Angus, I watch a lot of 3d printing videos here on youtube. And I just want to say: yours are the best! I've watched you for a while now and you always have good interesting ideas, low editing no jump cuts and they are just quality. Thanks for all the videos you make we out here in the 3d print community appreciate it and you might not hear it enough but these videos are critical to people like us. Sometimes i feel like you have a much more technical degree than just an industrial design degree, you're a smart dude! Be proud of the following you've created, and the good you've done for the community in general. I'm sure many others feel this way but for whatever reason have never taken the time to vocalize it to you. Cheers from California!
Hi, Thanks for sharing. There is a real problem when doing "small perimeters": this induces vibrations because the speed is constant so is the spacial frequency. One solution is to use a no-constant speed. I've worked in the past on big CNC machines, and we had the same problem (obviously not at the same level) when trying to go too fast. One solution has been to do a random variation of the speed when the direction of machining (print in the case of a 3D printer) has changed of more of 90°. This way, without any mechanical dumper, we have some kind "software dumping" [tm] and the quality is greatly improved. The speeds have to be choose to depend on the spacial frequency and eingen modes of the machine.
Interesting idea! This is similar to the spread-spectrum techniques we use in communications to spread the power out over a range of frequencies rather than concentrating it all at a single frequency. This can help avoid creating interference with other signals and also makes the signal more tolerant to outside interference.
@@reverse_engineered The problem is when the acceleration is very high. This acts like a dirac pulse and excites the resonant frequencies of the whole machine.
Angus, one other thing you can do to add strength without adding time is push your extrusion width to .6mm - your top & bottom layers will print faster with fewer passes & you can get even more strength with just those 2 perimeters. If you normally print with Z hop & don't have any delicate shapes that may get knocked over or extreme overhangs that will curl up, you can remove Z-hop for even more time savings. The .6 width with a .3 layer height will not do overhangs past 45 degrees well - so be warned! Those curved undersides of your frame would benefit from a vertical "wall" of trimmable support (similar to what you did for the horizontal bridges).
One thing to be aware of when using larger nozzles, wider extrusions, and thicker layers is that you will eventually be limited by the volumetric flow rate that your hot end can provide. For a lot of cheaper printers like the Ender 3 clones, you won't be able to squeeze out much more than about 10-15mm^3/s. However, with better nozzles, heat breaks, and heaters, you can push this much further, with some going well beyond 30mm^3/s.
@@barbarosbozkurt758 cubic (aka to the power of 3 which is what ^3 means) millimeters is a measurement of volume just like cubic centimeters or cubic inches, cubic feet etc.
I know that but I usually see the speed of the printer which is in mm^2, I didn't ask what it is, I asked why, again thanks for clearing it up@@negativenarwhals
I've been using similar settings on my MK3 for quite a while now, I use 0.32mm layer heights and get great results, print times easily twice as fast as the default 0.15mm Optimal and still way faster than the 0.20mm Draft as well. Figuring this out really helped keep the design hobby going for me - I was getting very frustrated with the iterative design approach and having to wait half a day for my part to print before testing.
Super useful tip. I mainly print organizers and inserts for board games so the quality sacrifice isn’t really a big deal and this cuts my print times by 50-60%. Thanks!
Anny, you didn't did the one and only thing that really makes the difference: to round all your sharp corners so machine won't have to slow down on every sudden direction change. There is a lot of corners in your model that you don't need for anything. Also I do extrude my PLA at 220-230°C and have zero issues with supports or bridges. And it makes layer adhesion only better. Also put on that 0.8-1.0 mm wide nozzle, this will make more difference than anything else if you are in a hurry.
Angus ,I don't own a Prusa but I found this very helpful in navigating the frustrating world of slicer settings. This 3d Printing hobby is turning out to be more like becoming an expert in a specific field of industrial engineering than just "printing stuff". I'm certainly not complaining mind you. I find it amazing that we at home can learn something so complicated. Now that is the true purpose of the Internet. I have an Anycubic I3 Mega as I mentioned above.
I long ago settled on a 0.5mm nozzle and 0.3mm layer heights, and running PLA @210C for speed. Very seldom less than 3 shells, and usually at least 30% infill. My parts are strong and suitable for mechanical use with any minor blemishes irrelevant most of the time.
Hi Angus. The way your part is designed has a significant impact on the part's rigidity as well. I believe that that's why you couldn't break it apart. Your part design is sound, and will withstand the stress put on it. In fact, your part is so well designed that when you attempted to flex it you couldn't.
The explanation of why it is way longer than you would think, for anyone who’s curious, is that if you double the size in all 3 dimensions of a 1in cube, it is then 2 in by 2in by 2in the area of a 2in by 2in square is 4in squared, so it increases by a factor of 4, likewise, the volume of a 2in by 2in by 2in cube, is 2x2x2, so 8in cubed, meaning the volume is increased by a factor of 8. The high amount of filament that needs to be extruded is the main cause of long print times, and the total amount of filament that needs to be extruded is directly proportional to the volume of the part. This still applies without 100% infill because you are still increasing the volume of the walls themselves. Hope this helps anyone who was curious! 😃
Great video! One note: as size increases the volume increases cubically, not exponentially. It's important not to confuse polynomial and exponential orders of magnitude.
A trick that i like to take advantage of when i don't want to increase layer height is printing at 0.16mm layer height and doubling the infill and support layer height in Simplify3d. 0.32 is the highest you can go with a standard 0.4mm nozzle and that trick works wonders with lots of infill and support compared to printing everything at 0.2.
Just got back into 3D printing (today) after my first failed attempt with the EZ3D printer. I tried your settings on my new Ender 3 Pro and they came out great! I honestly think I could push it further, but with it working this good so far I will stop here for now. It's crazy how the price has changed so much since I first tried 3d printing.
Thank you for making this video. it convinced me to try tweaking settings that aren't factory. I am often prototyping and don't need crazy amount of detail, but i do need speed. Just successfully printed PETG at a .3 layer height with a line width of .5 at 75 mm/s on an ender 3. What would have taken me 9.5 hrs at cura's standard quality .2mm layer height only took 5 hrs
Max travel speed 200 for prusa doesn't have any negative effects for me, it's pretty safe. Also I change the min retraction distance to 3mm and remove a layer change retraction requirement, it makes prints much faster too.
More tips if you are using Cura: use adaptive layer height so you can push the maximum layer heights a bit higher without losing as much detail on curves and holes, Cubic infill with several gradual infill steps (to leave it more hollow where it doesn't matter as much) without risking pillowing or gaps like going straight low % infill, and 10 or so degrees hotter temps. Pushing line width can make a big difference as well, just make sure not to exceed your extruder's max volumetric flow (usually 8-15 mm3/s on std all-metal heater blocks, half that for PTFE sleeved). Volcano with larger nozzle is better, obviously, but I tend to run volcanos on everything, regardless of nozzle diameter because the flow is just so much better and more consistent (and much better volumetric flow, up to 35 mm3/s on my setups).
Dude, you're a legend for just brazenly showing your settings and walking through them. I can't tell you how many UA-cam videos I've seen up until finding yours where it amounted to a nothing burger with the instructor just dancing around showing some hard settings. Instant like and sub!
Hi Angus. You inspired me to try a stupid fast setting myself on my Anycubic Kossel Linear Plus. After some issue’s with bed adhesion and upgrading to PEI powder coated magnetic bed, I found good results with first layer and my own stupid fast settings (using Cura 4.8, 4.12.1 or Arachne Beta2) which I’d like your take on. My layers are 0.32mm with 1st layer 0.3mm. I set my wall widths to 0.8mm using a 0.4mm nozzle. 1st layer Extrusion at 130% (for guaranteed bed adhesion) then back of to 100% extrusion for the rest. Hot end settings at the highest end of recommended filament heat range(I use 230°C). Speed wise I’m ranging between 60 and 120 mm/s with acceleration at 2000 mm/s/s. I was surprised that the 0.8mm line width extruded so well from a 0.4mm nozzle and resulted in very good layer coverage and layer adhesion. A calibration cube in vase mode at these settings strarted me off on using this setting as my own goto stupidly stupid fast larger model settings. The 300% vase mode calibration cube (60mm x 60mm x 60mm with a 0.8mm wall width) using Curas ‘0.2mm Draft mode’ takes just over 2hrs to print at the default settings. This one takes 21 minutes. Not bad.
