've not been called ASMR before but I'll take it 👍🤣
This video is amazing and deserves way more views. There are actually a lot of applications where printing super wide lines like that could make sense. Especially if the filament extruding heats the layer below it well enough to have good adhesion, this has lots of fun written all over it.
If extruder slippage is causing the under extrusion that implies that hot end pressure is getting too high. Maybe increasing the temperature would make the filament runner and reduce the pressure on the extruder. That would have the effect of introducing even more heat into the part which might cause other problems. Good video. Thanks!
Yes - absolutely, this is why I invested the time to write the software, to make a basis for further tests including temperature.
@@LostInTech3D Mhh maybe it has something to do with mass correction. If you calibrate your printer for 0815 printers and prints you can use some kind of flowrate ofset for the filament. However if the filament is off by 1-2% this gap gap will get bigger up the line width, but maybe just one handicap up the way.
Anyway very interesting for me as a newby on printing. Kinda like those test because i will mostly print stuff to use at the end.
That color changing filament is amazing. When the heat can propagate down not just to the previous layer but 2 and 3 layers deep, do you think that it significantly helps with the layer adhesion?
Perhaps a collaboration with CNC kitchen to test this.
That's an interesting thought actually!
The most important bit for reproducing this is really @16:00 here. Right in there Lost In Tech points out that a good starting point is Height = (2/5)*Width. When increasing extrusion width too much, I've always gotten "blotchy" prints like what you've shown. Thanks for burning through some plastic waste on our behalf :) I just gave up when I got those results. You pushed through!
So for example with a 0,4mm nozzle I would want a 1mm layer width with 0,4mm height?🤔 Can that the 0,4mm nozzle handle still.
Height is 40% of your width @@nAcolz
I love these videos! Keep pushing the limits and documenting it for us.
Will do!
This is an amazing video. I watch several different well known TH-camrs and yes have learned quite a bit over the last year I have been in this but this video really takes the cake. Not even joking this is like the Nobel Prize Video for 3d printing. I have watched the videos, had some vague level of understanding on the topic maybe but after watching this video twice I finally feel like I actually have some level of control, a level of confidence about how to make the printer print like I want it to. Thank you!
Glad to hear, thanks! 👍
great stuff! I understand that people who make prosthetic limbs print the custom piece that fits the body in vase mode using big nozzles.
Do they? I didn't know that
Sounds like a perfect case for CHT nozzles.
Maybe 😁
I don't know how many times I've watched this video. There are so many insights from the heat diffusion between layers to the H/W ratio to the critical nature of flow rate. It seems like a gimmick, but turns out to be one of the most important printing videos ever.
looking back, I have to agree - a lot of this stuff I didn't really fully understand at the time either.
Vase mode 3 soon maybe :)
It's amazing the extruded filament spreads sideways so much without balling up on the nozzle.
yeah I totally didn't expect that!
I’m pretty new so most of what you said I don’t get but I am at a loss for words about your expertise and how well you do explain things. Thank you for what you do.
This video was so incredibly interesting, loved it!
My guess is the loss comes from the over-height you get above your final line level. basically, it's like having a stack of thick baloney/ham slices, and bending them upwards. it's not that the slices become less wide, but each slice's width moves upward over the rest of the lines. It's pretty much the issue you're solving when choosing the layer height, and you're having with too little of a layer height. Such a cool little study!
Anyone who says "you can't extrude thicker than the nozzle" has never used spray whipped cream.
A good analogy!
My takeaway from this video is that there is still a whole lot more to learn about the dynamics of filament extrusion and a lot of opportunities for individuals to make a significant contribution with their experiments. Well done on this, I've not seen anyone else look at extreme extrusion like this - I think I recall Stephan doing a video showing extrusion at twice nozzle diameter in one video and asking if its worth getting larger nozzles but I think this was just something he noticed while looking at something else and he didn't seem to follow it up (unless I missed that video).
