Somehow this reminds me of an overclocking competition. I can totally see it now. Buncha guys sitting around with their secret sauce hot ends and name brand steppers getting just those few seconds off of each print. I dont know what the analog to liquid nitrogen would be though. This thing is getting closer to imagining and item and the item appearing.
difference is, the overclockers stop at the screen. imagine how much faster we'd advance if we had a VR CAD program and a printer that can move this fast under detail.
it is going to be stronger motors. then 2 motors per axis running parallel. then instead of flimsy rubber belts its goint to be chains to withstand the hard pulls and then nozzles with 2 mm. diameter and a feedrate of 1200mm/sec. Benchy made in >10 min. (Just my opinion)
superconductor overdriven bldc motors. you could get better speeds with ballscrews I'd wager. but probably the highest speed is using a coarse and fine adjustment (very high speed gantry and a small 10x10mm one to compensate.) also, mechanical brakes would be a must on faster machines.
lol at the liquid nitrogen. I was watching it print and thinking the only way he's going to get good overhangs is with an ac/refrigeration system for part cooling.
What this seems really useful for is large parts, where even at these insane speeds the layers can fully cool in a reasonable amount of time. Seems like smaller parts have significant quality losses, but large parts could be printed extremely fast with no quality reductions, and maybe even stronger because of better layer adhesion!
I told you that you should be the President of the 3D printing Planet. I admire your perseverance and generosity of sharing your knowledge. Bless you and the best wishes to you from now on. Salutations from France.
@@MirageC +1 Another vote added, for president of all 3d printing. However unlike the nobel prize, it's just a honorary title. So no monetary prize i'm afraid we are all too busy spending our own money trying to follow these upgrades
Everything is sleek and lightweight. This thing demonstrates you can have excellent print speed with the right part cooling. Direct drive extruder weight is no problem if done right. Small hotend, low thermal mass. My bet is, some time in the near future we wont have heater blocks, but directly heated nozzle cartridges. Thick heavy engineering can be scaled down to the minimum required if understood right and be more efficient
A heater nozzle sounds just like what's being done with soldering irons now, the tip itself has the heater and thermistor/thermocouple built in. The issue is that it's considerably more expensive of course, but that's where the market enters, one might want the best performance, one might want a lower price. And indeed, the soldering irons with built-in heater in the tip perform amazingly. No cranking the temperatures up to high heavens to get the right one at the tip, which will eventually fall quickly with a low thermal mass tip. This makes me wonder how well would you be able to compensate this with good thermal compound between all the parts that need to be hot.
Hes got the local comressor cooling and custom coolers moumted the each side of the plate as well as you can see an extra fan in the reflection of the build plate this is now a limitation of the material not his cooling which is bonkers 😂
I originally thought that making hevacs produce more laminar flow would be good to reduce the amount of heat it takes from the bed, but now I'm wondering if it would be better to try and introduce vortices to encourage better cooling on the sides of the print that it isn't facing. Lots to think about!
Hey! Engineer here. Why not add a couple 300watt PTC air heaters at the bottom of the printer near the fan intakes to heat the fan air. The point of cooling is just to get the plastic under its glass transition temperature so it can accept the next layer and nothing more. Once you’ve got enough cooling to do that, any additional cooling on top of that is hurting you. The thermodynamic ideal you should be aiming at is to print one entire benchy as fast as possible, but to keep that benchy at a warm, uniform temperature that’s just below the plastic’s glass transition temp and not cooler. That way when the print is complete its entire thermal mass can cool down together and shrink uniformly. With HevACS in its current state you have the bed keeping the bottom layer warm, the hotend printing the hot top layer, and the HevACS desperately rushing to cool everything in between all the way down to room temperature. That hot-cold-hot gradient in the middle of a live print is adding to the artifacts on the finished model. This is why industrial 3D printers all use heated chambers and no heated bed. They’re basically ovens with corexy motion components inside. Try removing the berd air system all together and adding heaters to the fan intakes. Other than that I’d say all the only other source of print artifacts would be extrusion slipping and machine rigidity. Keep up the awesome work!! 🤟🏼
The thing is that the amount of energy that needs to be pumped out of the part could be done using two methods: If the material is warping sensitive, using thicker (more dense) air that is closer to the glass transition temp would be required in order to achieve the required cooling. For other non warping material like PLA, we can use cooler air (more delta in temp) to achieve the cooling. I had to add two additional fans to the build area in order to achieve proper cooling of the material during that print (see time stamp 5:19). The big quantity of material in so little time over such a small area makes it very difficult for only thin air to be cooling properly.
@@MirageC Ah interesting! So for such a small part, the most 'sensible' solution would be to print multiple copies of the small item, batch printing. Then for each layer, the print head will be going away to lay down filament for all the other parts. By the time it gets back to the original part, it would then have been left alone a much longer time waiting, and plenty long enough for that layer to have cooled. In that situation, you then will benefit from a heated chamber (for those warping filament materials), to give a more uniform base temperature. That is closer to (but still less than) the glass transition temperature. AND Ideally, the slicer should itself include an algorithm, to artificially add some extra delays for specific layers. But only for the smaller layers, which are too small to give enough time for cooling over a 'threshold amount'. This artificial delays would of course lengthen the overall print time. But only on those layers that actually needed it. So printing would still be fast, especially if you oriented the prints in an optimal way to minimize short layers. It could in fact be an optional feature, reserved only for those objects which you want to print the most accurately. And at other times you could just disable the feature, for the very fastest printing speeds.
@@dreamcat4 exactely, those are two strategy that effectively work. I wanted to show the cooling capacity by printing one unit. I struggled, but got it ok. As for the time of specific layers, Cura and other slicers can do this by several means. There is a specific minimum layer time parameter that you can use, but you can also use specific zone settings by using a support blocker and changing the settings for the overlapping portion of the model.
Would love to see this with some closed loop brushless motors, but you may need to re-spec your timing belts for the torque and consider a gear ratio that balances armature moment of inertia with the inertial mass of your print head. Also, you probably are already pushing your hot-end to its limit so it might not really add much benefit other than a smaller layer heights. That aside this really is an impressive machine and I can appreciate how much work must have gone into it
Wow glad i requested this one, that was sooo fast. Only some minior quality defects at that speed unreal! You would have to slow down for other plastics though? I can see PETG printing this fast but the others maybe not?
Dont get me wrong here. The printer IS super loud. Motors, Fans, Berd Air..... Sound was reduced for the video so your ears would not be hurt. You dont want that machine in the house ;)
I understand the curiosity and lessons learned from printing at superfast speeds, but for prints that need higher accuracy and good looks, what print speeds do you use? I saw your video about the part cooling tests and am curious how fast you can print a quality part with your setup and what you'd recommend as the ultimate part cooling solution. Merci beaucoup!