Rectilinear infill DOES add strength in all 3 directions. It adds the most strength in the Z, that's true. But adds significant strength in X and Y too. Maybe thats a nitpick. Good video 👍
this saved so much time thank you! Also not just for test prints but loads of functional prints. i needed to print 240 time a print that normaly takes 13 hours now 3 hours each with still reasonable quality some slight imperfections but for a functional print perfect. this is done using ABS bdw.
Thanks Angus - and here are some things to maybe think about: Most printer frames are pretty flimsy - so much so that a relatively light printer head moving around with really no impediment to it moving is often causing flex and resonance. On the other hand, CNC router machines are swinging a full on router around and the tool resistance when cutting is very much higher. So lets start by making the frame of the printer more solid and able to absorb any forces applied without flexing and/or juddering all over the place. That theoretically will mean that you can jack the speed way up. The days of printer frames being made out of acrylic are going to come to an end. The next point may or may not be possible but here goes... I have heard that it is possible to monitor the amp usage of a stepper to determine the load being experienced by that stepper. If this is indeed the case then would it not be possible to monitor real time the loading inside the extruder? If this is possible then would it not be possible to jack up the temperature as you have done and use a PID loop to monitor the "back pressure" on the extruder (high back pressure means that the filament is not being melted fast enough) and if that is the case, slow the movement of the head down to keep the extrude coming through the nozzle able to lay down the optimum layer. Which basically means that the printer should always be moving the extruder at the maximum speed and because the machine frame rigidity is so much better, all the ghosting/artifacts etc should be greatly reduced and issues like layer adhesion should also be optimised. An idea of how to achieve the "optimum variable speed" would be to have an Uno monitoring the extruder motor load and adjusting the speed percentage setting on the main machine driver board on the fly as needed. If all this can be done then an optimal print will happen every time and the only thing that will change quality of the print will be the layer height setting. Just an idea... How about passing this along to Joseph? His guys should be able to make it all work - if it is possible ;>}
Most frames are already metal, only dirt cheap machines are acryllic (and even then there are cantilevers and Ender 3 for under 200 dollars that are metal). 20-series aluminium extrusions seem to be a new gold standart for 3D printers, and at least from personal experience (going from A8 to a custom frame with V-slots instead of linear rods) the frame rigidity skyrockets. A8 has almost vibrated itself to pieces (the screws were coming loose all over the frame), the AM8-custom can go over 100 mm/s without even reprogramming it. The age of acryllic frames is pretty much already over, and it won't be missed.
MK4Duo optionally does thermal compensation for the heat lost as extruded material. It's physically based and easy to configure. I'll try and port it over to Marlin2. Fundamentally temperature control overrides such an adjustment long term, but it can help avoid short term fluctuation in pressure difference. I don't see the point of measuring extruder current. Basically you have two fundamental issues, slipback and grinding, which are related, because the motor doesn't feel the filament not being advanced enough. You can trivially avoid stalls just by setting drive current high enough, and you have to because extruder stall is pretty much never a good thing except in emergency. Extruder operates at holding torque, essentially near standstill, so Trinamic current sensing is not possible. Prusa has a sensor identical to one you'd find in a laser mouse trained at passing by filament, so it can measure actual advancement speed. However a reliable way to integrate it into the firmware beyond trivial go/no-go detection has not been found. Once you integrate corrective measures such as negative feedback, the system becomes fundamentally sensitive to noise and loses stability, because of control delay. Solutions to that include PIDs and neural networks. Problem with PID is its narrow operating range. Problem with neural networks is that it takes a long time to get useful results out of them, computationally intensive, highly opaque, a scary option all around. Any intermediate solution trying to combine benefits of both is likely to exhibit the drawbacks of both to a degree, and it's too risky for a business anyway.
Newbie maker here, this was very helpful. I've been wasting so much time on small prototype prints, it's getting disheartening. Thanks for the tips. Subbed!
Use a bigger nozzle and it’s even faster. I do all my prints with a 0.6mm nozzle, 0.32mm layer height, 3 perimeters, and 5% infill. Also, a bigger nozzle extrudes a thicker continuous line width, and results in better mechanical properties.
I've done a lot of printing with a 0.6mm nozzle, and I love it. A big nozzle works really well with flexible TPU prints. Some things can be printed in vase mode but still be nearly impossible to tear apart at the layers. Additionally, because the bigger nozzle lets you cover more area, it helps make up for the slow print speed required by TPU.
holy shit these settings are incredible... I'm making fishing rod holders to keep my dad's truck organized during his annual trip, and this literally doubled my productivity... what was once a 5 hour print became 2 hours 35 minutes, and because they're simple with no fine details the quality remains virtually the same. The only thing I changed was keeping the walls at 3 layers, because those surf rods are heavy and will be jostling around a bit.
I started printing with these settings, everything has been ok except one. My extruder motor is skipping every now and then, is this fine? the print is coming out ok but it doesn't feel right. never happened to me before. I already tryed heating up the nozzle a few degrees, which helped a little but it is still happing. any tips?
Thanks for including the settings for slicer. I used the most current version of Prusa Slicer, with the 3.5 firmware on the Prusa MK3 printer. I used PTEG (Amazon Basics) 1.75mm and the nozzle that came on the printer. I raised the nozzel temp by 5 degrees F. 235/245. The first layer was a little sketchy so I turned the speed down to 40% and then after a couple of layers set it back to 100%. Seems to be working well. My project was 1day+ with the draft settings, and now its about 11 hrs. Thanks for the video.
Really helped speed up my printing. A short video on how to change the home setting in the Prusa MK3+ to the back corner would make printing really wonderful.
I printed a case for Raspberry Pi (model 1-b) at 0.3mm. I thought it was nice to work with the 'natural' grain of 3D printed parts instead of trying to hide it! :)
I agree Igor. For personal items I actually like the lines yet with a somewhat smooth finish. I feel like my piece has character and that I created an authentic piece. Now for things I want to sell. I had read that I need to make them injection mold quality for customers. I need to remember people not going consider less than quality as adequate. That's understandable when they are paying money for it.
I've also found that rotating the part around the z-axis decreases print time. Not 100% sure why it works, but my guess is that the stepper motors don't have to stop. If you print a rectangle aligned with the X and Y axis, the the Y axis has to come to a stop while printing along the X axis and vice-versa. Since your example print is pretty much the size of the printer, it wouldn't be an option for that print; however, try it with another model and see what happens. I tried it on a ridiculously small print and watched 14m27s go to 14m11s where a 45 degree turn around the Z axis is the only thing I did different. A 1% improvement on an 11 hour print would shave off 7-10 minutes. The more rectangular your model, the better the trick works.
Ryan Henning, to most people, anything with rapidly increasing gradient - f'(x), is "as good as exponential" Yes we know what the difference is.. Magnitudes.
@@PaulKozlovitch did you watch the video? He says by scaling up a model 200% it takes 8 times more time/material. He refers to this as "exponentially longer time". Nowhere in the video or in my comments did anyone suggest that volume and build time were exponentially related.
Even though it is 2020, thank god for this video. I tried to print a spool holder for my Tronxy x1. It took about 12 hrs with my normal settings, and this video helped cut this down to just 3 hrs!
The settings might be conservative for a Prusa printer, but when you have a cheaper one like mine, sometimes they can even be a bit much for it.. definitely worth considering what kind of printer you have
Dang! I've been doing all my infill at 50% because I just wanted it to be "better." But then, I'm a newbie. Can't wait to do some tweaking. Love your channel.
Depends on the part, but more wall/ceiling/floor layers generally is better for strength. CNC kitchen has some great videos on part strength testing and what settings have an effect
This saved me sooo much time on my build. I went from about ~50 hours in total parts down to ~24 hours. That's over 50% faster! Didn't save a whole lot on material but that's ok.
@@Nichealys His math is correct. Faster indicates you are comparing speed. 50 hours = 2% an hour speed 24 hours = 4.1666% an hour speed Speedup = (4.166 - 2.0) / 2.0 = 1.083 or 108.3%
Strength in 3D prints come from the walls. Infill resists compression/expansion, but for the most part it just provides a framework for the walls to be built on.