I love this video! I figured this out myself a couple years ago and glad you have laid out the facts. There are two ways to reduce print time: reduce the printed volume and print that volume faster. I print fat lines slow even with a small nozzle and knock my print times way down. If you know the max useful print head speed of your printer, you can cut down print time by making thicker extrusions. The sweet spot seems to be a a layer that is 2 to 3 times as wide as it is high. I rarely print with lines below 1 mm wide anymore.
Good to see someone discovering my own print philosophy about flow rate
I can tell you exactly where the extra width is going. If you watch it extrude, especially wider than the outer diameter of the nozzle tip, you'll notice that the plastic begins to flow upwards around the nozzle. So the same amount of plastic is going out, it's just going up as well, not to the side like you would want it to.
This is an amazing video , with superb results. I will reuse it.
You say that nobody talks about flowrate, then look for the voron community and the speedboat benchy challenge, where folks print benchys (regular size) in less than 5 minutes.
They max out regular nozzles and volcano too, so they have alternatives.
They address the flowrate issues, as well as cooling issues.
Some made air jet cooling, from a compressor, others used large industrial fans with a shroud to send a blanket of compressed air on everything below the print head, effectively cooling everything fast.
Obviously you'd run in some limitations in the heat transfer capacity of the plastic with thick beads, but I bet you could see an improvement.
Also cht nozzles beats volcano at regular size. CNC kitchen tested it recently and even found a way to DIY an equivalent.
Look for :
-speedboat benchy challenge
-cht nozzles
-voron printers
-hevort printer (this guy made some of the insane cooling solutions for insane flow rates, over 50mm cube per second).
Cheers!
We talk about all this stuff all the time on discord! Even the air pump cooling :) I just don't have the time to untangle it into sensible videos because it's all a bit crazy 😂
@@LostInTech3D After writing it, i figured that considering what you did, there was actually no chances that you didn't hear about all of this :)
Sure, crazy stuff, but really interesting, i sure would watch more videos about it :)
14:55 That was sick!!! Reminds me of the hours-long microscope-sessions I used to do (cancer cells), hoping to observe something new... Only the patient get rewarded!
I thoroughly enjoy your method, curiosity and deadpan humor :D
thank you!
Crazy! I never knew you could cleanly extrude anything significantly above the nozzle diameter, much less several times over. I think you could use this to greatly speed up 100% infill, printing the outer walls with the natural nozzle size but then doing infill with double or triple diameter. Cura can do infill layers thicker than wall layers, and can use a different extruder with larger nozzle for infill. So you could probably slice it like that, and then post-process the gcode to make it use a single extruder for everything. You'd also need to have relative extrusion enabled, since the commands generated for separate extruders would have different absolute positions.
So excited for future videos on this!
Fun things. As it happens I printed a ~2.3mm TPU "string" the other day from a 0.35mm nozzle using custom G-code. I needed this as a crucial repair for our caravan before a weekend trip.
I took a _lot_ of experiments, and extremely slow speeds to succeed. My plan was for a "normal" print first, but the TPU was too soft to be able to print such a slender print. But doing it in a single extrusion move, slowly, worked beatifully! Greater strength is a plus as well for a structural print like that.
This video is awesome! Thanks for explaining/documenting all of this.
After a while it's hard to grasp the scale of the prints/extrusion lines until you pick them up. Very cool!
What I took from this video is, to be able to print PET not PETG well and reliably is a shorter area for the filament to be melted in, basically making every bit of the filament that goes into the melting point need to be pushed out of that area faster because PET likes to crystallize the longer it is I contact with heat hot enough to melt it. Thank you so much for giving me this thought process! Just a theory now, but it does follow the science around PET and is the reason PETG is even a thing in 3D printing.
I'm trying to get consistently good prints out of PET (I have gotten exactly two prints, an AR lower and a grip, out of half a roll, with a lot of hard jams). If you figure out a flow rate range that works, especially with retraction, let me know!
In my experience, it's not the time spent hot, but the time spent cooling that matters. You have to cool fast enough to lock in the glassy state.
@@paulstrealer5414 it varies some with the geometry of your part. As Lost in Tech points out, the thicker your your walls, the harder it is to dissipate heat. I routinely print at 5 mm^3/s and 265 C on my Prusa MK3. If I notice any crystallization I increase the part cooling. You absolutely have to dry your PET to get a good print.
