Is fast the new big? What do you think about this new speed-printing trend? Don't forget to like & subscribe and share this video on Facebook, Reddit, Twitter, and other platforms!
stephan, i have the exact same wera toque wrenches here. they are awesome. you have made a pretty excelllent creep test in my opinionwith the dials, and also nice to hear about annealing the cf nylon compared to for pla. thing is... cf nylon is still too expensive for most of us. so maybe something else like that greentec pro cf is a bit cheaper? since we have to anneal them whichever option is chosen. at least the greentec cf pro is more environmentally friendly. but is it as strong as the claims? also have you gotten a sample of the brand new nonoilen (=pla and phb mix). because that is also claiming to be a strong material too, whilst also entirely environmentally safe and biodegradable. at least for that same sort of prices this is a genuine extra feature makes it feel a better 'value' (not for us, but for nature). so is cf nylon really worth it (for fdm purpose) by comparison? this is a genuine question i want to know the answer for! and doesnt seem like anybody else can convincingly answer it as for speedboat race, main casualty is the inter layer adhesion. so finding out about that is... important to put it into context. although maybe if you do anneal afterwards a speedbenchy then perhaps it doesnt matter? although annealingstep itself adds on a lot more of extra time. hehe [added] lol sorry i forgot to say something really important: there is this guy i know, and after finishing the print (in his same enclosed chamber) he then ramps up the temperature several times. for example 3 seperate times, starting about 60c and going up higher each time. ending at the highest annealing temp. maybe ending 100c or more. just whemever it makes the sense to stop cycling. this ramping type of work hardening / annealing process is better, in particular for these enhanced 'high temp' pla types. so this is also the recommended way for those greentec pro cf, and the nonoilen, and the esun pla+. any enhanced pla. just something you needed to know about, if you are struggling for annealing pla. doing the ramping program is essential. it keeps a partial crystalinity and then the object does not melt and loose its shape. which was in your previous experiences for annealing pla. so please try this method in future. many thanks
@@AlexusMaximusDE That thing is a BEAST and the guy is really cool, always interacts with the viewers too. It's always fascinating to see what the heck is the next physical limit he's gonna hit with that machine. At this point I'm thinking the plastic itself is close to being a limiting factor.
A "5 minutes 3D Benchy" video might be more clickable but doesn't really teach anyone something. My goal is to fail, learn, repeat and eventually succeed. That's how the ones how want to learn something benefit the most!
Very nifty test, but the type of creep on the printer bed is constant force. This creep test is constant strain. The bolt tension translates to a fixed compression. As the material experiences creep the force on the material will approach 0. An example experiment for a constant force test might be to hang a weight off of a sample instead.
Hi Stefan, I really appreciate the rigorous approach you take to testing different aspects of 3D printing. It must be very time consuming to test and record each variable and I am grateful to you for sharing your results with us.
Not sure how you arrived at 1. At those speeds, you need every bit of cooling you can get, even for abs. It's pretty much printed faster than it can warp. An enclosure would only hurt.
Usually it is very important for ABS to be printed in a warm enclosure (that also keeps the nasty smell in a little), though for speed printing, that's exactly the wrong thing because you want to cool the material down as fast as possible. The problem with my basement was rather that the heating system wasn't properly able to cope with the lower temperatures.
hmmm... this makes me wonder if there should be a 'feed forward' control system for the nozzle temperature that takes the current extrusion rate into account. Basically preemptively increase the hotend power if there is a fast extrusion move in the near future.
This is in fact called feed forward, or "2 dof control" and shouldn't be hard to implement because you only need to add the feedforward term (theoretical power for melting and convection) to the feedback part (PID control for the error between set and actual temperature). If you only use feedback control (the way it's currently done), the true temperature will always "lag behind" the reference to some degree.
I prefer bang bang for the hot end control over PID. It doesn't work in Marlin though because it doesn't sample often enough. In my own firmware I sample very quickly so that close to the set point noise gives a small range of PWM over a fraction of a degree.
@@nophead Haha wow it's like talking to a celebrity, thanks for everything you've done for the community since I got into this hobby 10 years ago :-) I think bang bang control is actually a great choice. PID just tends to work with most simple systems (and is more forgiving if you're slower than your system's dynamics) although it's certainly not the perfect controller for this type of system. What I'd like to see also is a controller for the nozzle temperature of the plastics instead of the temperature of the NTC. With a thermal model of the hotend and plastics and an observer this should yield near-perfect results. Although it's probably not really worth the effort.
That's an interesting idea. Another way to think about it is the filament has a cooling effect on the hotend that is proportional to feed rate. So it only makes sense to counteract this cooling with equal and opposite heating from the hotend. Normally this is handled lazily by a separate hotend temperature control PID loop but since heating happens slow and changes in extrusion rate happen fast, there is always some delay in the heater cartridges response. Also it takes some time for the thermistor to register that the hotend is cooling down before it can react.