From my experiences, any parts that crucially need a layer height below .3mm are purely aesthetic, non-functional parts. I print essentially all parts on my delta at .32mm layer height, and am soon upgrading to a .8mm nozzle to go even higher.
Thank you. Still new. Have watched and implemented your Inkskape and f360 tuts. Easy to follow. New to 3d design, or any design on pc. I can draw on paper pretty weel, then try to copy on pc. Some handshake so my drawing on phone is pretty bad. Tried mouse drawing with inkskape, and it's equally shaky. Short of getting a drawing pad with a rough screen, cannot think of fix. Doing curves with aid of bezier is ok , I guess.
at the 1:32 mark you show a Raise 3D with I guess is a 1mm nozzle according to the title description (although it doesn't look like a 1mm nozzle from the amount of extrusion coming from it). I have one of these machines and I have NEVER seen my models shake, shudder and move around as much as your print showed and with such crappy results from a print. I will say it looked as if you were pushing the speed a bit too high for a 1mm nozzle. Maybe I am missing something but my Raise 3d prints amazingly well and I almost always set my speed to 150% of normal.
@@MakersMuse My apologies for my remarks. I made a grievous error and somehow, thought that it was a .01 mm nozzle. Your remark about drilling out a nozzle made me realize my error. Now I can see why it didn't look as good as it should.
you can "overprint" much further than one might think. the biggest extrusion i got to work with a 0.4mm nozzle were 0,6mm layers at 1.4mm line width, before things turned into squiggles. (which are interresting in theyr own right) the usefullness is situational and limited, but it's interresting to see how far you can push things, and big simple structural parts and vase mode objects print in no time.
Gotta say thanks! I used these settings, but limited all printing speeds to 100mm/s and 500mm/s acceleration for my PETG printing set to 260 degrees does a great job to quickly print things without it looking terrible!
Doubling the size of a cube doesn’t technically increase the print time by 8x as the surface area only increases by 4x and the perimeter takes longer to print than the infill.
Vitalik Gula surface area increases with the square of the scale factor. Since creating the surface is most of the print and only some time is spent on the infill, the change in print time would be between 4x and 8x but would only reach 8x with 100% infill. This is also ignoring acceleration which would decrease the print time even more
@@keco185 so surface is in ^2, i agree. If we do as you say 20x20x20 cube print 4 time faster than 40x40x40 one. But I say, that 20x20x20 cube print 4 times faster than 40x40x20 one. You are right, that it take diferent time to move Z axis and X(or Y) but it's even slower, so it will be more than 8 times longer.
I've been printing 0.3mm since watching your video - I'm mass producing some little jigs to hold LEDs in place for soldering at maker fairs. I was amazed to find my stock Anet A8 works with the cheapest filament at 0.3mm *and* 250% speed! Next I think I'll install the Drylin bearings I bought to quiet it down. P.S. the PEI bed accessory I bought was not good, but smearing a 50% mix of PVAc (white glue) and water while at 60ºC made a good bed - things don't slide and come off easily.
Your comment on temps is interesting. On my Flashforge Dreamer with PLA I've always extruded at 215 (runs at around 20-213) with a bed temp of 50. The only time this has been an issue is if I'm printing something that comes to a point- usually gets a twist in it due to not being able to cool down enough between layers. Also have found best adhesion by wiping bed with metho just before it heats up, but leave it wet & allow the bed temp to dry it off rather than trying to wipe it. Doing that I virtually never have to raft.
Holy Cow Brother!!!! That cut my 6 Hour print job down to 1hr 45 min. It's not beautiful, but it's functional and its just a Scotty Mount for my Kayak.
A lot of that ghosting issues can be taken care of if you use dampening. SURE, that's a bit of a test and trial situation, and that takes time to figure out but it's worth it. On stock settings (but 0 infill) my CR-10 (first generation) can print reliably at 150% on the dial with the obvious exception of the starting phase where i dial it to 20%. I added some weight, removed some other, and used a dampening rubber joint on the table. It has removed almost all vibration, regardless of how close quarter the motion is. It does seem to have issues with the standard nozzle, i use a stainless steel one (.4).
Creality is excellent bang for the buck. Only probs I've seen with those is those damn warped beds. But once fixed up they print beautifully. Also under extrusion can happen but it's an easy fix as well. Basically they might need a bit of tuning sometimes.
Angus, you will gain MUCH MORE speed with a thicker nozzle. Line 0.6mm nozzle will save you about 50% of print time or so. All the things you said is also work. But for the prints not needing high precision thicker nozzle is the way to go!
Have you tried Klipper on any of your printers? It uses a Raspberry Pi to calculate the paths instead of any 8bit board. This is another way to get a better finish on fast prints, since 8bit MCUs like to skip some paths to save time and keep up with the workload. In very complicated models (say round bits) this makes older boards even stall for a second while the mcu tries to calculate where to go next. Klipper takes all those calculations off the board and just sends the solutions to the printer. Much smoother, even somewhat quieter since the movements aren't micro-stuttering at speeds over 100mm/s. Worth a shot on older/cheaper printers (if the frame is sturdy enough to cope with the forces of that chunk of metal being thrown around quite fast)
I'm also using klipper, but if the time estimation was correct for him, he didn't bump into such issues. And microstutters are only absent if you're using a virtual sd card.
@@istvanmaasz6771 at least i can print faster now without too much stuttering... not sure why you think there's no advantage, but after testing I'm very pleased with the outcome. It really does make a difference. 🤷♂️
@@therealpanse I also use klipper as I said, and it works fine, but for Angus, if he did such high speeds and the estimation in the slicer was correct, but means that the hardware was able to keep up, that's why I said for him (for a prusa with simple kinetics) it wouldn't make much difference. On a delta it does, on a corexy, also, since motions are more complex to calculate.
@@roolin2 I already checked it on multiple prints. Gyroid IS definetly slower then cubic/rectilinear, but not so much as honeycomb. If you want to go for speed only, gyroid is not the way to go. Though it definetly looks really nice and is pretty strong and versatile.
@@MakersMuse i have but it's 10 mins tops with a trimmer and 6mm distance bit mounted will you do a quality check on the printer (the fall out one)? i wonder how good it prints and if worth the time contra what you would get i dont have a fallout case but electronics could maybe be mounted under the base with some extra feet under it, or on the side
The fallout printer was a learning experience but certainly not worth the time investment for a good quality printer, more for the fact it exist and was a blast to design! You're much better off buying an established kit these days I think.
You were given a good name... a "Full Angus" sounds WAY more, take 'em to the "mattresses", impressive than a "Full Joel". I'd be much more nervous if I were told that I was going to meet Angus in the back alley!! hehe... good on ya!
Bondy. Where is the material science holdup... There are people out there who are doing this. (ps. it is merely a manufacturing technique, no need for any new technology.)
I'm surprised that all of the replies universally condemn the epoxy resin approach, but completely failed to suggest a workable solution, which I will now offer/propose: Low-expansion closed-cell foam as found in any home store, which can be trimmed easily using a saw blade or knife.
i might be crazy but i printed my project with 0,8mm nozzle at 230c with stock settings for 06mm nozzle from creality only increased the extrusion width and layer heignt to 0,5mm i also cut my print in parts so i can print one in under 3 hours and gyroid infill if i ever decide to fill it with resin the aded bonus was i cut in in places i could drop in a motor or other parts to be fully enclosed , a way that would be impossible if the model was solid all the way trough
For someone who wants to appear knowledgeable regarding 3d printing this video does not seem well informed at all. An E3D V6 will never be able to extrude 0.3mm tall and 0.4/0.48mm wide layers at 250mm/s. When you run those speeds you will push at speeds between 30 and 36 mm³/s when the hotend is only rated to 15 mm³/s. Also, rated speeds don't compare to real life and I can tell from my own and other people's experiences that the upper limit is around 11-12 mm³/s for reliable extrusion. This relates well to the 200mm/s infill at 0.15mm layers with optimal settings. The only reason you were able to print succesfully is because with those low acceleration values the printer will never reach 250mm/s or maybe just barely on some moves. You need 25mm to reach 250mm/s at 1500mm/s2 acceleration and also need that same distance to slow down again. So any move shorter than 50mm will not even reach that speed. Looking at your part with all the small infill sections I can image most moves being a lot shorter. Anyone who will try this on large flat parts will encounter massive under extrusion. Also, setting your travels to 170mm/s when you print with 250mm/s doesn't make sense. If you have the slightest bit of z-hop you wont collide with the parts at 170mm/s or 250mm/s. I run my travel moves at 1000mm/s with 12500mm/s2 acceleration and 0.15mm z-hop and never have any collisions. I have a demo video on my channel if you're interested.