These days my hard jams with PET mostly from not gripping the filament tightly enough (it's slippery!). I use 2 mm retraction with a direct drive extruder, but I think that's more by chance than by tuning.
interesting, layer height of 0.16 has always been recommended for 0.4 by all slicers for "quality" prints. Since your last video about vase mode i changed my widths to 0.45mm and then had to bump my print height to 0.18mm to get clean walls I wasn't sure why but i was able to drop infill to 8% from 15% without losing rigidity using my calibrated hands, so I wasn't complaining.
that checks out then doesn't it, because 2/5ths of 0.45 is 0.18! Nice! 👍
@@LostInTech3D So... 0.25mm nozzle/width -> 0.1mm layer height? I usually go for 0.08mm but I guess those "rules/estimates" probably don't apply all that much to small nozzles/widths.
I'm not sure I'm willing to commit to the idea that the rule works for all nozzle widths, but hey you never know!! That layer height does look about right for a 0.25 nozzle!
Taking flow rate to a new layer. Excellent work! Thankyou
Amazing video, you helped us solve a mystery when pushing vase mode.
Thank you!
I think that thermosomething filament deserves a Short of its own ;)
That was my plan actually...! 👍
Excellent work!
Thank you.
Just throwing my two cents in, when I print in Vase mode I use a cheap chinese 0.8mm nozzle on a mostly stock prusa mini (other than a bondtech extruder upgrade). I print a 0.6mm layer height with 2.4mm extrusion width and it prints beautifully at 2.35-2.4 measured layer width. My one trick is limiting the flow rate to 12mm³/s and the speeds sort themselves out. Love your content!
Coming back to this video a couple years later. I think it would be very interesting to see you do a video on purposeful overhang drooping behavior that is common to see in 3d printed lamps and textured prints especially in the context of how different layer thicknesses and heights impact this behavior. And what about doubling up on wall stacks etc? It’s super aesthetic and I’d love to learn from the best 💪
This is great, never seen a video that covers this.
super cool shots...i print a lot in vase mode with my sidewinder and the 1mm nozlzle... with silk on some geometrics i had issues, looking to me like heat creep with pla plus that didn´t happen.... thats explaining it really good... the flex amount in silk curls/blobs up again when it gets heated from the above layer and kind a wrinkles...think thats was happend to my prints, i goes...especially at small vases
This is great! Keep going !
Subscribed! Awesome and very interesting content! Thank you!
This is incredible
Awesome video, really helped me out.
Glad to hear it! More vase mode shenanigans are on the way 👀
@@LostInTech3D Haha great.
The only problem I still got is stringing but after fumbling around with retraction wiping and alike I am pritty sure it is because my PETG is wet. I should probably store it outside, we got - 5°C or so. I will probably get a filament heater and or storage
Subscribed! Always looking for the fastest way to print stuff!
Very much impressed. 👌
Vase mode is great. Adaptive layer height can give great overhangs!
Heat propagation was very cool!
Love it!
I am running a 0.4 nozzle and compared my values H=0.3 W=0.7. It seems close to your H=2/5*W formula. Before I had bad layers, now they are quite good.
I actually am as surprised as everyone else that the 2/5 rule appears to work for all sizes!
very very good video, thanks
I believe that nozzle diameter is a constraint also, because force required to squeeze a viscous fluid through an aperture at a given flow rate is inversely proportional to aperture diameter. The effect becomes noticeable as early as 0.4mm nozzles, which don't seem to achieve the same results (proportionally) as a 0.8mm nozzle. I have found that I need to use a 0.8mm nozzle to achieve the full flow rate of the hot end. There is probably a crossover point somewhere between 0.6mm to 0.8mm nozzles where the limiting factor switches from being the nozzle diameter to the melt zone characteristics.
yes I agree - and my data did support this in the graphs, if I had more sensitive equipment (a micrometer screw gauge for example) I would probably have seen that more in the lower nozzle sizes. A 0.1mm resolution isn't really enough to make assertions about this, perhaps I will pencil that in for the future.