For your creep testing, how high of a temp are you testing too? In most vorons enclosed the ambients can sit around 50c when printing ABS. Noticing the layer shifting? Are you tweaking accel/decel values as well as IS For replacing the collet, ya that part kinda sucks, i popped mine out too originally, best recommendation is install it BEFORE screwing both halves together so its sandwiched in place instead of pressing it in, seems to grab it a bit tighter. (or go with the direct feed toolhead cus 0.1 is out now )
"Increased" temperatures will, for the first round, be only 40°C to stay far away from any glass transition temperatures. Believe it or not but that already drastically changes the results!
I'm a 3D printing neophyte compared to all the testing you perform, but the two things that have solved layer shifts for me have been putting loctite on the motor pulley grub screws, and increasing belt tension. I'm enjoying how you are sharing your progress on this challenge.
This feels like your boss fight! All the research you've put in for years culminating in the speed benchy challenge. Good luck and thanks for sharing your findings!
Hi Stefan, your scientific approach to 3D printing is just on another level. Thank you for sharing with us hours of testing and valuable pieces of information.
Good point - especially what and who they compare. Not everyone has a setup to go to 10 minutes so something slower might give them a good baseline for their setup.
I love the artful display of the failed benchies - definitely highlighting the issues is a fantastic way to help us non-professionals feel less bad about our mangled prints :D
Oh boy, I've been looking forward to the creep testing for years! I know it's got to be a pretty massive amount of work, so I really appreciate you doing it for us! I also LOVE the speed printing trend!
Also, yes, silicone socks are great and pretty much essential imo. My temperature goes all over the place without it on my x1. Just bought another one. Thanks for the reminder and great video as always. Resubscribed after 2 years.
Beyond high volume extruding, I think high speed kinematics would be a cool series... Basically increasing print speeds and trying to determine what needs to be improved
At around 12:02 and and 13:01 there is vapour coming out from the nozzle if i see it correctly. Due to very high latent heat of vaporasition of water, i suspect it might increase your heating load dramatically. Amazing work btw thank you :)
Fantastic video! These Voron videos are great. If the Prusa XL wasn't around the corner I would build a 2.4 yesterday, it seems so great. Thank you for taking us on this journey of insane speed.
Accuracy, print stability and consistency - not speed is what I am trying to achieve. After I master those - I will probably work on some speed improvements. Thanks for showing us your methodology - you teach me (us) a lot when you do that - so we can do it for our setups. Again, Thank You!
I think the most interesting thing about this speed printing challenge is that it looks like we've reached a point where the mechanical structure of the system is able to better handle the speeds than the other components. CoreXY and a lot of the clever implementations of it have really brought the general structure of 3D printers to a new level.
Steffan - The Mosquito hotend is amazing! I've been using it for around 7-8 months now. I have some of my own designs exclusively up on thingiverse, search "blood drive ender 3", my mod is purposed around a linear rail converted Ender 3. I've been able to achieve 25mm/3 flow rate with PETG, PLA has more difficulties. Somethings I'd recommend - You should be using boron nitride paste that you can get from Slice Engineering. Infact I think this would be a good test to conduct on its own for your channel. The thermistor and heater core should be coated with it for maximum heat transfer. Also, the threads in the Mosquito hotblock as well as the threads on your nozzle should be coated as well. It is significant, I believe. Additionally, nickel plated brass nozzles will further your heat transferrence even more.
I enjoyed this video. Thanks for all the analysis and lessons. One note, for your extrusion rate graphs, the legend says NF for all plots, even those that are for the high flow heat break.
Awesome!! Comparing Polymaker materials TDS, PA12-CF has almost double the melt index of ABS, and becasue of the fibers, it will hold its shape up to very high temperatures. Maybe a filled PA is a good candidate for the fast benchy :)
@@User-nc2lf Thank you for your interest. (^‿^) I researched creating metamaterials for high frequency components using FDM printing. Even though I finished my PhD I continued working on that topic, because it is so interesting. Let alone, I love the topic of 3D printing. Not a day goes by without me working or creating things with the 3D printer. This alone made it all worth it, even though it was very difficult.
Very interesting video, thanks as you answered my question in the video on why the filament starts to skip. 👍 I’m a new to 3D printing, I spent some time in the oil field as a machinist and can see it’s no different lol it all takes time practice and patients. Thanks again
I use socks for exactly that reason. My MK2.5s won't keep heat steady without one. I do tend to attach it with a few dabs of high temp silicone. Keeps it from falling off unintentionally.
Tiny nit pick - the flow graphs such as at 6:23 list all of the traces as "NF", I guess Normal Flow, and none as "HF", so that there are two traces for each temperature listed as "NF".
Interesting video. I designed a heater sock and removable nozzle sock with a small air gap to reduce heat conduction. Easy to see the results with an IR camera or measure the mains power reduction.