I have no issues extruding over 35³mm/s with my E3D v6, just crank the temperature up and use quality infill. Also volcano exists if you need a lot more.
@@cthulpiss From the plastics poor heat conductivity, it needs to be heated through and through. One of the reasons a volcano hotend has a much longer melt zone.
1) its not sq mm/s, is cubic mm/s (mm^3/s) 2) you can do 20mm^3/s @240ºC on a all metal V6 if your fillament isnt cheap shit, am talking from experience. Faster and the plastic will extrude cold and wont stick. 3) yeah, his settings have lots of noobish mistakes. He doesnt mass print parts 24/7 for a living. 5) he didnt do 250mm/s in the infill but not because of the low acceleratio, the Prusa PLA profile itself has a mm^3/s cap set at 15mm^3/s, so anything faster gets slowed down to match that. It's usefull when you just want to set something to print as fast as possible or if you are tunning a part and didnt notice you were already at the limit the hotend could melt. 6) For comparison you can do from 22mm^3/s @ 218ºC up to 34mm^3/s @ 240ºC on an all metal vulcano.
I will certainly put to test your valuable insights but using gyroid infill. I think it looks fantastic and prints smoother without many accelerations and decelerations . Many claim that it offers a good ration between strength/material. Slightly better than cubic. Have you experimented with it at all? Edit: Grammar
@@MakersMuse It's awesome, and I find it's a lot stronger at low infills - be aware though, Slic3r PE can take a huge amount of time to slice gyroid if the perimeters are complex/bumpy. You'll also want to disable "Only infill where needed" on with gyroid - it increases the slicing time even further and adds weird splits that actually increase the print time....
Recently saw some published research explaining why a nearly-flat inner surface instead of conical has better heat transfer & extruding performance (the melting interface of the filament and the nozzle). I wonder how much mroe speed you could squeeze out if you could get ahold of a brass nozzle that's milled instead of drilled. I'd find all the info if you'd be willing to give it a shot in a video.
Cura has a cool feature called jerk control, it allows you to set the speed at witch the print head changes directions. At say 100mm/s the print head jurks when changing directions but you can make it change direction at say 8mm/s so it slows it down just before direction change and eliminates jerk. It hardly adds any print time to.
I just got my first we printer, and what surprised me the most was how incredibly strong the parts were, like for example yesterday I printed a small tower that is 10x10x50mm 50mm high, this is not an optimal orientation for layer adhesion and strength, it was also only 2 perimeters and 10% infill. I tried as hard as I could to snap it in half, however I was completely unable to do it with my hands alone.
I’m a new subscriber, so please forgive me if this question has been answered. Can you do a video on how you use interference fits to put pieces together?
I was playing myself on timesaving settings just recently, as I print a full size surfboard. I came along with very similar settings, only additional even lower infill at 6% to reduce weight even more.
That's what the Reprap project is all about! :-) Thanks to the open hardware Reprap, the 3D printing movement got possible in the first place. Before 2010, printer's for under $1000 were pretty much unknown of, unless you got cheap parts or stuff laying around, had the right machines and tools to make parts already, and knowledge of the Reprap project. Look on Thingiverse, UA-cam or other sites for the Snappy 3D printer, which is one of the mostly printed printers.
metacollin, very useful insights. You discuss significant rheological issues and good that you have. Sadly, the open source slicers are still quite entry level, I will follow up with your settings photo. Just watched a Univ of Michigan 'promo' video discussing acceleration control which is also in the correct direction but your points are very significant and hope that more get to read them. Congratulations, great comments. BTW, I have changed the E6-like extruder for a Volcano design to up the maximum volumetric flow rate. OK, this is a more advanced change to a 3D Printer. However, its not that difficult. My next change is to precompute the path profile for better flow control vs. acceleration. Have you considered that point?
I am not surprised it was a strong part because in some ways a coarser layer height increases part strength because there are less layers to break. The weakest area is in between the layers so less layers = less weak spots!
4 years later, PrusaSlicer has advanced quite alot and allowed for much more adjustability. Has anything changed in your settings on Prusa Slicer 2.5.0?
Angus,
I watch a lot of 3d printing videos here on youtube. And I just want to say: yours are the best! I've watched you for a while now and you always have good interesting ideas, low editing no jump cuts and they are just quality. Thanks for all the videos you make we out here in the 3d print community appreciate it and you might not hear it enough but these videos are critical to people like us. Sometimes i feel like you have a much more technical degree than just an industrial design degree, you're a smart dude! Be proud of the following you've created, and the good you've done for the community in general. I'm sure many others feel this way but for whatever reason have never taken the time to vocalize it to you. Cheers from California!
Hi, Thanks for sharing.
There is a real problem when doing "small perimeters": this induces vibrations because the speed is constant so is the spacial frequency. One solution is to use a no-constant speed. I've worked in the past on big CNC machines, and we had the same problem (obviously not at the same level) when trying to go too fast. One solution has been to do a random variation of the speed when the direction of machining (print in the case of a 3D printer) has changed of more of 90°. This way, without any mechanical dumper, we have some kind "software dumping" [tm] and the quality is greatly improved. The speeds have to be choose to depend on the spacial frequency and eingen modes of the machine.
Interesting idea! This is similar to the spread-spectrum techniques we use in communications to spread the power out over a range of frequencies rather than concentrating it all at a single frequency. This can help avoid creating interference with other signals and also makes the signal more tolerant to outside interference.
@@reverse_engineered The problem is when the acceleration is very high. This acts like a dirac pulse and excites the resonant frequencies of the whole machine.
Wow that's fascinating
Eigen modes… phew if only my teacher could’ve come up with a single reason we would need to learn them as students… and here we are!
Angus, one other thing you can do to add strength without adding time is push your extrusion width to .6mm - your top & bottom layers will print faster with fewer passes & you can get even more strength with just those 2 perimeters. If you normally print with Z hop & don't have any delicate shapes that may get knocked over or extreme overhangs that will curl up, you can remove Z-hop for even more time savings. The .6 width with a .3 layer height will not do overhangs past 45 degrees well - so be warned! Those curved undersides of your frame would benefit from a vertical "wall" of trimmable support (similar to what you did for the horizontal bridges).
One thing to be aware of when using larger nozzles, wider extrusions, and thicker layers is that you will eventually be limited by the volumetric flow rate that your hot end can provide. For a lot of cheaper printers like the Ender 3 clones, you won't be able to squeeze out much more than about 10-15mm^3/s. However, with better nozzles, heat breaks, and heaters, you can push this much further, with some going well beyond 30mm^3/s.
why is it mm^3?
@@barbarosbozkurt758 cubic (aka to the power of 3 which is what ^3 means) millimeters is a measurement of volume just like cubic centimeters or cubic inches, cubic feet etc.
I know that but I usually see the speed of the printer which is in mm^2, I didn't ask what it is, I asked why, again thanks for clearing it up@@negativenarwhals
I've been using similar settings on my MK3 for quite a while now, I use 0.32mm layer heights and get great results, print times easily twice as fast as the default 0.15mm Optimal and still way faster than the 0.20mm Draft as well. Figuring this out really helped keep the design hobby going for me - I was getting very frustrated with the iterative design approach and having to wait half a day for my part to print before testing.
Super useful tip. I mainly print organizers and inserts for board games so the quality sacrifice isn’t really a big deal and this cuts my print times by 50-60%. Thanks!
increasing line width a bit helps a lot too. set a 0.4 nozzle to like 0.8 width in the slicer does wonders
Anny, you didn't did the one and only thing that really makes the difference: to round all your sharp corners so machine won't have to slow down on every sudden direction change.