This is brilliant 😀
I wish more stuff was rated in flow. Pretty much every filament had recommended print speeds, but even those are kind of irrelevant depending on your slicer. The thing I finally learned in Cura was to show all the speeds on the settings so I could set them the same. I was having issues where it was printing inside walls way faster than my hot end could flow. Now I just set it all the same and target common flow rates. I have an old Flashforge Creator clone with a mk10 hot end, but it's hard finding data out there for flow rates. Almost everything online for testing is based on the E3D style hot ends. With my all metal upgrade though I can reliably print PETG at 12-15 mm3/s, so I've been reasonably happy.
I realize I'm getting to view this after awhile, but blame it on the YT algo if you ask me. What sort of data are you looking for from your "test slicer"?
This ( Thermochroic filament )is fascinating to me, and one of the reasons I tried to argue ( with myself ) that in order to properly print with ABS, I need to immediately hit it with the fan to cause the latest filament printed to cool after attaching the previous layer. I have only managed to print with ABS using an ooze shield about 4mm from the model. I have tried building a box, but that doesn't work for me. I am using a Creality 10S, 300mm x 300mm x 400mm, direct drive, BLTouch, other than that pretty much stock. I have PETG, ABS, ASA, PLA, TPU, and NInjaFlex filaments.
Useful video
Glad to see Steve's name is sticking
I'm doing everything I can to make Steve a thing 🤣
Really good video again. Those lines are thicc
Nice use of color changing filament! That looks great. I have some that I might break out for fan duct testing because that looks a lot better than a 100x100 pixel thermal camera image.
I'll see about using your script to benchmark the 1.8mm CHT. Nice work, looking forward to whats next (mr shakey?)
Cheers for sharing, informative and entertaining at the same time great video .
The colour changing filament was impressive to watch.
Flow rate varies based on what plastic you're melting (PLA melts far easier than PETG, so it gets a higher flow rate). Also, temp affects it, higher temp, higher flow rate. So all printers have a range of flow rates (Prusa specifies approximate numbers, not sure about Crealities.)
I think the overextrusion you see for some prints is related as you mentioned to the fact, that the extruder is calibrated for the force of a 0.4mm nozzle. However, this just means that the extruder is bad in the first place, as it already slips at nominal speed. It slips less on low flow rates with large nozzles than calibrated for. I'd like to see if a quality extruder, e.g. Bondtech would exhibit the same effect.
For the underextrusion at higher widths I can imagine two effects. First, the extruded material might bulge upwards when it's wider than the flat part of the nozzle reducing the cross section. Second, it might already have cooled down so the extruder cannot make the extrusion much wider (trying to deform rigid material).
Also all hotends tend to underextrude a little when approaching their maximum flow rate. However with 10mm3/s on a volcano you should have some headroom.
Another remark: you can print pretty fast with Prusa Slicer if you set all speeds to your travel speed and just limit the print speed via the volumetric flow of the filament. That way you never have to worry again if you overload your printer. Changes to layer heights and line widths will change your print speed automatically to arrive at your desired flow rate.
To increase your flow rate significantly on standard hotends look into Bondtech CHT nozzles. They perform better than a volcano hotend with a non CHT nozzle.
It seems that most of those color change filaments change color somewhere roughly between room temperature and life threatening fever temperature.
Would be nice to have one that changes color within a few degrees of turning solid or maybe heat deflection temperature. This would not only be really interesting, but help with tuning part cooling.
The big CHT nozzles are the way to go.
May test it in an upcoming :)
@@LostInTech3D I've played a bit with the 1.4mm one, you can go through soooo much material, blink and the roll is empty :D
yeah that was happening a bit with the volcano, luckily it's old roll ends I used for this...that vase at the end is a literal paperweight!
Try the Bondtech CHT nozzles if you want to increase flowrate dramatically.
First time watcher of your channel, and a great video. My sidewinder with a 0.6 nozzle is itching to try this.
Welcome, and sorry for the temptation 😁👍
Here's an idea the thick layers seem to not be able to cool fast enough before the next layer extrudes which is why it doesn't look the best. so what if we cool down the layers with a fan?