Replacing the 2 toothed idlers (at the ends of the X gantry) with bearing stacks, loosing belts, increasing current on the A/B motors, cooling them or upgrading to the V0.1 motors are things that helped me with better motion
This makes me think about two things: - heated chamber would help - pre-heating filament would help (also useful for drying the filament or keeping it dry)
@@CNCKitchen I'm mostly happy, the FORMBOT kit from aliexpress gives you the bare minimum of pieces to get it built. I made the mistake of plrinting the parts in PLA with the idea of re-printing them on the voron in ASA... that was a challenge! If i was to build it again i would make the 2020Vslot variant with absolutely NO M2 SCREWS :-)
I had thermal runaway problems with long fast movements on my MK3 too when i relocated it to the basement. Capping the max fan speed to 80% solved it without having a heater next to the printer or adding a sock.
It makes sense, Stefan. A sub-10' benchy is quite a big challenge. I didn't attempt it yet, not having finished to tune Klipper for my printer (not sure I'll attempt it, "speed contests" are not my bag). What you may try, just as a comparison point, is an E3D Volcano hotend. As you said, if one wants to print fast, the filament must have enough time to melt. Volcano has a larger heatblock, so it might help. It could be interesting to compare how it performs alongside the Mosquito in the same context. Also, depending on the filament material, adequate cooling is required. Don't know, maybe use a 5020 intead of a 5015. Something like that. Just my 2 cent. ;)
It's really awesome to see how things like temperature and extrusion speed affect what actually gets extruded. I'd like to see a printer firmware in the future that takes these things into account. For instance, increasing extrusion multiplier and hotend power with high extrusion speeds.
5:38 You should change your graph description. In booth diagrams there are labeled as [Temperature] & NF. For Normal Flow it makes sense, but not for High Flow.
I think the magnum is slower because there is more friction in the system due to the high viscosity of molten plastic. Also bowden tube friction will be more of an issue at high feed rates. Could you do a graph of extruder stepper motor power consumption? Could shed some light on the situation.
If you can, I would love to see how a Nova Hotend would work on the V0. I dont see too many people reviewing it and would love a proper review aswell. Thank you!!
I don't care about fast but... I appreciate your content and can't wait to get all my goodies to build my little V0.1 . Most of what I print are small parts so this little machine just makes sense. Or maybe 2 or 3 of these little guys =)
I was going to mention until you said it, and it is what i do, increase the temperature above recommended when printing fast because as we know the viscosity of a polymer is indirectly proportional to the temperature however, heat might start travelling back up the heat break...
While rebuilding the voron, please check voron 0.1, which was released about 20 days ago. There is a direct drive extruder option as well as some extra/updated parts. I'm building my Voron v0.0 and found quite usefull linear rail stoppers, for example. Plus more detailed manual
Amazing stuff! And cool setup with the creep test - I'm looking forward to seeing the results :) Instead of putting the printer in your basement, you should wait for Summer and put it on the attic ;)
a few things. - the nylon is only going to waste your time and make you furious in the long run. Id personally suggest the use of ASA as its less prone to degrading from the hydrocarbons that will eventually leak out of the cheap idlers. But abs is fine, as it creeps very little. Theres a reason nylon usually has a ton of glass in it (not the dust you see in filament), the long strands reduce the creep over time. Its also almost never used in shear or tension unless its molded with a glass content of 40%+ - invest in a spool of MG94 abs, its a base resin by sabic, a few vendors sell it - interpolate off, spreadcycle, and run the motors at 80% of peak current - stick a box fan in front of the printer once it hits layer 5 this should get you down to 8 minutes, assuming the frame is square. the rest youll have to figure out on your own we look forward to seeing an official submission on our leaderboards
I printed all my parts for a corexy mod I did for the ender 5 with CF nylon(sainsmart). I annealed all the parts in the oven(most structural parts and the carrier were printed at 100%) and then proceeded to paint all the parts with either black or silver stove paint. Still going strong after a year of printing.
Those standing waves at around 11:30 are very interesting. It shows that the liquid shows some high viscosity effects with non-linear behaviour. It is quite similar to how honey behaves and the really bad part about it is, that it shows reflection properties. If you use honey and let a steady streamflow at the right height, you can create a reflection arch forming a bend parabola that oscillates periodically but stable. Works with some shampoo too. However, as interesting as this is from a physicists point of view, as long as the material is in this window of operation, you will not be able to get reliable prints. Dependencies here are unfortunately not only temperature and surrounding factors, but the flow speed itself. As this changes during the print, the material property does as well.
next voron themed video : after finding that annealing the nylon worked well, I was bored and just annhealed the full printer without dissassembling it ... Just kidding, thanks a lot for all your well made, well documented, well explained, and well engineered videos
Pro tip for silicone socks: Take some narrow kapton tape, and give it a couple wraps around the sock once it's on the hotend. The kapton won't stick to the silicone sock, but it'll stick to itself once you get past the first wrap. I've never had a sock come off since doing that!