There is a lot of corners in your model that you don't need for anything.
Also I do extrude my PLA at 220-230°C and have zero issues with supports or bridges.
And it makes layer adhesion only better.
Also put on that 0.8-1.0 mm wide nozzle, this will make more difference than anything else if you are in a hurry.
Angus ,I don't own a Prusa but I found this very helpful in navigating the frustrating world of slicer settings. This 3d Printing hobby is turning out to be more like becoming an expert in a specific field of industrial engineering than just "printing stuff". I'm certainly not complaining mind you. I find it amazing that we at home can learn something so complicated. Now that is the true purpose of the Internet.
I have an Anycubic I3 Mega as I mentioned above.
I long ago settled on a 0.5mm nozzle and 0.3mm layer heights, and running PLA @210C for speed. Very seldom less than 3 shells, and usually at least 30% infill. My parts are strong and suitable for mechanical use with any minor blemishes irrelevant most of the time.
Hi Angus. The way your part is designed has a significant impact on the part's rigidity as well. I believe that that's why you couldn't break it apart. Your part design is sound, and will withstand the stress put on it. In fact, your part is so well designed that when you attempted to flex it you couldn't.
Thank you for shortening my 2 days of printing to 14 hours. ✅
This is perfect for me! I've got a lot of things I'm 3d printing for Christmas and these settings are just what I needed!
I can imagine your setting for 'Stupid Fast' is great for prototyping, but if you're in a real hurry, you will need to select 'Ludicrous Mode' 🚀
But then all your prints come out plaid and no one wants that! ;P
@@ghoast59 ahh, u beat me to it! !
Also, you'd have to buckle up to go ludicrous speed :)
Don't forget to close all the shops in the mall and shut down the 3 ring circus....
NO SIR I DID NOT SEE YOU PLAYING WITH YOUR DOLLS AGAIN SIR
The explanation of why it is way longer than you would think, for anyone who’s curious, is that if you double the size in all 3 dimensions of a 1in cube, it is then 2 in by 2in by 2in the area of a 2in by 2in square is 4in squared, so it increases by a factor of 4, likewise, the volume of a 2in by 2in by 2in cube, is 2x2x2, so 8in cubed, meaning the volume is increased by a factor of 8. The high amount of filament that needs to be extruded is the main cause of long print times, and the total amount of filament that needs to be extruded is directly proportional to the volume of the part. This still applies without 100% infill because you are still increasing the volume of the walls themselves. Hope this helps anyone who was curious! 😃
11:44 "Fooking paris man"
Yeah I went back and heard it correctly this time but man was I surprised for a bit
had to turn on subtitles for that bit lol
@@76aussieguy "For Comparison"
The Aussie accent coming out.
@@acestu I clearly heard Fuck Paris Man.
😂😂😂😂
Great video! One note: as size increases the volume increases cubically, not exponentially. It's important not to confuse polynomial and exponential orders of magnitude.
A trick that i like to take advantage of when i don't want to increase layer height is printing at 0.16mm layer height and doubling the infill and support layer height in Simplify3d. 0.32 is the highest you can go with a standard 0.4mm nozzle and that trick works wonders with lots of infill and support compared to printing everything at 0.2.
Just got back into 3D printing (today) after my first failed attempt with the EZ3D printer. I tried your settings on my new Ender 3 Pro and they came out great! I honestly think I could push it further, but with it working this good so far I will stop here for now. It's crazy how the price has changed so much since I first tried 3d printing.
Never go full Joel. Those highfives can be powerful.
I applied all your suggestions and cut a 7hr print down to just under 2. Brilliant! Thank you.
Thank you for making this video. it convinced me to try tweaking settings that aren't factory. I am often prototyping and don't need crazy amount of detail, but i do need speed. Just successfully printed PETG at a .3 layer height with a line width of .5 at 75 mm/s on an ender 3. What would have taken me 9.5 hrs at cura's standard quality .2mm layer height only took 5 hrs
hey bro could you share me you profile, i cannot find some settings in cura , thanks
Max travel speed 200 for prusa doesn't have any negative effects for me, it's pretty safe.
Also I change the min retraction distance to 3mm and remove a layer change retraction requirement, it makes prints much faster too.
Sweet! Good tips to push it even further. Cheers.
you can do 250mm/s but if your zhop is too low (
More tips if you are using Cura: use adaptive layer height so you can push the maximum layer heights a bit higher without losing as much detail on curves and holes, Cubic infill with several gradual infill steps (to leave it more hollow where it doesn't matter as much) without risking pillowing or gaps like going straight low % infill, and 10 or so degrees hotter temps. Pushing line width can make a big difference as well, just make sure not to exceed your extruder's max volumetric flow (usually 8-15 mm3/s on std all-metal heater blocks, half that for PTFE sleeved). Volcano with larger nozzle is better, obviously, but I tend to run volcanos on everything, regardless of nozzle diameter because the flow is just so much better and more consistent (and much better volumetric flow, up to 35 mm3/s on my setups).
Dude, you're a legend for just brazenly showing your settings and walking through them. I can't tell you how many UA-cam videos I've seen up until finding yours where it amounted to a nothing burger with the instructor just dancing around showing some hard settings. Instant like and sub!
Wow I just put in these settings into my TEVO Tornado and all I can say is WOW! I have to crank the temp way up to prevent a jam! Thanks Angus!
Awesome!
@@MakersMuse Can you post the part please want to print this thing
Hi Angus. You inspired me to try a stupid fast setting myself on my Anycubic Kossel Linear Plus. After some issue’s with bed adhesion and upgrading to PEI powder coated magnetic bed, I found good results with first layer and my own stupid fast settings (using Cura 4.8, 4.12.1 or Arachne Beta2) which I’d like your take on.
My layers are 0.32mm with 1st layer 0.3mm. I set my wall widths to 0.8mm using a 0.4mm nozzle. 1st layer Extrusion at 130% (for guaranteed bed adhesion) then back of to 100% extrusion for the rest. Hot end settings at the highest end of recommended filament heat range(I use 230°C).
Speed wise I’m ranging between 60 and 120 mm/s with acceleration at 2000 mm/s/s.
I was surprised that the 0.8mm line width extruded so well from a 0.4mm nozzle and resulted in very good layer coverage and layer adhesion. A calibration cube in vase mode at these settings strarted me off on using this setting as my own goto stupidly stupid fast larger model settings.
The 300% vase mode calibration cube (60mm x 60mm x 60mm with a 0.8mm wall width) using Curas ‘0.2mm Draft mode’ takes just over 2hrs to print at the default settings. This one takes 21 minutes. Not bad.
Rectilinear infill DOES add strength in all 3 directions. It adds the most strength in the Z, that's true. But adds significant strength in X and Y too. Maybe thats a nitpick. Good video 👍
this saved so much time thank you!
Also not just for test prints but loads of functional prints. i needed to print 240 time a print that normaly takes 13 hours now 3 hours each with still reasonable quality some slight imperfections but for a functional print perfect. this is done using ABS bdw.
Thanks Angus - and here are some things to maybe think about:
Most printer frames are pretty flimsy - so much so that a relatively light printer head moving around with really no impediment to it moving is often causing flex and resonance.
On the other hand, CNC router machines are swinging a full on router around and the tool resistance when cutting is very much higher.
So lets start by making the frame of the printer more solid and able to absorb any forces applied without flexing and/or juddering all over the place.
That theoretically will mean that you can jack the speed way up.
The days of printer frames being made out of acrylic are going to come to an end.
The next point may or may not be possible but here goes...
I have heard that it is possible to monitor the amp usage of a stepper to determine the load being experienced by that stepper.
If this is indeed the case then would it not be possible to monitor real time the loading inside the extruder?
If this is possible then would it not be possible to jack up the temperature as you have done and use a PID loop to monitor the "back pressure" on the extruder (high back pressure means that the filament is not being melted fast enough) and if that is the case, slow the movement of the head down to keep the extrude coming through the nozzle able to lay down the optimum layer.
Which basically means that the printer should always be moving the extruder at the maximum speed and because the machine frame rigidity is so much better, all the ghosting/artifacts etc should be greatly reduced and issues like layer adhesion should also be optimised.