Instead of re-slicing the gcode to gain speed, I can set the Duet speed multiplier on the fly via the web board's web interface. This multiplication will be limited and impacted by Acceleration and Jerk settings. If the move is too small for the acceleration to reach the desired speed, then it will not. Other longer moves and travel moves will be able to get faster.
If you put a wood box form on the table 150mm tall and fill with cement as a base for the printer to anchor to it should give the mass needed to dampen the frame motion.
Why did you change from Zesty Nimble (in your early videos) to BMG Extruder, what are the disadvantages of a Nimble? Thanks for the video it is as always great.
Gearing ratio of the Zesty V1 (30:1) would not keep up with such speed and extruder acceleration. And the cable was creating ripple artifact. Great concept, but not adequate for this application.
How you get it to stick to the bed with that initial first layer printing speed? My bed is 400mm² aluminum and sanded glass and I have to spray hair sprayer over it. Power supplies are 2 One for the board running steppers Second for the heat bed only, voltage reduced to 22v and consumes ~30A it takes about 3 min to heat up but more 2min for the heat to reach to glass surface.
It's three times slower than the 6 minute Benchy some guy got with a delta printer, but it's probably ten times better quality, so I'd say this is a win.
How is the layers adhesion? I guess the part is more brittle than printed at lower speed? It would be interersting to print some samples, and send them to Stefan (CNC Kitchen) so he can test them on its traction machine...
I was watching the print, and a thought popped into my head. I thought I would bounce it off you. Occasionally I've seen people talk about the heat break getting too hot, and filament melting in there so that it clogs. What do you think about running an extra duct of air onto the heat break (if there's enough of it exposed), or on the bottom couple fins of the heat sink? Most printers I've seen have a fan for the heat sink, but nothing compared to the experimentation you've done with cooling. This may just be a solution looking for a problem that doesn't exist. Any thoughts?
That is mostly a problem for very low extrusion rates, since the filament moves slow and the heat creeps up, but he is extruding so damn fast that he would probably be all right even without the heatsink
With the flow of air from the HevACS and the moving speed no fan is required on the heatsink. Just need to manage preheat and cool down sequences. Vlad from 3dPassion recommended that i try this with the Nova HotEnd. Works great ;)
Super cool and fast. I'd love to see a 0.4mm nozzle though. Just doing a wider nozzle takes off a ton more time than boosting the speed. I would also love to see how fast you can print with a 1.2mm nozzle. Watching the print, it seems like it was having serious problems bridging. Have you tried using the bed cooler you have with the berd air or a 5015? Might help to have the fan blowing closer to target the location you're extruding specifically, and also the bed cooler to cool the entire print. Also with the top layers, it seems like maybe the flow rate and infill overlap is a bit too low?
The main issue comes from the fact that this hotend is significantly shorter than what I was using (super volcano) for that print head. This has made the bed to come up by at least 20mm. Completely blocking off the side ducts of the HevACS (th-cam.com/video/65FVQ1jArME/w-d-xo.html) for the first few layers, and then skewing off the intended flow angle. I usually have them pointing downwards at the bed for better internal cooling. I am creating a new head and revised air nozzles to address some of that.
@@MirageC yeah it's always tricky changing configurations coming with side effects like that. Why the change in hotend? I'm guessing the heater cartridge is much more powerful, but with a smaller melting zone? Assuming you could get a more powerful cartridge for the super volcano, wouldn't that have even better flow rate? I suppose that comes at a cost of weight. Very interesting to hear about and see the potential tradeoffs you're having to deal with at these speeds!
@@lllllllllllillllllll Super Volcano is powerfull, but it is heavy machinery. Plus the size of the melt zone (diameter and length) makes it less reactive on smaller nozzles which require higher pressure. Super Volcano also requires very slow retraction in order not to break the fusion bath and let material free fall form the nozzle during retraction. It weighs 100G and it creates quite a momentum on the heatbreak and carriage. Nova HotEnd is at the complete opposite of that spectrum. Flow comparision are actually not that different at smaller nozzle sizes. I am planning a flow race testing between the 2 this weekend.
@@MirageC that makes a lot of sense, thanks for filling me in. Didn't realize there would be a difference in retraction, I look forward to further testing/comparisons
Do you ever list everything for a build ? I been looking for kits, whatever info I can find on HevORT and it all points to video's and a community I can't find. Secret squirrel society ? I'm ready to pull a trigger and would love to build my own. However not enough knowledge to piece it myself. I build crazy PC's, liquid cooled mining rigs... I'm ready for 3D printing ! Some direction on all the parts ?
@@MirageC Ok, so from scratch ? Or do you recommend a base unit to build off of ? I see parts being printed, so I'm wondering if I should start with something cheap Ender 3 V2 to print parts with ? Just don't want to waste money in the wrong direction. Love the work your doing, and thanks for the reply. I know it's not easy with only so many hours in a day :)
Let's get this speed dialed in to perform. There are 2 things preventing this technology from enabling even the layman to get off the planet. One is speed, the other is material. VegOilGuy found a way around material; you just might give us the last key to common space travel.
Get Klipper on that beast and tune resonance compensation and your ringing will be gone. Probably use something a bit more powerfull than a RPI for ultra highspeed printing though. even with mainsail.
I use a Duet control board and the Dynamic Acceleration Adjusmtent (M593) duet3d.dozuki.com/Wiki/Gcode#Section_M593_Configure_Dynamic_Acceleration_Adjustment
Hey @MirageC I was just wondering if those ClearPath Servo motors included closed loop on the spindle? I am curious if closed loop is beneficial at these speeds, and also if you had seen the closed loop controller chip by Trinamic: it is the TMC4361A chip. There is dev board for this product, and Trinamic says it's compatible with the TMC5160 drivers. However I also am aware that there is no commercially sold 'ready made' product based on these 2 chips. And that you are using Duet boards... it just seemed like a worthwhile question to ask you. Since you have to also spend $300 x2 = $600 in order to purchase those ClearPath servos. It just seems like the cost of a close loop system would not add too much more to that investment cost (relative to the cost of the motors itself). If the ClearPath is or is not closed loop already? BTW over in the world of CNC machines, those guys seem to like to use motors from StepperOnline, for example the "Nema 23 Closed Loop Stepper Motor 3.0Nm/425oz.in Encoder 1000CPR" is a lot cheaper than the ClearPath. So how suitable would those be? Aside from the 1.8 degrees. How much worst performance? Because they seem to be a lot cheaper from stepperonline. Which is why the CNC guys seem to like them. I just was wondering if you knew about them / had already researched them for your HevOrt platform. Perhaps they also offer closed loop controllers too. Are they too slow, being meant for CNC application instead? IDK. Just that I cannot afford those ClearPath stuff... it's too expensive for me to justify! BTW thank you so much for the recommendation for hot end. Now that is something which I can afford. Thanks so much for all your research, all these investigations. It really is amazing work.