Stefan, for the parts - it will be really great to get another filament comparison. Specifically the Priline Polycarbonate Carbon Fiber with the Nylon Carbon Fiber and PETG CF variants. I found the Priline quite reasonably priced, easy to print and seems to be quite rigid but not sure of its actual mechanical strength - it will also be really great if you can compare annealing vs not as the printed parts can be annealed with 100 degrees oven without deformation and supposed to help quite a bit. I am printing parts for my printer but dont have actual proof this stuff is really living to the theory. Thanks!
I'd call that creep test a stress relaxation test, but it'll give good comparative results as long as the screws are clean and consistent. Comparing PLA, PLA+, PETG, and Nylon for use in plastic springs I've found that PETG makes the best springs. It creeps less than PLA or Nylon so you can use it at higher stress levels, and the low stiffness means that you can also use it at higher strain levels.
I really like this type of content: intelligent and well researched. One comment about your creep test - bolt tightening torque is quite a poor indicator of the clamping force, the tightening torque is mostly governed by the friction between the nut and bolt thread. If you lubricate the thread you will get a vastly higher clamping force for the same tightening torque. Basically different bolts and nuts out of the same batch might have quite a bit different clamping force for the same tightening torque. Something to be aware of.
As someone with a partly built Voron 0 / 0.1 hybrid (it's a long story...) my sympathies at needing to touch the motion system. I have a V0 Z motor assembly in an otherwise V0.1 kit, and disassembly to remove the *one* rail I need to move to insert two T nuts the older assembly requires (Z rail & leadscrew is a much earlier step in V0) will make me take everything apart from the XY motion system to the little rubber feet. For the week between that discovery and now, I consider it an art piece as I wait for an integrated lead screw to be available to buy.
I'm curious about the annealing of the nylon-carbon fiber. A video about how annealing changes the nylons properties, similar to your PLA video, would be very appreciated. I wonder if maybe the reduction of creep is just a result of drying the nylon, and might revert again? I've noticed with my own pa-cf prints, that while they start of very rigid, they become more flexible over time again as the printed part absorbs water.
Is fast the new big? What do you think about this new speed-printing trend?
Don't forget to like & subscribe and share this video on Facebook, Reddit, Twitter, and other platforms!
It's extremely fascinating to watch someone like MirageC print the 5min benchy on his HervORT. Looks like a timelapse if you turn the sound off.
stephan, i have the exact same wera toque wrenches here. they are awesome. you have made a pretty excelllent creep test in my opinionwith the dials, and also nice to hear about annealing the cf nylon compared to for pla. thing is... cf nylon is still too expensive for most of us. so maybe something else like that greentec pro cf is a bit cheaper? since we have to anneal them whichever option is chosen. at least the greentec cf pro is more environmentally friendly. but is it as strong as the claims? also have you gotten a sample of the brand new nonoilen (=pla and phb mix). because that is also claiming to be a strong material too, whilst also entirely environmentally safe and biodegradable. at least for that same sort of prices this is a genuine extra feature makes it feel a better 'value' (not for us, but for nature). so is cf nylon really worth it (for fdm purpose) by comparison? this is a genuine question i want to know the answer for! and doesnt seem like anybody else can convincingly answer it
as for speedboat race, main casualty is the inter layer adhesion. so finding out about that is... important to put it into context. although maybe if you do anneal afterwards a speedbenchy then perhaps it doesnt matter? although annealingstep itself adds on a lot more of extra time. hehe
[added]
lol sorry i forgot to say something really important: there is this guy i know, and after finishing the print (in his same enclosed chamber) he then ramps up the temperature several times. for example 3 seperate times, starting about 60c and going up higher each time. ending at the highest annealing temp. maybe ending 100c or more. just whemever it makes the sense to stop cycling.
this ramping type of work hardening / annealing process is better, in particular for these enhanced 'high temp' pla types. so this is also the recommended way for those greentec pro cf, and the nonoilen, and the esun pla+. any enhanced pla. just something you needed to know about, if you are struggling for annealing pla. doing the ramping program is essential. it keeps a partial crystalinity and then the object does not melt and loose its shape. which was in your previous experiences for annealing pla. so please try this method in future. many thanks
@@AlexusMaximusDE That thing is a BEAST and the guy is really cool, always interacts with the viewers too.
It's always fascinating to see what the heck is the next physical limit he's gonna hit with that machine. At this point I'm thinking the plastic itself is close to being a limiting factor.
Good for improving prototyping with FFF, especially if it leads to printers/parts focused on consistent throughput without sacrificing quality.
@@AlexusMaximusDE 3 minutes incoming :)
What I love about Stefan is he discusses his own errors. I wish more people would behave this way.
A "5 minutes 3D Benchy" video might be more clickable but doesn't really teach anyone something. My goal is to fail, learn, repeat and eventually succeed. That's how the ones how want to learn something benefit the most!