An idea of how to achieve the "optimum variable speed" would be to have an Uno monitoring the extruder motor load and adjusting the speed percentage setting on the main machine driver board on the fly as needed.
If all this can be done then an optimal print will happen every time and the only thing that will change quality of the print will be the layer height setting.
Just an idea...
How about passing this along to Joseph? His guys should be able to make it all work - if it is possible ;>}
Most frames are already metal, only dirt cheap machines are acryllic (and even then there are cantilevers and Ender 3 for under 200 dollars that are metal). 20-series aluminium extrusions seem to be a new gold standart for 3D printers, and at least from personal experience (going from A8 to a custom frame with V-slots instead of linear rods) the frame rigidity skyrockets. A8 has almost vibrated itself to pieces (the screws were coming loose all over the frame), the AM8-custom can go over 100 mm/s without even reprogramming it. The age of acryllic frames is pretty much already over, and it won't be missed.
You screwed up few people's who are currently writing Patent with your public disclosure!
That is great for the Open Source community. Thanks!
MK4Duo optionally does thermal compensation for the heat lost as extruded material. It's physically based and easy to configure. I'll try and port it over to Marlin2.
Fundamentally temperature control overrides such an adjustment long term, but it can help avoid short term fluctuation in pressure difference.
I don't see the point of measuring extruder current. Basically you have two fundamental issues, slipback and grinding, which are related, because the motor doesn't feel the filament not being advanced enough. You can trivially avoid stalls just by setting drive current high enough, and you have to because extruder stall is pretty much never a good thing except in emergency. Extruder operates at holding torque, essentially near standstill, so Trinamic current sensing is not possible.
Prusa has a sensor identical to one you'd find in a laser mouse trained at passing by filament, so it can measure actual advancement speed. However a reliable way to integrate it into the firmware beyond trivial go/no-go detection has not been found. Once you integrate corrective measures such as negative feedback, the system becomes fundamentally sensitive to noise and loses stability, because of control delay. Solutions to that include PIDs and neural networks. Problem with PID is its narrow operating range. Problem with neural networks is that it takes a long time to get useful results out of them, computationally intensive, highly opaque, a scary option all around. Any intermediate solution trying to combine benefits of both is likely to exhibit the drawbacks of both to a degree, and it's too risky for a business anyway.
Thanks so much for all the great details. I've been printing for a year now and I feel like you just filled so many gaps in my knowledge!
To keep things accurate, I think we need a Full Angus to Full Joel conversion chart
Newbie maker here, this was very helpful. I've been wasting so much time on small prototype prints, it's getting disheartening. Thanks for the tips. Subbed!
Use a bigger nozzle and it’s even faster. I do all my prints with a 0.6mm nozzle, 0.32mm layer height, 3 perimeters, and 5% infill. Also, a bigger nozzle extrudes a thicker continuous line width, and results in better mechanical properties.
I've done a lot of printing with a 0.6mm nozzle, and I love it. A big nozzle works really well with flexible TPU prints. Some things can be printed in vase mode but still be nearly impossible to tear apart at the layers. Additionally, because the bigger nozzle lets you cover more area, it helps make up for the slow print speed required by TPU.
holy shit these settings are incredible... I'm making fishing rod holders to keep my dad's truck organized during his annual trip, and this literally doubled my productivity... what was once a 5 hour print became 2 hours 35 minutes, and because they're simple with no fine details the quality remains virtually the same. The only thing I changed was keeping the walls at 3 layers, because those surf rods are heavy and will be jostling around a bit.
I started printing with these settings, everything has been ok except one. My extruder motor is skipping every now and then, is this fine? the print is coming out ok but it doesn't feel right. never happened to me before. I already tryed heating up the nozzle a few degrees, which helped a little but it is still happing. any tips?
Tighten up the direct or bowden drive screw wich increases the pressure it grabs the filament and should hold better
Thanks for including the settings for slicer. I used the most current version of Prusa Slicer, with the 3.5 firmware on the Prusa MK3 printer. I used PTEG (Amazon Basics) 1.75mm and the nozzle that came on the printer. I raised the nozzel temp by 5 degrees F. 235/245. The first layer was a little sketchy so I turned the speed down to 40% and then after a couple of layers set it back to 100%. Seems to be working well. My project was 1day+ with the draft settings, and now its about 11 hrs. Thanks for the video.
Tried these values. Works as advertised. Thank you!
Cool! Thanks for letting me know :)
I pinged you on ig for advice.
Very helpful. Used the suggested setting on my Mk3s and got a great print in 6 hours that previously took 40 hours.
thank you!!! I just got an order of 160 parts, that are huge, and a very short time to do them in! This literally saved our life here lol
Really helped speed up my printing. A short video on how to change the home setting in the Prusa MK3+ to the back corner would make printing really wonderful.
I printed a case for Raspberry Pi (model 1-b) at 0.3mm. I thought it was nice to work with the 'natural' grain of 3D printed parts instead of trying to hide it! :)
I agree Igor. For personal items I actually like the lines yet with a somewhat smooth finish. I feel like my piece has character and that I created an authentic piece. Now for things I want to sell. I had read that I need to make them injection mold quality for customers. I need to remember people not going consider less than quality as adequate. That's understandable when they are paying money for it.
I've also found that rotating the part around the z-axis decreases print time. Not 100% sure why it works, but my guess is that the stepper motors don't have to stop. If you print a rectangle aligned with the X and Y axis, the the Y axis has to come to a stop while printing along the X axis and vice-versa. Since your example print is pretty much the size of the printer, it wouldn't be an option for that print; however, try it with another model and see what happens. I tried it on a ridiculously small print and watched 14m27s go to 14m11s where a 45 degree turn around the Z axis is the only thing I did different. A 1% improvement on an 11 hour print would shave off 7-10 minutes. The more rectangular your model, the better the trick works.
Do You realize... Your chanel is .... by far... the best 3D printing chanel on YT :D
#fact
Dude, this profile is mental. Printing me a small spool for the sample packs. It's crazy! This printer rules
Cubic, not exponential.
(At first, I thought "quadratic". But it is worse than quadratic, it's cubic. But not exponential.)
I was looking for this comment :)
Ryan Henning, to most people, anything with rapidly increasing gradient - f'(x), is "as good as exponential"
Yes we know what the difference is.. Magnitudes.
You're confusing the "exponential function" in mathematics with the word "exponential". A cubic function shows exponential (not linear) increase.
8x volume takes 8x time? Sounds like linear to me
@@PaulKozlovitch did you watch the video? He says by scaling up a model 200% it takes 8 times more time/material. He refers to this as "exponentially longer time". Nowhere in the video or in my comments did anyone suggest that volume and build time were exponentially related.
Even though it is 2020, thank god for this video. I tried to print a spool holder for my Tronxy x1. It took about 12 hrs with my normal settings, and this video helped cut this down to just 3 hrs!
When the video is so good that you can safely like it before watching 30% of it.
The settings might be conservative for a Prusa printer, but when you have a cheaper one like mine, sometimes they can even be a bit much for it.. definitely worth considering what kind of printer you have
Angus "Let's get started" Deveson
I just bought my first 3d printer and the information available on your channel is really great. Subbed so I can always find your channel.
Dang! I've been doing all my infill at 50% because I just wanted it to be "better." But then, I'm a newbie. Can't wait to do some tweaking. Love your channel.
Depends on the part, but more wall/ceiling/floor layers generally is better for strength. CNC kitchen has some great videos on part strength testing and what settings have an effect
I might be late but thank you. Now that i have my own printer this video helped me print stuff in general.
This saved me sooo much time on my build. I went from about ~50 hours in total parts down to ~24 hours. That's over 50% faster! Didn't save a whole lot on material but that's ok.
The rigidity is the same?
It's 108% faster not just 50%.
Carl Siemens Either quarantine has sizzled my mind, or your math is *wrong*
@@Nichealys His math is correct.
Faster indicates you are comparing speed.
50 hours = 2% an hour speed
24 hours = 4.1666% an hour speed
Speedup = (4.166 - 2.0) / 2.0 = 1.083 or 108.3%
@@pallavagarwal07 Yeah, not too sure what I was thinking.