You have to understand that a Servo moto absolutely needs encoder to function. Servos are unstepped motors. They have no holding toque. The only way they can hold position is by contanly compensating according to the encoder reading. The clearpaths have all of this closed loop system integrated to them. So you feed them requests as you would with stepper motors. Now the whole concept of closed loop with stepper motors on a 3D printer should not be used to improve performance. It is only a crash recovery safety feature. If you go to fast so the stepper skips, then closed loop will slow down the print, recover the steps and move on... no gain. A proper selection of the motor in function of its torque and speed capacity is the way to go. NEma 23 are strong but slow by design. The coil diameter makes their inductance value higher as well as their inertia. If you want fast with high torque past the capacity of Nema17s, Servo are the next logical upgrade.
@@MirageC Thanks so much for explaining this! It really helps to understand better the component choice upgrade path for the steppers and drivers. I also noticed that these nema 17 you use are less than 2 amp. My own question is whether the extra expense of buying TMC5160 drivers is actually worthwhile here... Because it seems TMC2130 can do 2 amps which should be enough for these e3d 'high torque' nema17 at least. But the next upgrade looks to increase the driving current required for the ClearPath Servo. As the only next worthwhile upgrade for the x-y motors. The ClearPath ones you have listed on github, in the BoM are... 'CPM-SDSK-2321S-RQN'. I am looking at their product page and it says '173 w' constant power, when driving them at 75v. But only '39w' for driving at 24v... this then makes me wonder if it can be driven at a higher voltage above the usual 24v, and with tmc 5160 being worthwhile / most suitable component choice to drive them? It would seem that putting TMC5160 onto a 2nd power rail to the rest of the printer's electronics might give more juice? Because the datasheet of the tmc5160 stepper drivers says it can go 'up to 60v' and deliver 20a of peak coil current. But I then wonder if there are other factors and printing considerations, which will be limiting the speed below those maximum ratings.
Are the steppers closed loop if not have you had any troubles with layer shift, and do the fans stay level with the bed? Lol and lots of oomph from the glitter 🤣
Simple nema 17 0.9degree are used in this video. They are High Torque version from E3D. No closed loop system involved. No layer shift either. Jerk, Accel and motor amp were tuned to avoid Jerk while maintaining good performance. The side fans are stable to the current print layer. Bed goes down as part is being built.
@@tcratius1748 I Dialed in 160mm/s on the slicer, but cranked the printer's controller to 280% of that speed. So true speed was what ever the printer could accelerate to ;)
@@MirageC sure beats my A10M, it was slipping so I increased the torque and lost speed significantly, takes about double what Cura says. Crazy what you learn as you go and wish you knew earlier. I think you have a good product that could be massed produced via kickstarter. Cheers for the replies. :)
If you are trying to shave grams off the weight of the machine you can buy ready made carbon fibre box section to mount your rail to instead of the heavy aluminium extrusion. In the UK I get mine from easycomposite but there should be something similar in your country.
I have fast steppers that can rotate up to 50rpm same as hybrid stepper servos, but the cooling and the hotend is my bottleneck Duet2 wifi CORE XY acceleration X5000 Y4000 FR 333mm/s Dual color & 16mm ball screws for Z those alone costs me $75
How would to more motors shorten the belt length? Unless you are referring changing type of gantry completely. For Sherpa and Orbiter, yes they are on the list!
@@speadskater You must be referring to CroXY. CroXY has same total amount of belt length. These have more in quantity and they are smaller. But they have 2 more rails and 1 more Cross member. As well as two more motors.
The cause of the watch rattling might be affecting print yes. I now need to anchor the printer like a CNC would be anchored. Moving the print head mass quickly into one direction pushes the printer into the other one..... causing vibration.
4:20-20:23 =16 minutes and 3 seconds. a little bit slower than nero 3dp on his voron with 14 minutes :D but your print is way better quality especially on the chimney. nice printer and nice video ! :)
@@MirageC great idea. my setup is a bit different to yours, running a mosquito magnum (new version) a BMG with 25mm NEMA 17, all genuine MGN 12 rails on a diy corexy, 310x310x350 build volume. cooling wise I guess it's very much dependent on the filament used, some PLAs are pretty finicky where others perform well hot on hot. what actually caught my attention though was your servo drive attempt. I'm going to try a odrive with brushless out runners, still loads of firmware tweaking though since I want to use a native interface instead of step/dir, cause I hope that'll eliminate some of the positioning errors introduced by the step/dir latency.
While I really admire your quest against print time by building this monster, I wonder if the material can take the abuse. Try "recrystallizing" those fast prints against some slowly printed ones in a oven. Would be interesting to see if they have divergent warping characteristics.
Cooling the part by sinking the heat into the part rather than cooling the filament being deposit is keey in getting stronger parts while printing fast as per my observations.
@@MirageC This is just such interesting topic and I'm not aware of much research about this. You have a really capable machine maybe I can nudge you to do some tests about layer bonding and internal stresses depending on print vs cooling speed? It would benefit the entire community. I experimented a bit to dial in some filaments but I've always done that by gut feeling, so a bit more organized testing would be appreciated 😃.
Fairly comparable however the hevort has 1 thing that allows non planar printing that voron doesn't, the 3z bed with linear rails. So once a slicer is out that can utilize 3z then it will be able to do that and do high speed. Other than that they have theoretical same top speed within constraints of frame vibration
@@animeandrew18 there has been a lot of talk lately in the Voron community, a triple Z version of the VORON1 is coming soonish, hopefully it’s as cool as it sounds
@@MirageC thats great but how? How many grams did you save with your changes with hotend , extruder and rails? Because in 800mm/s video , hevort was missing steps at even lower accelerations than 20000
@@furkanbicer154 The main answer is in the Jerk! That thing is nasty! get it as low as you can and pump the accels! but I am also over 125g lighter with this config than using Hemera / Volcano / MGN 12.
It would be really cool if the big youtubers made it a competition for fastest benchy or something. I think it would create a lot of innovation in the 3d printing world. Also, what was the print speed accel and jerk?