Maybe just one that stands away from the printer or a hotend mount that uses a computer fan?
Yes - and it works too, I used that trick to make the "wave" pattern more obvious when filming. It would be worth testing on the thicker filament, for sure!
I make hollow statues and looking to print a thicker layer for strength and speed of print. Would love to see any special options for something like that? Thanks again for the vid! 😀
After watching this video as well as CNC Kitchen's line width video, I am starting to wonder if measurement of extrusion force needs to become part of the discussion. The bed leveling springs have a finite stiffness, as does the bed and the machine frame. At these extreme widths, the pressure applied to the bed and in reverse to the extruder, even if the extruder does not skip, must be large enough to deform the bed springs at least.
I wonder if the linearity in line width underdelivery is really just measuring the spring rate of the bed leveling and machine stiffness, and the wander is due to the variability of stiffness at different parts of the build plate.
Great video. 👍
Thanks 👍
Can you share some cura profiles for the .8 mm test? I would love to try this on some of my printers
I wonder if the line with is because of the surface tension of the plastic. When your printing wider than the the flat part of the nozzle it looks like the plastic is "springing" up and trying to reduce surface area, aka become more circular. You might be able to see this by checking if the layer heights.
If I understood CNC Kitchens verdict properly then using a Bondtech CHT Nozzle is a viable solution to increasing your stock Hotends Melting capability to Volcano like levels. Maybe worth a consideration? You could technically even combine that concept with the Volcano adapter CNC Kitchen sells in their store allowing the use of a CHT Nozzle in a Volcano Hotend.
Yeah I need to at some point use a bondtech nozzle, but CNC kitchen has done a lot of work here as you say, possibly not much else to learn from it
The bottleneck becomes the ability to cool the printed part. Thickening the extrusion increases the volume much faster than it increases the surface area where cooling can happen. So at some point the material's ability to conduct heat to the cool surface, combined with the total size of the model, will be a limit on the useful extrusion rate.
Any chance you could run this test using the new Sunlu Meta PLA, it's capable of much higher flow rates Vs other filaments and has a huge temperature range, starting much lower than usual filaments.
Oh yes. That has to happen.
@@LostInTech3D awesome 😊 Looking forward to seeing it! Thanks 😊
good content!
Thank you 👍
2:08 you said 0.8mm but the video shows a 0.6mm nozzle?
Yes, more please.
Noted 🙂
I've always tought line width wasn't a parameter meant to produce such an output, but the feedback you give to the slicer regarding how much material is squeezed under the noozle and spreads around and sideways accordibgly to the material fluidity, the flow rate, and the layer height, so the slicer knows how much to spread appart parallel lines as a baseline on top of wich to match the actual parameters like, overlaping, for there's no other way for the slicer to know the outcome. At least I made some tests (although not so thoroughly as yours) with Simplify 3D a couple of years ago and they led to my conclussion, but never tested it in other slicers, and now I wonder how it does Cura to know the actual outcome for a given parameters set. I'm puzzled now with your video.
I have another video called "flow math" that may explain this
im expecting this too be goooooood
I hope you weren't disappointed 🤣👍
What slicer were you using initially before you wrote your own? Can this be accomplished simply by increasing the line width setting in Cura? Or are there other settings that need to be changed too?
I don't want to create ridiculously thick lines like demonstrated here, I just want to make a vase with 1mm wall thickness using a 0.4mm nozzle, for instance.
You can use any slicer, you might need to set the nozzle to 1mm, to make it accept the numbers.
YES
YES
Enjoyable, but how did non-vase mode turn out? I print a lot of stuff with 2mm walls, and while I have used .8 and 1.0mm nozzles, swapping around is a pain. if I could make 3 passes with a .4 or 2 passes with a .4 and get a reasonable wall, I'd be a happy camper. Most of the stuff is structural and goes inside things, so the lines/edging aren't a real big deal.
Have you given thought to machining a custom nozzle with a large flat width? The edges would need to be chambered with a sharp edge, I guess just like a normal nozzle.