40 seconds in and this is the most defeated sounding intro I've ever heard. You can do it Stefan! I believe in you
At the end now. You've learned so much in this video. I've picked up a lot too!
I don't have a 3D printer, but I enjoy the learning and entertainment
That "I failed" thumbnail too
Let us take a moment of silence for all those benchies that fell in the line of printing...
That "sunk" in the line of 3d printing
What is that intro music? Reqieum For A Dream, The Fountain? EDIT: Ahh Max Richter not Clint Mansell.
That creep test is genius!
Thanks!
We have to thank you for this nice footage!
@@CNCKitchen eta on the video?
Very nifty test, but the type of creep on the printer bed is constant force. This creep test is constant strain. The bolt tension translates to a fixed compression. As the material experiences creep the force on the material will approach 0. An example experiment for a constant force test might be to hang a weight off of a sample instead.
@@ojhall3164 weight tests are good but, a creep test such as this one allows you to place the item wherever you like
Hi Stefan, I really appreciate the rigorous approach you take to testing different aspects of 3D printing. It must be very time consuming to test and record each variable and I am grateful to you for sharing your results with us.
He reminded makers of two things: 1 if you're printing abs then use an enclosure
2. For better print quality use a silicon heat block sock
Not sure how you arrived at 1. At those speeds, you need every bit of cooling you can get, even for abs. It's pretty much printed faster than it can warp. An enclosure would only hurt.
Yes but clearly he saw a difference when he moved his printer to the basement. An enclosure would have solved that problem
Usually it is very important for ABS to be printed in a warm enclosure (that also keeps the nasty smell in a little), though for speed printing, that's exactly the wrong thing because you want to cool the material down as fast as possible. The problem with my basement was rather that the heating system wasn't properly able to cope with the lower temperatures.
He should probably use proper mounting hardware for his sock.
hmmm... this makes me wonder if there should be a 'feed forward' control system for the nozzle temperature that takes the current extrusion rate into account. Basically preemptively increase the hotend power if there is a fast extrusion move in the near future.
That might be interesting, or DUET/RRF can uses Non-linear extrusions to compensate for that effect (M592).
This is in fact called feed forward, or "2 dof control" and shouldn't be hard to implement because you only need to add the feedforward term (theoretical power for melting and convection) to the feedback part (PID control for the error between set and actual temperature). If you only use feedback control (the way it's currently done), the true temperature will always "lag behind" the reference to some degree.
I prefer bang bang for the hot end control over PID. It doesn't work in Marlin though because it doesn't sample often enough. In my own firmware I sample very quickly so that close to the set point noise gives a small range of PWM over a fraction of a degree.
@@nophead Haha wow it's like talking to a celebrity, thanks for everything you've done for the community since I got into this hobby 10 years ago :-)
I think bang bang control is actually a great choice. PID just tends to work with most simple systems (and is more forgiving if you're slower than your system's dynamics) although it's certainly not the perfect controller for this type of system.
What I'd like to see also is a controller for the nozzle temperature of the plastics instead of the temperature of the NTC. With a thermal model of the hotend and plastics and an observer this should yield near-perfect results. Although it's probably not really worth the effort.
That's an interesting idea. Another way to think about it is the filament has a cooling effect on the hotend that is proportional to feed rate. So it only makes sense to counteract this cooling with equal and opposite heating from the hotend.
Normally this is handled lazily by a separate hotend temperature control PID loop but since heating happens slow and changes in extrusion rate happen fast, there is always some delay in the heater cartridges response. Also it takes some time for the thermistor to register that the hotend is cooling down before it can react.
OH MAN I love the creep test concept. That is really cool.
like a hidden code on ps5...
For your creep testing, how high of a temp are you testing too? In most vorons enclosed the ambients can sit around 50c when printing ABS.
Noticing the layer shifting? Are you tweaking accel/decel values as well as IS
For replacing the collet, ya that part kinda sucks, i popped mine out too originally, best recommendation is install it BEFORE screwing both halves together so its sandwiched in place instead of pressing it in, seems to grab it a bit tighter. (or go with the direct feed toolhead cus 0.1 is out now )
"Increased" temperatures will, for the first round, be only 40°C to stay far away from any glass transition temperatures. Believe it or not but that already drastically changes the results!
I'm a 3D printing neophyte compared to all the testing you perform, but the two things that have solved layer shifts for me have been putting loctite on the motor pulley grub screws, and increasing belt tension. I'm enjoying how you are sharing your progress on this challenge.
That's a lot of analysis to print a benchy in less than 10min ! Keep the good work!
I'm so excited for a series on high speed printing!
This feels like your boss fight! All the research you've put in for years culminating in the speed benchy challenge. Good luck and thanks for sharing your findings!
Hi Stefan, your scientific approach to 3D printing is just on another level. Thank you for sharing with us hours of testing and valuable pieces of information.