Impressive Benchy bro! How do you effectively cool the part? You got the airflow over the whole print surface, but is that additional compressed air blowing directly under the nozzle? Edit: Okayokayokay. Had a look at your last video, which is great btw. You earned a sub. This in very interesting. I always thought that having that much air blasting over your heated bed might end up warping the part or at least give the heatbed itself a really hard time keeping up a certain temperature. Im currently working on retrofitting a coordinate measurement machine (1.3ton) to an FDM 3d printer (and also some more stuff). Its going down in the next weeks. If you want to have a look at it and maybe leave some information about hotend/cooling design, that would be awesome and highly aprechiated! Just check the vids on my channel. Since I'm on granite and airbearing I do have compressed air anyways. btw, thats also how I found your video in the first place :D
I have yet to try material prone to warping yet. My theory is that I could use hot air to manage the temp delta. The HevACS system relies on sinking heat in the part. This provides better bonding of layers while still cooling efficiently. I want to try fast printing on ABS see if this would still work. I will go and check your channel now :)
Nova hotend seems like a great hotend, but the creator isn't exactly the nicest of people, he threatened to sure someone for saying that the Mosquito heatbreak had a better surface finish.
@@MirageC That's fair, just felt like I should say my experience with him. Don't exactly know why. Anyways, congratulations on this fantastic print! Amazing seeing a FDM 3D printer print this fast!
First of all, non linear 3D printing or some name like that, that thing where it prints an outside curves in vertical rows instead of layers. Second, why printer angery? T.T
@@un-review I'll have to make one. You can find me on Facebook by my name here if you want, TH-cam comments isn't really the best way to keep in touch.
@@un-review I just realized 1) it was at higher layer height like this one and 2) it was a very specially tuned file, lots of tweaking to be able to cheat, like placing infill in strategic locations. So, I take back what I said. It is possible, but it won't be the same quality.
@@un-review I just ran some tests, basically, it all comes down to cooling. servos are unnecessary. I can print at 350mm/sec, 10,000 accel, 1500 jerk, 270C and get just as fast of a print, but it won't solidify.
Agreed with you. Steppers can outrun currently available hotend/extruder solutions on bigger parts. But I was far from reaching any close to the Nova HotEnd flow maximal capacity on this video. Even with more moving power, the next show stopper is cooling.... again! what about a raft screwed down the build plate, then waterfill the bed plate all the way close to the last layer. then add air... :/ Brain Fart! Stinky one! haha!
Hahah! You man! Lets make this model 4 times bigger and increase layer height. Then we can see where the chain breaks.... Printer or Hotend? :) Thank you so much for creating such a "hot" product!
Trinity: How did you do that?
HevORT: Do what?
Trinity: You moved like they do. I've never seen anyone move that fast.
HevORT: Wasn't fast enough.
That would be NEOrt ;) hahah good one :)
Meanwhile my hotbed is heating up
It's so quick, it almost looks like a time-lapse!
Somehow this reminds me of an overclocking competition. I can totally see it now. Buncha guys sitting around with their secret sauce hot ends and name brand steppers getting just those few seconds off of each print. I dont know what the analog to liquid nitrogen would be though.
This thing is getting closer to imagining and item and the item appearing.
difference is, the overclockers stop at the screen. imagine how much faster we'd advance if we had a VR CAD program and a printer that can move this fast under detail.
it is going to be stronger motors. then 2 motors per axis running parallel. then instead of flimsy rubber belts its goint to be chains to withstand the hard pulls and then nozzles with 2 mm. diameter and a feedrate of 1200mm/sec. Benchy made in >10 min. (Just my opinion)
superconductor overdriven bldc motors.
you could get better speeds with ballscrews I'd wager. but probably the highest speed is using a coarse and fine adjustment (very high speed gantry and a small 10x10mm one to compensate.)
also, mechanical brakes would be a must on faster machines.
lol at the liquid nitrogen. I was watching it print and thinking the only way he's going to get good overhangs is with an ac/refrigeration system for part cooling.
well i mean the analog to liquid nitrogen would probably be liquid nitrogen
What this seems really useful for is large parts, where even at these insane speeds the layers can fully cool in a reasonable amount of time. Seems like smaller parts have significant quality losses, but large parts could be printed extremely fast with no quality reductions, and maybe even stronger because of better layer adhesion!
🤣 I knew it! Of course you are going to go ahead and beat even the impossible 17min I forecasted... Congrats... Looking amazing..
Doesnt even seem real watching it print that fast. Very impressive @MirageC!
Show us vase mode! Should be amazingly quick at these speeds
Hasn't he done this alteady with 800mm/s? Or are you perhaps referring to circular vase? Yah, that would be insane.
I told you that you should be the President of the 3D printing Planet. I admire your perseverance and generosity of sharing your knowledge. Bless you and the best wishes to you from now on. Salutations from France.
Merci! :)
@@MirageC +1 Another vote added, for president of all 3d printing. However unlike the nobel prize, it's just a honorary title. So no monetary prize i'm afraid we are all too busy spending our own money trying to follow these upgrades
@@dreamcat4 hahah! Thank you, I am just having fun :)
Quick and impressive machine ! Nice work !
It’s not just that this is amazing it’s a lot of fun too!
Everything is sleek and lightweight. This thing demonstrates you can have excellent print speed with the right part cooling. Direct drive extruder weight is no problem if done right. Small hotend, low thermal mass. My bet is, some time in the near future we wont have heater blocks, but directly heated nozzle cartridges. Thick heavy engineering can be scaled down to the minimum required if understood right and be more efficient
A heater nozzle sounds just like what's being done with soldering irons now, the tip itself has the heater and thermistor/thermocouple built in. The issue is that it's considerably more expensive of course, but that's where the market enters, one might want the best performance, one might want a lower price. And indeed, the soldering irons with built-in heater in the tip perform amazingly. No cranking the temperatures up to high heavens to get the right one at the tip, which will eventually fall quickly with a low thermal mass tip.
This makes me wonder how well would you be able to compensate this with good thermal compound between all the parts that need to be hot.
Impressive!!
Seams as one side local cooling is not enough
Thanks for sharing👍😀
Hes got the local comressor cooling and custom coolers moumted the each side of the plate as well as you can see an extra fan in the reflection of the build plate this is now a limitation of the material not his cooling which is bonkers 😂
It's incredible!
Imagine prototyping on this beast ♥️
I originally thought that making hevacs produce more laminar flow would be good to reduce the amount of heat it takes from the bed, but now I'm wondering if it would be better to try and introduce vortices to encourage better cooling on the sides of the print that it isn't facing. Lots to think about!
Hell yeah man!
Hey! Engineer here. Why not add a couple 300watt PTC air heaters at the bottom of the printer near the fan intakes to heat the fan air. The point of cooling is just to get the plastic under its glass transition temperature so it can accept the next layer and nothing more. Once you’ve got enough cooling to do that, any additional cooling on top of that is hurting you.