Strength in 3D prints come from the walls. Infill resists compression/expansion, but for the most part it just provides a framework for the walls to be built on.
From my experiences, any parts that crucially need a layer height below .3mm are purely aesthetic, non-functional parts. I print essentially all parts on my delta at .32mm layer height, and am soon upgrading to a .8mm nozzle to go even higher.
delta + wide nozzle , that must go insanely fast
You know that useful parts can also look good?
I love your videos. I have been watching for a while trying to soak up as much knowledge as I could for when I finally got my 3D printer
My god he’s taught the machines to reproduce! The end is near!
It all started with the RepRap.
Thank you. Still new. Have watched and implemented your Inkskape and f360 tuts. Easy to follow. New to 3d design, or any design on pc. I can draw on paper pretty weel, then try to copy on pc. Some handshake so my drawing on phone is pretty bad. Tried mouse drawing with inkskape, and it's equally shaky. Short of getting a drawing pad with a rough screen, cannot think of fix. Doing curves with aid of bezier is ok , I guess.
at the 1:32 mark you show a Raise 3D with I guess is a 1mm nozzle according to the title description (although it doesn't look like a 1mm nozzle from the amount of extrusion coming from it). I have one of these machines and I have NEVER seen my models shake, shudder and move around as much as your print showed and with such crappy results from a print. I will say it looked as if you were pushing the speed a bit too high for a 1mm nozzle. Maybe I am missing something but my Raise 3d prints amazingly well and I almost always set my speed to 150% of normal.
Probably, it was an experiment where I drilled a nozzle to 1mm. Some prints worked great, others were terrible
@@MakersMuse My apologies for my remarks. I made a grievous error and somehow, thought that it was a .01 mm nozzle. Your remark about drilling out a nozzle made me realize my error. Now I can see why it didn't look as good as it should.
you can "overprint" much further than one might think. the biggest extrusion i got to work with a 0.4mm nozzle were 0,6mm layers at 1.4mm line width, before things turned into squiggles. (which are interresting in theyr own right)
the usefullness is situational and limited, but it's interresting to see how far you can push things, and big simple structural parts and vase mode objects print in no time.
Will there be files made available for that printer??
Awesome job
i also looking one. that found it?
me too……
Gotta say thanks! I used these settings, but limited all printing speeds to 100mm/s and 500mm/s acceleration for my PETG printing set to 260 degrees does a great job to quickly print things without it looking terrible!
Doubling the size of a cube doesn’t technically increase the print time by 8x as the surface area only increases by 4x and the perimeter takes longer to print than the infill.
He's not very strong at math
And it's 2 times taller, you forgot!
Vitalik Gula surface area increases with the square of the scale factor. Since creating the surface is most of the print and only some time is spent on the infill, the change in print time would be between 4x and 8x but would only reach 8x with 100% infill. This is also ignoring acceleration which would decrease the print time even more
@@keco185 so surface is in ^2, i agree. If we do as you say 20x20x20 cube print 4 time faster than 40x40x40 one. But I say, that 20x20x20 cube print 4 times faster than 40x40x20 one. You are right, that it take diferent time to move Z axis and X(or Y) but it's even slower, so it will be more than 8 times longer.
Vitalik Gula you’re thinking of a solid cube
I've been printing 0.3mm since watching your video - I'm mass producing some little jigs to hold LEDs in place for soldering at maker fairs. I was amazed to find my stock Anet A8 works with the cheapest filament at 0.3mm *and* 250% speed! Next I think I'll install the Drylin bearings I bought to quiet it down. P.S. the PEI bed accessory I bought was not good, but smearing a 50% mix of PVAc (white glue) and water while at 60ºC made a good bed - things don't slide and come off easily.
Any chance on the project files for that printer your building? I see reading below many have asked about it! Thanks for responding
Your comment on temps is interesting. On my Flashforge Dreamer with PLA I've always extruded at 215 (runs at around 20-213) with a bed temp of 50. The only time this has been an issue is if I'm printing something that comes to a point- usually gets a twist in it due to not being able to cool down enough between layers.
Also have found best adhesion by wiping bed with metho just before it heats up, but leave it wet & allow the bed temp to dry it off rather than trying to wipe it. Doing that I virtually never have to raft.
Holy Cow Brother!!!! That cut my 6 Hour print job down to 1hr 45 min. It's not beautiful, but it's functional and its just a Scotty Mount for my Kayak.
A lot of that ghosting issues can be taken care of if you use dampening. SURE, that's a bit of a test and trial situation, and that takes time to figure out but it's worth it. On stock settings (but 0 infill) my CR-10 (first generation) can print reliably at 150% on the dial with the obvious exception of the starting phase where i dial it to 20%. I added some weight, removed some other, and used a dampening rubber joint on the table. It has removed almost all vibration, regardless of how close quarter the motion is. It does seem to have issues with the standard nozzle, i use a stainless steel one (.4).
After a month of viewing your videos I finally got a 3d printer. I hope it doesn't have issues
Muhammad-Ilyaas Padia what 3d printer?
@@therollo9 Creality Ender 3
Creality is excellent bang for the buck. Only probs I've seen with those is those damn warped beds. But once fixed up they print beautifully. Also under extrusion can happen but it's an easy fix as well. Basically they might need a bit of tuning sometimes.
@@The_sound_Of_Thunder I only levelled my bed based on the corners and my prints stick well. *YAY*
& no underextrusion
Abandon all hope all ye who enter here
I have been watching you a couple years now i really liked your fallout printer. Thanks my friend
6:10 I swear, i can see your hair changing colors as your flexing to try to brake the part lol.
Super Saiyan Angus is among us
@@ValueTechTube and still isn't strong enough to break that part
Angus, you will gain MUCH MORE speed with a thicker nozzle. Line 0.6mm nozzle will save you about 50% of print time or so.
All the things you said is also work. But for the prints not needing high precision thicker nozzle is the way to go!
Have you tried Klipper on any of your printers? It uses a Raspberry Pi to calculate the paths instead of any 8bit board. This is another way to get a better finish on fast prints, since 8bit MCUs like to skip some paths to save time and keep up with the workload. In very complicated models (say round bits) this makes older boards even stall for a second while the mcu tries to calculate where to go next. Klipper takes all those calculations off the board and just sends the solutions to the printer. Much smoother, even somewhat quieter since the movements aren't micro-stuttering at speeds over 100mm/s. Worth a shot on older/cheaper printers (if the frame is sturdy enough to cope with the forces of that chunk of metal being thrown around quite fast)
I'm also using klipper, but if the time estimation was correct for him, he didn't bump into such issues. And microstutters are only absent if you're using a virtual sd card.
I'd not heard about this, sounds like it could be useful for my 8bit delta...
Pook365 for delta it’s a must, also very good for corexy types, but for simple cartesian, it doesn’t have too much advantage
@@istvanmaasz6771 at least i can print faster now without too much stuttering... not sure why you think there's no advantage, but after testing I'm very pleased with the outcome. It really does make a difference. 🤷♂️
@@therealpanse I also use klipper as I said, and it works fine, but for Angus, if he did such high speeds and the estimation in the slicer was correct, but means that the hardware was able to keep up, that's why I said for him (for a prusa with simple kinetics) it wouldn't make much difference. On a delta it does, on a corexy, also, since motions are more complex to calculate.
0:52 the square cube law at it's finest
Is the Printer that you printed on Thingeverse?
Can you share the link?
I want print this too.
That cubic makes sense, triangles and all that. Hoping to get a cheap 3d printer in the new year, so keep the good videos coming!!!!
Gyroid infill in the newest version is so nice
Yeah it's on the cards to test it out next week! Looks sexy :)
@@MakersMuse gyroid also looks amazing with top and bottom layers turned off.
its also bad if you dont have enough infill on a small part. Also its way slower than cubic.
Gyroid??
@@roolin2 I already checked it on multiple prints.
Gyroid IS definetly slower then cubic/rectilinear, but not so much as honeycomb.
If you want to go for speed only, gyroid is not the way to go.
Though it definetly looks really nice and is pretty strong and versatile.