Stefan, your stuff just makes me smile, Your either a savant or an engineer, maybe even a research scientist, cus this stuff is just sensational.
The intro music and panning over failed benchy boats was an immediate Thumbs Up from me :P :D
Thanks!
I can’t wait for the creep tests. Love to see ASA vs abs and even annealed pla.
Might be interesting to see some different timed ones, like a 30 min, 20 min. benchy etc. in the long run.
Good point - especially what and who they compare. Not everyone has a setup to go to 10 minutes so something slower might give them a good baseline for their setup.
the test for creep is brilliant! i'm really curious to see your progression on this project!
I love the artful display of the failed benchies - definitely highlighting the issues is a fantastic way to help us non-professionals feel less bad about our mangled prints :D
Oh boy, I've been looking forward to the creep testing for years! I know it's got to be a pretty massive amount of work, so I really appreciate you doing it for us!
I also LOVE the speed printing trend!
Thank you so much Stefan! With your channel I finally fixed all problems of my homemade mini 3d printer! ❤️
Love this vid as I've been looking to building a voron 2.4. Very informative to have you answer questions I hadn't even thought of yet.
Also, yes, silicone socks are great and pretty much essential imo. My temperature goes all over the place without it on my x1. Just bought another one. Thanks for the reminder and great video as always. Resubscribed after 2 years.
Sorry, you mentioned this later in the video as i continue to watch. My bad.
Beyond high volume extruding, I think high speed kinematics would be a cool series... Basically increasing print speeds and trying to determine what needs to be improved
At around 12:02 and and 13:01 there is vapour coming out from the nozzle if i see it correctly. Due to very high latent heat of vaporasition of water, i suspect it might increase your heating load dramatically. Amazing work btw thank you :)
Good point as well and that also get's worse the higher the temperature is.
All of your wonderful videos are an introduction to methods and theory of mechanical engineering. Please keep up the good work.
Fantastic video! These Voron videos are great. If the Prusa XL wasn't around the corner I would build a 2.4 yesterday, it seems so great. Thank you for taking us on this journey of insane speed.
I've learned more experimentation techniques from one of Stefan's videos than my entire undergrad in physics. So much good stuff.
Great to hear!
Accuracy, print stability and consistency - not speed is what I am trying to achieve. After I master those - I will probably work on some speed improvements. Thanks for showing us your methodology - you teach me (us) a lot when you do that - so we can do it for our setups.
Again, Thank You!
Good job, Stephan. All your tests are very interesting.
Always glad to see the Voron 0 out.
after coffee i find my extrusion rate is greatly increased.
definitely a reduction in quality though
Can relate.
I think the most interesting thing about this speed printing challenge is that it looks like we've reached a point where the mechanical structure of the system is able to better handle the speeds than the other components.
CoreXY and a lot of the clever implementations of it have really brought the general structure of 3D printers to a new level.
Steffan - The Mosquito hotend is amazing! I've been using it for around 7-8 months now. I have some of my own designs exclusively up on thingiverse, search "blood drive ender 3", my mod is purposed around a linear rail converted Ender 3. I've been able to achieve 25mm/3 flow rate with PETG, PLA has more difficulties.
Somethings I'd recommend - You should be using boron nitride paste that you can get from Slice Engineering. Infact I think this would be a good test to conduct on its own for your channel.
The thermistor and heater core should be coated with it for maximum heat transfer. Also, the threads in the Mosquito hotblock as well as the threads on your nozzle should be coated as well. It is significant, I believe. Additionally, nickel plated brass nozzles will further your heat transferrence even more.
Hi Stephan, as part of the rebuild, you may want to consider going to the V0.1, with a DD extruder that eliminates that collet issue.
0.1 also has a bunch of nice quality of life improvements and is easier to service.
I enjoyed this video. Thanks for all the analysis and lessons. One note, for your extrusion rate graphs, the legend says NF for all plots, even those that are for the high flow heat break.
I'm loving this series!
The deformation test looks awesome, very clever!
Awesome!!
Comparing Polymaker materials TDS, PA12-CF has almost double the melt index of ABS, and becasue of the fibers, it will hold its shape up to very high temperatures. Maybe a filled PA is a good candidate for the fast benchy :)
you´r always doing such amazing tests, keep it up!
Seems an excellent way to measure creep. Looking forward to the results.
Cool - I discovered the same thing with the sock within my PhD too. 🤓 Good job!
You got to study this in university? Jealous man
@@User-nc2lf Being more precily, had to, because it still was very difficult though.
@@jankohler562 What is your PhD today? Do you still 3d print yourself or was it mostly for educational purposes?
@@User-nc2lf Thank you for your interest. (^‿^) I researched creating metamaterials for high frequency components using FDM printing. Even though I finished my PhD I continued working on that topic, because it is so interesting. Let alone, I love the topic of 3D printing. Not a day goes by without me working or creating things with the 3D printer. This alone made it all worth it, even though it was very difficult.