The thermodynamic ideal you should be aiming at is to print one entire benchy as fast as possible, but to keep that benchy at a warm, uniform temperature that’s just below the plastic’s glass transition temp and not cooler. That way when the print is complete its entire thermal mass can cool down together and shrink uniformly. With HevACS in its current state you have the bed keeping the bottom layer warm, the hotend printing the hot top layer, and the HevACS desperately rushing to cool everything in between all the way down to room temperature. That hot-cold-hot gradient in the middle of a live print is adding to the artifacts on the finished model. This is why industrial 3D printers all use heated chambers and no heated bed. They’re basically ovens with corexy motion components inside. Try removing the berd air system all together and adding heaters to the fan intakes. Other than that I’d say all the only other source of print artifacts would be extrusion slipping and machine rigidity. Keep up the awesome work!! 🤟🏼
The thing is that the amount of energy that needs to be pumped out of the part could be done using two methods:
If the material is warping sensitive, using thicker (more dense) air that is closer to the glass transition temp would be required in order to achieve the required cooling.
For other non warping material like PLA, we can use cooler air (more delta in temp) to achieve the cooling.
I had to add two additional fans to the build area in order to achieve proper cooling of the material during that print (see time stamp 5:19). The big quantity of material in so little time over such a small area makes it very difficult for only thin air to be cooling properly.
@@MirageC Ah interesting! So for such a small part, the most 'sensible' solution would be to print multiple copies of the small item, batch printing. Then for each layer, the print head will be going away to lay down filament for all the other parts. By the time it gets back to the original part, it would then have been left alone a much longer time waiting, and plenty long enough for that layer to have cooled. In that situation, you then will benefit from a heated chamber (for those warping filament materials), to give a more uniform base temperature. That is closer to (but still less than) the glass transition temperature.
AND
Ideally, the slicer should itself include an algorithm, to artificially add some extra delays for specific layers. But only for the smaller layers, which are too small to give enough time for cooling over a 'threshold amount'. This artificial delays would of course lengthen the overall print time. But only on those layers that actually needed it. So printing would still be fast, especially if you oriented the prints in an optimal way to minimize short layers. It could in fact be an optional feature, reserved only for those objects which you want to print the most accurately. And at other times you could just disable the feature, for the very fastest printing speeds.
@@dreamcat4 exactely, those are two strategy that effectively work. I wanted to show the cooling capacity by printing one unit. I struggled, but got it ok.
As for the time of specific layers, Cura and other slicers can do this by several means. There is a specific minimum layer time parameter that you can use, but you can also use specific zone settings by using a support blocker and changing the settings for the overlapping portion of the model.
That's my first print in 60 mm/s :)
Holy s*** this is real time? Damn I got to finish mine.
bro, it's so amazing 🔥
Would love to see this with some closed loop brushless motors, but you may need to re-spec your timing belts for the torque and consider a gear ratio that balances armature moment of inertia with the inertial mass of your print head. Also, you probably are already pushing your hot-end to its limit so it might not really add much benefit other than a smaller layer heights. That aside this really is an impressive machine and I can appreciate how much work must have gone into it
Amazing speed. Is the copper tube air? Would you not need all around the nozzle?
This is impressive... And now: again please with the ClearPath Servos mounted!
Yes, I need to move on that project!
@@MirageC and what you thing about using a stepper motor with closed loop instead a servos? It should be cheaper, and have also good results
He had steppers on before making them closed loop would make them more reliable but would not improve acceleration. Those servos are in another league
@@nicholasjacob3594 stepper motors were used for that video. No servo involved ;)
@@MirageC wow I feel stupid I wrote the comment while watching and haven't seen that yet.
This is awesome. I am really impressed. Not saying me being impressed means much LOL but I am. WOW!
Wow glad i requested this one, that was sooo fast. Only some minior quality defects at that speed unreal!
You would have to slow down for other plastics though? I can see PETG printing this fast but the others maybe not?
Possibly faster. Materials like ABS don't like fan cooling.
The speed is awesome but what's unbelievable is how quite it is, super nice work on the rigidity.
Dont get me wrong here. The printer IS super loud. Motors, Fans, Berd Air..... Sound was reduced for the video so your ears would not be hurt. You dont want that machine in the house ;)
@@MirageC I mean no knocking or rattling, I'm sure it sounds like a jet with all the fans.
@@timmturner well... maybe it rattles... I just cover up the noise with fans! Hahah
I understand the curiosity and lessons learned from printing at superfast speeds, but for prints that need higher accuracy and good looks, what print speeds do you use? I saw your video about the part cooling tests and am curious how fast you can print a quality part with your setup and what you'd recommend as the ultimate part cooling solution. Merci beaucoup!
Hola, soy de argentina y me encantan tus videos
sounds like R2 got a speed and cooling upgrade
I love your work. Do you have an explanation video about 280% vs printer multiplier speed vs 100% but faster gcode settings?
Instead of re-slicing the gcode to gain speed, I can set the Duet speed multiplier on the fly via the web board's web interface. This multiplication will be limited and impacted by Acceleration and Jerk settings. If the move is too small for the acceleration to reach the desired speed, then it will not. Other longer moves and travel moves will be able to get faster.
The hotend goes brrrr!
Now replace the steppers with servos. You can get plug-and-play stepper replacements from Clearpath.
:) I know! th-cam.com/video/m6DoKoESPdg/w-d-xo.html
I would love if you could capture any 60fps footage of that beast print
940fps in the works ;)
If you put a wood box form on the table 150mm tall and fill with cement as a base for the printer to anchor to it should give the mass needed to dampen the frame motion.
Better to allow frame movement as is than anchor it at bottom. Top of frame (XY plane) is place where we want to add mass, maybe.
Why did you change from Zesty Nimble (in your early videos) to BMG Extruder, what are the disadvantages of a Nimble?
Thanks for the video it is as always great.
Gearing ratio of the Zesty V1 (30:1) would not keep up with such speed and extruder acceleration. And the cable was creating ripple artifact. Great concept, but not adequate for this application.
Ninja 3D printing haha!!
How you get it to stick to the bed with that initial first layer printing speed?
My bed is 400mm² aluminum and sanded glass and I have to spray hair sprayer over it.
Power supplies are 2
One for the board running steppers
Second for the heat bed only, voltage reduced to 22v and consumes ~30A it takes about 3 min to heat up but more 2min for the heat to reach to glass surface.
My bed is 340 x 340 using a 300X300 750W 120VAC keenovo heatpad. I use a mirror with a coating of Magigoo bed adhesive. This stuff rocks!