Good to see the Aussie Mullet back in full swing. Takes me back to a childhood of watching Neighbours/Home & Away!
who has time for a hair cut lol
@@MakersMuse i have but it's 10 mins tops with a trimmer and 6mm distance bit mounted
will you do a quality check on the printer (the fall out one)? i wonder how good it prints and if worth the time contra what you would get
i dont have a fallout case but electronics could maybe be mounted under the base with some extra feet under it, or on the side
The fallout printer was a learning experience but certainly not worth the time investment for a good quality printer, more for the fact it exist and was a blast to design! You're much better off buying an established kit these days I think.
Link to this 3d printer stl? Would be a great gift to give someone that does not have a printer.
I want print this too.
You were given a good name... a "Full Angus" sounds WAY more, take 'em to the "mattresses", impressive than a "Full Joel". I'd be much more nervous if I were told that I was going to meet Angus in the back alley!! hehe... good on ya!
Can you just print a shell then pour in casting resin?
Quick: Yes, basically toolless "moulding".
Bondy. Where is the material science holdup... There are people out there who are doing this.
(ps. it is merely a manufacturing technique, no need for any new technology.)
Yes, someone on the internet did that with epoxi
interesting
I'm surprised that all of the replies universally condemn the epoxy resin approach, but completely failed to suggest a workable solution, which I will now offer/propose:
Low-expansion closed-cell foam as found in any home store, which can be trimmed easily using a saw blade or knife.
i might be crazy but i printed my project with 0,8mm nozzle at 230c with stock settings for 06mm nozzle from creality only increased the extrusion width and layer heignt to 0,5mm
i also cut my print in parts so i can print one in under 3 hours and gyroid infill if i ever decide to fill it with resin
the aded bonus was i cut in in places i could drop in a motor or other parts to be fully enclosed , a way that would be impossible if the model was solid all the way trough
For someone who wants to appear knowledgeable regarding 3d printing this video does not seem well informed at all. An E3D V6 will never be able to extrude 0.3mm tall and 0.4/0.48mm wide layers at 250mm/s. When you run those speeds you will push at speeds between 30 and 36 mm³/s when the hotend is only rated to 15 mm³/s. Also, rated speeds don't compare to real life and I can tell from my own and other people's experiences that the upper limit is around 11-12 mm³/s for reliable extrusion. This relates well to the 200mm/s infill at 0.15mm layers with optimal settings.
The only reason you were able to print succesfully is because with those low acceleration values the printer will never reach 250mm/s or maybe just barely on some moves. You need 25mm to reach 250mm/s at 1500mm/s2 acceleration and also need that same distance to slow down again. So any move shorter than 50mm will not even reach that speed. Looking at your part with all the small infill sections I can image most moves being a lot shorter. Anyone who will try this on large flat parts will encounter massive under extrusion.
Also, setting your travels to 170mm/s when you print with 250mm/s doesn't make sense. If you have the slightest bit of z-hop you wont collide with the parts at 170mm/s or 250mm/s. I run my travel moves at 1000mm/s with 12500mm/s2 acceleration and 0.15mm z-hop and never have any collisions. I have a demo video on my channel if you're interested.
I have no issues extruding over 35³mm/s with my E3D v6, just crank the temperature up and use quality infill.
Also volcano exists if you need a lot more.
Where is the limitation of E3D v6 coming from?
Mass of the heatblock? Power of the heater?
@@cthulpiss From the plastics poor heat conductivity, it needs to be heated through and through. One of the reasons a volcano hotend has a much longer melt zone.
@@goury I'm having trouble believing this. What settings do you print at?
1) its not sq mm/s, is cubic mm/s (mm^3/s)
2) you can do 20mm^3/s @240ºC on a all metal V6 if your fillament isnt cheap shit, am talking from experience. Faster and the plastic will extrude cold and wont stick.
3) yeah, his settings have lots of noobish mistakes. He doesnt mass print parts 24/7 for a living.
5) he didnt do 250mm/s in the infill but not because of the low acceleratio, the Prusa PLA profile itself has a mm^3/s cap set at 15mm^3/s, so anything faster gets slowed down to match that. It's usefull when you just want to set something to print as fast as possible or if you are tunning a part and didnt notice you were already at the limit the hotend could melt.
6) For comparison you can do from 22mm^3/s @ 218ºC up to 34mm^3/s @ 240ºC on an all metal vulcano.
Man this worked amazingly well. So glad I watched your video brutha
I will certainly put to test your valuable insights but using gyroid infill. I think it looks fantastic and prints smoother without many accelerations and decelerations . Many claim that it offers a good ration between strength/material. Slightly better than cubic. Have you experimented with it at all? Edit: Grammar
I'm keen to test it out next week once I update to the latest version.
@@MakersMuse It's awesome, and I find it's a lot stronger at low infills - be aware though, Slic3r PE can take a huge amount of time to slice gyroid if the perimeters are complex/bumpy. You'll also want to disable "Only infill where needed" on with gyroid - it increases the slicing time even further and adds weird splits that actually increase the print time....
Recently saw some published research explaining why a nearly-flat inner surface instead of conical has better heat transfer & extruding performance (the melting interface of the filament and the nozzle). I wonder how much mroe speed you could squeeze out if you could get ahold of a brass nozzle that's milled instead of drilled. I'd find all the info if you'd be willing to give it a shot in a video.
Is it possible to geht the stl Files of your Printer design ?
looks that it is not possible or have anyone it now?
Cura has a cool feature called jerk control, it allows you to set the speed at witch the print head changes directions. At say 100mm/s the print head jurks when changing directions but you can make it change direction at say 8mm/s so it slows it down just before direction change and eliminates jerk. It hardly adds any print time to.
"ludicrous speed, GO!"
I just got my first we printer, and what surprised me the most was how incredibly strong the parts were, like for example yesterday I printed a small tower that is 10x10x50mm 50mm high, this is not an optimal orientation for layer adhesion and strength, it was also only 2 perimeters and 10% infill. I tried as hard as I could to snap it in half, however I was completely unable to do it with my hands alone.
I love this project and can't wait to see your v2. Random question tho, what delta printer is that orange one in the background?
It's the 101 Hero. It makes a great prop, but made a terrible 3D Printer!
@@MakersMuse lol. Nice prop!
@@MakersMuse I disagree. It would look terrible on an airplane.
B Bowling I don't know why I laughed so much on this comment
Thank you!! This cut my 5 hour estimated print time in to 2 hours!
I’m a new subscriber, so please forgive me if this question has been answered.
Can you do a video on how you use interference fits to put pieces together?
I was playing myself on timesaving settings just recently, as I print a full size surfboard. I came along with very similar settings, only additional even lower infill at 6% to reduce weight even more.
Where I can found that project? I love that a 3dprinter print another 3dprinter...lol
That's what the Reprap project is all about! :-) Thanks to the open hardware Reprap, the 3D printing movement got possible in the first place. Before 2010, printer's for under $1000 were pretty much unknown of, unless you got cheap parts or stuff laying around, had the right machines and tools to make parts already, and knowledge of the Reprap project.
Look on Thingiverse, UA-cam or other sites for the Snappy 3D printer, which is one of the mostly printed printers.
metacollin, very useful insights. You discuss significant rheological issues and good that you have. Sadly, the open source slicers are still quite entry level, I will follow up with your settings photo. Just watched a Univ of Michigan 'promo' video discussing acceleration control which is also in the correct direction but your points are very significant and hope that more get to read them. Congratulations, great comments. BTW, I have changed the E6-like extruder for a Volcano design to up the maximum volumetric flow rate. OK, this is a more advanced change to a 3D Printer. However, its not that difficult. My next change is to precompute the path profile for better flow control vs. acceleration. Have you considered that point?
The full Joel would have taken care of it lol
Yup, Joel looks like he´s been spending more time at the gym, he has to be careful though, not to loose that natural "nerdness".
@@PaulojnPereira Buff nerds are still nerds! Fitness nerd.
I am not surprised it was a strong part because in some ways a coarser layer height increases part strength because there are less layers to break. The weakest area is in between the layers so less layers = less weak spots!
Can I get a CURA translation for all these settings? LOL :D
4 years later, PrusaSlicer has advanced quite alot and allowed for much more adjustability. Has anything changed in your settings on Prusa Slicer 2.5.0?