Im in love with Stefans brain.
thats the kind of Raw information i like
I see nova hotend with 400mm/s and its insane :o
PLEASE MORE :D
Your work is awesome!
They print PLA at like 260 °C
Very interesting video, thanks as you answered my question in the video on why the filament starts to skip. 👍 I’m a new to 3D printing, I spent some time in the oil field as a machinist and can see it’s no different lol it all takes time practice and patients. Thanks again
I use socks for exactly that reason. My MK2.5s won't keep heat steady without one.
I do tend to attach it with a few dabs of high temp silicone. Keeps it from falling off unintentionally.
Thanks for the info. Good that you tell us about your failures. This is very valuable.
Tiny nit pick - the flow graphs such as at 6:23 list all of the traces as "NF", I guess Normal Flow, and none as "HF", so that there are two traces for each temperature listed as "NF".
Interesting video. I designed a heater sock and removable nozzle sock with a small air gap to reduce heat conduction. Easy to see the results with an IR camera or measure the mains power reduction.
Replacing the 2 toothed idlers (at the ends of the X gantry) with bearing stacks, loosing belts, increasing current on the A/B motors, cooling them or upgrading to the V0.1 motors are things that helped me with better motion
That's not a fail, that is modern art!
Thanks for sharing your proces, really helpfull in understanding how a 3D printer works.
I will build a Voron 0 because of your excellent videos. :-)
OMG I'm so excited for results as well.
Thank you for your hard work on this !
My pleasure!
This makes me think about two things:
- heated chamber would help
- pre-heating filament would help (also useful for drying the filament or keeping it dry)
After you showed the V0 on the meltzone, i decided i definitely need one. now i have one. :-D
Great to hear! Are you happy with it?
@@CNCKitchen I'm mostly happy, the FORMBOT kit from aliexpress gives you the bare minimum of pieces to get it built. I made the mistake of plrinting the parts in PLA with the idea of re-printing them on the voron in ASA... that was a challenge! If i was to build it again i would make the 2020Vslot variant with absolutely NO M2 SCREWS :-)
Excellent video Stefan. Thorough and scientific as always 👍. I learn a lot from you so thank you😎
I had thermal runaway problems with long fast movements on my MK3 too when i relocated it to the basement. Capping the max fan speed to 80% solved it without having a heater next to the printer or adding a sock.
Great video as always 👍
I love your experience and systematic approach 👍
Thanks for sharing your thoughts with all of us 👍😀
the v0.1 update really improved my print speeds and quality
It makes sense, Stefan.
A sub-10' benchy is quite a big challenge. I didn't attempt it yet, not having finished to tune Klipper for my printer (not sure I'll attempt it, "speed contests" are not my bag).
What you may try, just as a comparison point, is an E3D Volcano hotend. As you said, if one wants to print fast, the filament must have enough time to melt. Volcano has a larger heatblock, so it might help. It could be interesting to compare how it performs alongside the Mosquito in the same context.
Also, depending on the filament material, adequate cooling is required. Don't know, maybe use a 5020 intead of a 5015. Something like that.
Just my 2 cent. ;)
It's really awesome to see how things like temperature and extrusion speed affect what actually gets extruded. I'd like to see a printer firmware in the future that takes these things into account. For instance, increasing extrusion multiplier and hotend power with high extrusion speeds.
Awesome video Stefan! I can't wait to see the next videos in this series!
Just awesome, i appreciate every second of your time you invest into your great video content. *chapeau* ❤
5:38 You should change your graph description. In booth diagrams there are labeled as [Temperature] & NF. For Normal Flow it makes sense, but not for High Flow.
I think the magnum is slower because there is more friction in the system due to the high viscosity of molten plastic. Also bowden tube friction will be more of an issue at high feed rates.
Could you do a graph of extruder stepper motor power consumption? Could shed some light on the situation.
Thumbs up on the creep test. Very interested to see the results for that.
Keep it up. Always very informative!
Best video I was just thinking about getting into this stuff
Is the creep test still ongoing? I'm really interested in the long term results :)
If you can, I would love to see how a Nova Hotend would work on the V0. I dont see too many people reviewing it and would love a proper review aswell. Thank you!!
Always love your videos!
As always, a ton of useful info here. Thanks!
Great informative video as always! Thanks Mr S
I don't care about fast but... I appreciate your content and can't wait to get all my goodies to build my little V0.1 .
Most of what I print are small parts so this little machine just makes sense. Or maybe 2 or 3 of these little guys =)
I was going to mention until you said it, and it is what i do, increase the temperature above recommended when printing fast because as we know the viscosity of a polymer is indirectly proportional to the temperature however, heat might start travelling back up the heat break...
While rebuilding the voron, please check voron 0.1, which was released about 20 days ago. There is a direct drive extruder option as well as some extra/updated parts.