It's three times slower than the 6 minute Benchy some guy got with a delta printer, but it's probably ten times better quality, so I'd say this is a win.
I need this... how much do you think I could do this entire build for? Mad props, brother!!!
you can figure a 2000 USD for a build like that.
How is the layers adhesion? I guess the part is more brittle than printed at lower speed?
It would be interersting to print some samples, and send them to Stefan (CNC Kitchen) so he can test them on its traction machine...
Or build a traction machine to do my own tests :)
Could ringing on the side be due to your watch being too close and generating air flow turbulence on the right side hevacs duct ?
You know what, that is very possible! Needed a witness... i have the witness to site elsewhere next time. :)
have you tried to do it with the volcano ?
I was watching the print, and a thought popped into my head. I thought I would bounce it off you. Occasionally I've seen people talk about the heat break getting too hot, and filament melting in there so that it clogs. What do you think about running an extra duct of air onto the heat break (if there's enough of it exposed), or on the bottom couple fins of the heat sink? Most printers I've seen have a fan for the heat sink, but nothing compared to the experimentation you've done with cooling. This may just be a solution looking for a problem that doesn't exist. Any thoughts?
That is mostly a problem for very low extrusion rates, since the filament moves slow and the heat creeps up, but he is extruding so damn fast that he would probably be all right even without the heatsink
With the flow of air from the HevACS and the moving speed no fan is required on the heatsink. Just need to manage preheat and cool down sequences. Vlad from 3dPassion recommended that i try this with the Nova HotEnd. Works great ;)
Super cool and fast. I'd love to see a 0.4mm nozzle though. Just doing a wider nozzle takes off a ton more time than boosting the speed. I would also love to see how fast you can print with a 1.2mm nozzle.
Watching the print, it seems like it was having serious problems bridging. Have you tried using the bed cooler you have with the berd air or a 5015? Might help to have the fan blowing closer to target the location you're extruding specifically, and also the bed cooler to cool the entire print. Also with the top layers, it seems like maybe the flow rate and infill overlap is a bit too low?
The main issue comes from the fact that this hotend is significantly shorter than what I was using (super volcano) for that print head. This has made the bed to come up by at least 20mm. Completely blocking off the side ducts of the HevACS (th-cam.com/video/65FVQ1jArME/w-d-xo.html) for the first few layers, and then skewing off the intended flow angle. I usually have them pointing downwards at the bed for better internal cooling. I am creating a new head and revised air nozzles to address some of that.
@@MirageC yeah it's always tricky changing configurations coming with side effects like that. Why the change in hotend? I'm guessing the heater cartridge is much more powerful, but with a smaller melting zone? Assuming you could get a more powerful cartridge for the super volcano, wouldn't that have even better flow rate? I suppose that comes at a cost of weight. Very interesting to hear about and see the potential tradeoffs you're having to deal with at these speeds!
@@lllllllllllillllllll Super Volcano is powerfull, but it is heavy machinery. Plus the size of the melt zone (diameter and length) makes it less reactive on smaller nozzles which require higher pressure. Super Volcano also requires very slow retraction in order not to break the fusion bath and let material free fall form the nozzle during retraction.
It weighs 100G and it creates quite a momentum on the heatbreak and carriage.
Nova HotEnd is at the complete opposite of that spectrum. Flow comparision are actually not that different at smaller nozzle sizes. I am planning a flow race testing between the 2 this weekend.
@@MirageC that makes a lot of sense, thanks for filling me in. Didn't realize there would be a difference in retraction, I look forward to further testing/comparisons
@@lllllllllllillllllll I was surprised to see how precise the Nova is. Feels like a syringe.
Do you ever list everything for a build ? I been looking for kits, whatever info I can find on HevORT and it all points to video's and a community I can't find. Secret squirrel society ? I'm ready to pull a trigger and would love to build my own. However not enough knowledge to piece it myself. I build crazy PC's, liquid cooled mining rigs... I'm ready for 3D printing ! Some direction on all the parts ?
Have you seen this page: miragec79.github.io/HevORT/
Follow the steps not all is there yet but the community can help with the gaps ;)
@@MirageC Ok, so from scratch ? Or do you recommend a base unit to build off of ? I see parts being printed, so I'm wondering if I should start with something cheap Ender 3 V2 to print parts with ? Just don't want to waste money in the wrong direction. Love the work your doing, and thanks for the reply. I know it's not easy with only so many hours in a day :)
please make a tutorial to create this beast ....please
try using a mini ac unit for cooling
i would like to see how good in terms of quality this mashine can print instead of a faster and a faster and a even faster print :D
Let's get this speed dialed in to perform. There are 2 things preventing this technology from enabling even the layman to get off the planet. One is speed, the other is material. VegOilGuy found a way around material; you just might give us the last key to common space travel.
no time for retractions? :)
Get Klipper on that beast and tune resonance compensation and your ringing will be gone. Probably use something a bit more powerfull than a RPI for ultra highspeed printing though. even with mainsail.
I use a Duet control board and the Dynamic Acceleration Adjusmtent (M593)
duet3d.dozuki.com/Wiki/Gcode#Section_M593_Configure_Dynamic_Acceleration_Adjustment
@@MirageC Are you using S Curve acceleration as well?
Csn you provide more settings? how many perimeters? bottom top layers? what infill? what speed and accels?
Settings added to the video description.
impressive! material and temp settings? thanks
@@makevoid Purple glitter PLA at 200C.
If you print fast enough you never have to worry about bed adhesion... Print is done before the dang thing has a chance to move....
Hey @MirageC I was just wondering if those ClearPath Servo motors included closed loop on the spindle? I am curious if closed loop is beneficial at these speeds, and also if you had seen the closed loop controller chip by Trinamic: it is the TMC4361A chip. There is dev board for this product, and Trinamic says it's compatible with the TMC5160 drivers. However I also am aware that there is no commercially sold 'ready made' product based on these 2 chips. And that you are using Duet boards... it just seemed like a worthwhile question to ask you. Since you have to also spend $300 x2 = $600 in order to purchase those ClearPath servos. It just seems like the cost of a close loop system would not add too much more to that investment cost (relative to the cost of the motors itself). If the ClearPath is or is not closed loop already? BTW over in the world of CNC machines, those guys seem to like to use motors from StepperOnline, for example the "Nema 23 Closed Loop Stepper Motor 3.0Nm/425oz.in Encoder 1000CPR" is a lot cheaper than the ClearPath. So how suitable would those be? Aside from the 1.8 degrees. How much worst performance? Because they seem to be a lot cheaper from stepperonline. Which is why the CNC guys seem to like them. I just was wondering if you knew about them / had already researched them for your HevOrt platform. Perhaps they also offer closed loop controllers too. Are they too slow, being meant for CNC application instead? IDK. Just that I cannot afford those ClearPath stuff... it's too expensive for me to justify! BTW thank you so much for the recommendation for hot end. Now that is something which I can afford. Thanks so much for all your research, all these investigations. It really is amazing work.