It's an interesting idea, not sure I'd be able to, I'm not exactly good with metalwork 😉
So refreshing to watch someone caring about details and testing. Reminds me about myself :-) Your Video should be made available in German language. The community would love it.
Would be glad to do the narrating.
My voice and speech is widely commented as being very appreciated :-)
So far I strictly film my very own ideas and still have plenty ideas to care about. Would be fun to do a translation.
I have no monetary intrests, so you would publish the German for fun on your channel. The extraordinary content is worth it to spread out.
Mike from Berlin Germany
I watched some of your videos and wish I understood more German! We need to collab some day, we are almost twin channels! :)
@@LostInTech3D Hope You don't mind featuring you in my todays video. The first 1000 viewers within 3 houres expressed great pleasure learning about your experiments. So don't get surprised to get some German viewers added to your upcoming statistics :-)
A theory seen from one of the sites I was reading is about being able to achieve much Steeper overhangs with wider lines but not taller.
Makes sense mathematically at least...wider lines should mean larger overhangs... you would think
With this, and high flow nozzles like the volcano, are people tuning their PID values with filament flowing through the nozzle at speed? It seems like that would make a big difference.
It's an interesting idea but I don't know if it would have much effect, based on octoprint thermistor graphs. Of course what's going on in the heater block itself is unknown, but the graphs are the same data PID uses.
I think (and have heard this from many sources) that the constraining factor is the filament's thermal conductivity.
Have you tried Orca Slicer? It has a flow rate calibration built in. It has different flow rates built in to one model. It has a ton of other calibrations too. If you don't like it and think I should switch, why? I just got my first 3D printer a month ago (BL P1S) and have only used Orca and Bambu slicers. I would love to know if you think there's something better out there!
This makes me wonder how big of a nozzle bore you can go. Can I take an old nozzle and ream it out to 1.75? Do I then have to step up to thicker filament? Or can you just pass the stuff straight through and have it heat up properly. There'd be theoretically no pressure in the hot end which in my limited print experience, seems like you do kind of need some. Maybe you don't idk. Anyways I love these experimental videos pushing boundaries that people get stuck on. I don't have time, money, or enough knowledge to test out a lot of my questions. This was one of those, I see people saying you can't do this or that just because that's the way it is. No one ever answers why or how, people just get stuck on the popular thought or something they heard someone else say. Question everything 😁
Hmm perhaps the use of a cooling duck that has 3 sharp point outlets(to increase pressure) ,1 for each side and 1 in the middle might cool the thicker layers more .🤔 it appears that the inside doesn't get contact with the cooling and is insulated keeping it hot for so long?
It's the thermal mass the longer hot end has. It's not that it's in it longer, it's because the thermal mass is bigger and won't cool as fast as small hot end nozzles.
The sound track of the thermochroic filament part made me feel like I'm watching some documentary about deep space on discovery channel.
PS: Tell us some examples using that magic number you found at 16:25
This is cool
I never liked printing in vase mode because things end up very flimsy. Now I know why. Duh! Flow is an interesting setting ni Cura. Most people thing 100% is the limit but it can go higher than 100%. That concept took a while to sink into my head.
Is there any chance you could reupload your flow test generator thing? The link appears to be broken :(
Still gotta sort my printers out but... I have 3 variations of the Anet A8, a JG Aurora, a 24V one I can't remember the name of, a wooden Geetech pro B i3 and a Geetech E180 modified with a CR10 hotend.
Would be happy to run tests for you once I'm set up again.
You have more than I do...I think! 😂
@@LostInTech3D Crazy thing is, took me a year to pay my mate of the first one (Patient Zero, it was the first printer and it's usually a bit sick, but keeps going) Bowden mod A8, printer 3 and 4 I bought but, 2, 5 6 & 7 were gifted to me.
Looking to learn more about my machines so once they are all up and running again I'll happily run tests and send you data!
The links to your website are broken
The captions for this video sure are... *Interesting.*
they are? 🤔 I usually turn off the auto ones, but I see they are still on
Weighing the samples in cnc kitchen style migh give you some nice data on underextrutio with increased width
My mind is just blown