I'm building my Voron v0.0 and found quite usefull linear rail stoppers, for example. Plus more detailed manual
I'm really looking forward to your creep test.
Amazing stuff! And cool setup with the creep test - I'm looking forward to seeing the results :)
Instead of putting the printer in your basement, you should wait for Summer and put it on the attic ;)
Stefan, you’re the best. This content 🔥🔥🔥🔥
Love that testing procedure
Good rundown. I got a Chinese mosquito clone and it's an odd threading but at least it came with a heat sock
a few things.
- the nylon is only going to waste your time and make you furious in the long run. Id personally suggest the use of ASA as its less prone to degrading from the hydrocarbons that will eventually leak out of the cheap idlers. But abs is fine, as it creeps very little. Theres a reason nylon usually has a ton of glass in it (not the dust you see in filament), the long strands reduce the creep over time. Its also almost never used in shear or tension unless its molded with a glass content of 40%+
- invest in a spool of MG94 abs, its a base resin by sabic, a few vendors sell it
- interpolate off, spreadcycle, and run the motors at 80% of peak current
- stick a box fan in front of the printer once it hits layer 5
this should get you down to 8 minutes, assuming the frame is square. the rest youll have to figure out on your own
we look forward to seeing an official submission on our leaderboards
great video, some lessons learnt here for sure
Proper engineering, love it!
Such thorough analysis!
I printed all my parts for a corexy mod I did for the ender 5 with CF nylon(sainsmart). I annealed all the parts in the oven(most structural parts and the carrier were printed at 100%) and then proceeded to paint all the parts with either black or silver stove paint. Still going strong after a year of printing.
You could try the Phaetus dragonfly HF HIC. It has a very large melt zone
Those standing waves at around 11:30 are very interesting. It shows that the liquid shows some high viscosity effects with non-linear behaviour. It is quite similar to how honey behaves and the really bad part about it is, that it shows reflection properties. If you use honey and let a steady streamflow at the right height, you can create a reflection arch forming a bend parabola that oscillates periodically but stable. Works with some shampoo too. However, as interesting as this is from a physicists point of view, as long as the material is in this window of operation, you will not be able to get reliable prints. Dependencies here are unfortunately not only temperature and surrounding factors, but the flow speed itself. As this changes during the print, the material property does as well.
next voron themed video : after finding that annealing the nylon worked well, I was bored and just annhealed the full printer without dissassembling it ...
Just kidding, thanks a lot for all your well made, well documented, well explained, and well engineered videos
Pro tip for silicone socks: Take some narrow kapton tape, and give it a couple wraps around the sock once it's on the hotend. The kapton won't stick to the silicone sock, but it'll stick to itself once you get past the first wrap. I've never had a sock come off since doing that!
Stefan, for the parts - it will be really great to get another filament comparison. Specifically the Priline Polycarbonate Carbon Fiber with the Nylon Carbon Fiber and PETG CF variants. I found the Priline quite reasonably priced, easy to print and seems to be quite rigid but not sure of its actual mechanical strength - it will also be really great if you can compare annealing vs not as the printed parts can be annealed with 100 degrees oven without deformation and supposed to help quite a bit. I am printing parts for my printer but dont have actual proof this stuff is really living to the theory. Thanks!
I'd call that creep test a stress relaxation test, but it'll give good comparative results as long as the screws are clean and consistent.
Comparing PLA, PLA+, PETG, and Nylon for use in plastic springs I've found that PETG makes the best springs. It creeps less than PLA or Nylon so you can use it at higher stress levels, and the low stiffness means that you can also use it at higher strain levels.
I really like this type of content: intelligent and well researched. One comment about your creep test - bolt tightening torque is quite a poor indicator of the clamping force, the tightening torque is mostly governed by the friction between the nut and bolt thread. If you lubricate the thread you will get a vastly higher clamping force for the same tightening torque. Basically different bolts and nuts out of the same batch might have quite a bit different clamping force for the same tightening torque. Something to be aware of.
As someone with a partly built Voron 0 / 0.1 hybrid (it's a long story...) my sympathies at needing to touch the motion system. I have a V0 Z motor assembly in an otherwise V0.1 kit, and disassembly to remove the *one* rail I need to move to insert two T nuts the older assembly requires (Z rail & leadscrew is a much earlier step in V0) will make me take everything apart from the XY motion system to the little rubber feet. For the week between that discovery and now, I consider it an art piece as I wait for an integrated lead screw to be available to buy.
Your struggle is our blessing! Keep failing and sharing please!
Really interested in the results of the creep test!
I'm curious about the annealing of the nylon-carbon fiber. A video about how annealing changes the nylons properties, similar to your PLA video, would be very appreciated.
I wonder if maybe the reduction of creep is just a result of drying the nylon, and might revert again? I've noticed with my own pa-cf prints, that while they start of very rigid, they become more flexible over time again as the printed part absorbs water.