You have to understand that a Servo moto absolutely needs encoder to function. Servos are unstepped motors. They have no holding toque. The only way they can hold position is by contanly compensating according to the encoder reading.
The clearpaths have all of this closed loop system integrated to them. So you feed them requests as you would with stepper motors.
Now the whole concept of closed loop with stepper motors on a 3D printer should not be used to improve performance. It is only a crash recovery safety feature. If you go to fast so the stepper skips, then closed loop will slow down the print, recover the steps and move on... no gain. A proper selection of the motor in function of its torque and speed capacity is the way to go.
NEma 23 are strong but slow by design. The coil diameter makes their inductance value higher as well as their inertia. If you want fast with high torque past the capacity of Nema17s, Servo are the next logical upgrade.
By the way, I am using simple Nema 17s in this video ;) No Servos were involved :)
@@MirageC Thanks so much for explaining this! It really helps to understand better the component choice upgrade path for the steppers and drivers. I also noticed that these nema 17 you use are less than 2 amp. My own question is whether the extra expense of buying TMC5160 drivers is actually worthwhile here... Because it seems TMC2130 can do 2 amps which should be enough for these e3d 'high torque' nema17 at least. But the next upgrade looks to increase the driving current required for the ClearPath Servo. As the only next worthwhile upgrade for the x-y motors. The ClearPath ones you have listed on github, in the BoM are... 'CPM-SDSK-2321S-RQN'. I am looking at their product page and it says '173 w' constant power, when driving them at 75v. But only '39w' for driving at 24v... this then makes me wonder if it can be driven at a higher voltage above the usual 24v, and with tmc 5160 being worthwhile / most suitable component choice to drive them? It would seem that putting TMC5160 onto a 2nd power rail to the rest of the printer's electronics might give more juice? Because the datasheet of the tmc5160 stepper drivers says it can go 'up to 60v' and deliver 20a of peak coil current. But I then wonder if there are other factors and printing considerations, which will be limiting the speed below those maximum ratings.
Are the steppers closed loop if not have you had any troubles with layer shift, and do the fans stay level with the bed? Lol and lots of oomph from the glitter 🤣
Simple nema 17 0.9degree are used in this video. They are High Torque version from E3D. No closed loop system involved. No layer shift either. Jerk, Accel and motor amp were tuned to avoid Jerk while maintaining good performance.
The side fans are stable to the current print layer. Bed goes down as part is being built.
Out of curiosity you said 160mm/s and 280%, what is the last part referring to?
@@tcratius1748 I Dialed in 160mm/s on the slicer, but cranked the printer's controller to 280% of that speed. So true speed was what ever the printer could accelerate to ;)
@@MirageC sure beats my A10M, it was slipping so I increased the torque and lost speed significantly, takes about double what Cura says. Crazy what you learn as you go and wish you knew earlier. I think you have a good product that could be massed produced via kickstarter. Cheers for the replies. :)
@@tcratius1748 Thank you for the kind words :).
Will that direct drive hold up for high temp pc printing?
For PC yes. but not for PEEK.
If you ever decide to sell one of these pre-assembled, let me know. We might be interested.
Or kits!
If you are trying to shave grams off the weight of the machine you can buy ready made carbon fibre box section to mount your rail to instead of the heavy aluminium extrusion.
In the UK I get mine from easycomposite but there should be something similar in your country.
yes, that is exactly what I have done for my second attempt. ;) th-cam.com/video/ZtpQwZyuZP8/w-d-xo.html
@@MirageC Lol, oops. I guess great minds do think alike!
I have fast steppers that can rotate up to 50rpm same as hybrid stepper servos, but the cooling and the hotend is my bottleneck
Duet2 wifi CORE XY
acceleration X5000 Y4000
FR 333mm/s
Dual color & 16mm ball screws for Z those alone costs me $75
50 rev per seconf i suppose. that sounds like a beast! Cooling wise have you tried side cooling? th-cam.com/video/65FVQ1jArME/w-d-xo.html
i cant believe my eyes. this is Sci Fi! Thanks for the watch. I would have thought it is time-lapse.
Are the clearpath servos running on this one?
No, simple E3D High torque Nema 17s 48mm 0.9deg ;)
Did you use the binder on glass bed？
Yes, I am using Magigoo bed adhesive.
@@MirageC Why not pei steel board?
@@user-vq2zl4zs4g yeah, what about it?
Next stop, 2 more motors for shorter belt length and a an orbiter/sherpa-mini extruder
How would to more motors shorten the belt length? Unless you are referring changing type of gantry completely.
For Sherpa and Orbiter, yes they are on the list!
@@MirageC one motor per corner instead of coreXY. It's less continuous belt, I guess it's more belt overall isn't it?
@@speadskater You must be referring to CroXY. CroXY has same total amount of belt length. These have more in quantity and they are smaller.
But they have 2 more rails and 1 more Cross member. As well as two more motors.
@@MirageC correct, a bit more weight for more control i think. One day I'd love to see or do a direct comparison.
Damn are you using input shaper?
I am using Dynamic Acceleration Adjustment from Duet. I need to give a try at Klipper one day.
Also I think the watch rattling on the bed could accidentally have caused the wringing or inconsistent layers.
The cause of the watch rattling might be affecting print yes. I now need to anchor the printer like a CNC would be anchored. Moving the print head mass quickly into one direction pushes the printer into the other one..... causing vibration.
Man that's the sound of nightmares
Yeah... its not a machine to rest next to your bed ;)
4:20-20:23 =16 minutes and 3 seconds. a little bit slower than nero 3dp on his voron with 14 minutes :D but your print is way better quality especially on the chimney. nice printer and nice video ! :)
I managed to get into the 12minutes, but now I need to try klipper. Machine is down for upgrades... will be back soon :)
not bad, my personal record is 21:44 but with 0.2 layer height. trying to beat your time with my next attempt with 0.24 layers😜
So much fun that race! :) I think we should come up with a set of rules and make it a true competition :) Cant wait to see your results :) Cheers!
@@MirageC great idea. my setup is a bit different to yours, running a mosquito magnum (new version) a BMG with 25mm NEMA 17, all genuine MGN 12 rails on a diy corexy, 310x310x350 build volume. cooling wise I guess it's very much dependent on the filament used, some PLAs are pretty finicky where others perform well hot on hot.
what actually caught my attention though was your servo drive attempt.
I'm going to try a odrive with brushless out runners, still loads of firmware tweaking though since I want to use a native interface instead of step/dir, cause I hope that'll eliminate some of the positioning errors introduced by the step/dir latency.
i feel this is the future of 3d printing
While I really admire your quest against print time by building this monster, I wonder if the material can take the abuse.
Try "recrystallizing" those fast prints against some slowly printed ones in a oven.
Would be interesting to see if they have divergent warping characteristics.
Cooling the part by sinking the heat into the part rather than cooling the filament being deposit is keey in getting stronger parts while printing fast as per my observations.
@@MirageC This is just such interesting topic and I'm not aware of much research about this.
You have a really capable machine maybe I can nudge you to do some tests about layer bonding and internal stresses depending on print vs cooling speed?
It would benefit the entire community.
I experimented a bit to dial in some filaments but I've always done that by gut feeling, so a bit more organized testing would be appreciated 😃.
Any thoughts on hevort vs voron?
Fairly comparable however the hevort has 1 thing that allows non planar printing that voron doesn't, the 3z bed with linear rails. So once a slicer is out that can utilize 3z then it will be able to do that and do high speed. Other than that they have theoretical same top speed within constraints of frame vibration
@@animeandrew18 there has been a lot of talk lately in the Voron community, a triple Z version of the VORON1 is coming soonish, hopefully it’s as cool as it sounds
20 000mm/s^2 ?? Are you using servo motors on this test?
Not even...! Those are E3D High Torque NEMA17s 48mm. I am amazed by what I can pull out from them.
@@MirageC thats great but how? How many grams did you save with your changes with hotend , extruder and rails? Because in 800mm/s video , hevort was missing steps at even lower accelerations than 20000
@@furkanbicer154 The main answer is in the Jerk! That thing is nasty! get it as low as you can and pump the accels! but I am also over 125g lighter with this config than using Hemera / Volcano / MGN 12.
@@MirageC thanks i will give a try!
It would be really cool if the big youtubers made it a competition for fastest benchy or something. I think it would create a lot of innovation in the 3d printing world.
Also, what was the print speed accel and jerk?
20,000 Accel and 800 Jerk. Added to the description of the video.
@@MirageC awesome, and GOOD LORD
Impressive Benchy bro! How do you effectively cool the part? You got the airflow over the whole print surface, but is that additional compressed air blowing directly under the nozzle?
Edit: Okayokayokay. Had a look at your last video, which is great btw. You earned a sub. This in very interesting. I always thought that having that much air blasting over your heated bed might end up warping the part or at least give the heatbed itself a really hard time keeping up a certain temperature.
Im currently working on retrofitting a coordinate measurement machine (1.3ton) to an FDM 3d printer (and also some more stuff). Its going down in the next weeks. If you want to have a look at it and maybe leave some information about hotend/cooling design, that would be awesome and highly aprechiated! Just check the vids on my channel. Since I'm on granite and airbearing I do have compressed air anyways. btw, thats also how I found your video in the first place :D
I have yet to try material prone to warping yet. My theory is that I could use hot air to manage the temp delta. The HevACS system relies on sinking heat in the part. This provides better bonding of layers while still cooling efficiently. I want to try fast printing on ABS see if this would still work.
I will go and check your channel now :)
Will you try to beat current under 8 min record?
Working on it! :)
Print it again at half speed and see if it looks flawless.
9 minute benchy here th-cam.com/video/A8-Ea4wwKtk/w-d-xo.html
At first glance I read "benchy rape", which, when you think about it, is not so far from the truth :D. Anyway, what an incredible printer. I love it!
Oops, a voron just beat you, time to bump it up 2 minutes.
I would be curious to compare time at similar quality ;) But I am happily working on a revised attempt! :)
lol I guess you have :P
Your SSL Certificate on Hevort.com is not valid anymore fyi
Thank you for notifying me. It should be fixed now.
Nova hotend seems like a great hotend, but the creator isn't exactly the nicest of people, he threatened to sure someone for saying that the Mosquito heatbreak had a better surface finish.
Can't really comment on that, Vlad has been super nice with me and others. Product delivers to specs and more, that's all I am asking. :)
@@MirageC That's fair, just felt like I should say my experience with him. Don't exactly know why. Anyways, congratulations on this fantastic print! Amazing seeing a FDM 3D printer print this fast!
this needs 60fps
First of all, non linear 3D printing or some name like that, that thing where it prints an outside curves in vertical rows instead of layers. Second, why printer angery? T.T
You don't need all that over power. I have a standard RailCore II and printed a 12 minute benchy at .2 layer height.
show us the video of that please, very curious.
@@un-review I'll have to make one. You can find me on Facebook by my name here if you want, TH-cam comments isn't really the best way to keep in touch.
@@un-review I just realized 1) it was at higher layer height like this one and 2) it was a very specially tuned file, lots of tweaking to be able to cheat, like placing infill in strategic locations. So, I take back what I said.
It is possible, but it won't be the same quality.
@@un-review I just ran some tests, basically, it all comes down to cooling. servos are unnecessary. I can print at 350mm/sec, 10,000 accel, 1500 jerk, 270C and get just as fast of a print, but it won't solidify.
Agreed with you. Steppers can outrun currently available hotend/extruder solutions on bigger parts. But I was far from reaching any close to the Nova HotEnd flow maximal capacity on this video. Even with more moving power, the next show stopper is cooling.... again! what about a raft screwed down the build plate, then waterfill the bed plate all the way close to the last layer. then add air... :/ Brain Fart! Stinky one! haha!
Why is everybody speaking of jerk in mm/s, isn't it mm*s^-3?
Duet Refers to Instant speed deviation more than real jerk. the "Instant" word removes a "/s" from the unit ;)
@@MirageC good to know but that's still mm*s^-2
@@unusualcomment9731 sorry.. i did not express myself right. Speed deviation is measeaured in terms of speed unit. Mm/s.
@@MirageC ok but that's still of a weird term if it's instantaneous acceleration
@@unusualcomment9731 instant speed deviation. Not acceleration.
Excuse me, WHAT?
What what? :)
@@MirageC What did i just witness?
Why so slow? :)
Hahah! You man! Lets make this model 4 times bigger and increase layer height. Then we can see where the chain breaks.... Printer or Hotend? :)
Thank you so much for creating such a "hot" product!
Awww you not showing the end result was probably the biggest anime betrayal of all tome
Not sure i am fully grasping your comment. Do you intend that i did not show the end result?? Did you watch the video?
10 out of 10 for speed, 6 out of 10 for quality.
Agreed!