Is a sub 1 min 3dBenchy possible? Hotend Build and First Print! (Episode 2)
Vložit
- čas přidán 7. 09. 2024
- This week we build the first iteration of a hotend that is capable of printing a sub 1 min 3dBenchy.
Feel free to post any question in the comments!
Discord:
/ discord
I appreciate how you grounded the mains hotend heater. Good job with safety. 😁
He is german, everything has to be followed by the rules! 😉 I'm one of this kind too.
@@tinom.2455 He is German buy says Aluminium the wrong way 🤣🤣🤣 (just joking)
This old Tony just released a video about gas struts explaining how they have constant force! I doubt you can find one with a low enough stroke or spring rate but it’s a possible option to overcome the spring rate issue. Best of luck, keep it up 👍
If this would be a thing, imagine how car engines would profit from it. But where are them? So.. no.. it wont work.
This is the first one of this guy's videos I have watched, but if the first comment is a fellow ToT enjoyer I will subscribe.
@@IFRYRCE this is only the second video I’ve watched from this channel but I think the dude definitely deserves a sub 👍
@@hanswurstusbrachialus5213 I have no idea what you mean?
@@jamieclarke321 Gas struts are way too slow. They are designed to dampen movement. If you think of your trunk lid, how it opens smooth but slow.
This has to be one of the most creative designs for a printer I have ever seen.
"spring" with constant force is a vacuum. you can machine airtight cylinders with pistons that you can use as springs, and force of the spring will be determined by area of the piston
@@IllSkillz I also suggested gas struts but like your thinking
Agreed. Except it is the atmosphere excreting the force, not the vacuum😉
I guess that is called damper
@@anonym9323 Only if you’re a biker😉 It is actually a spring (a “true” damper dissipates energy, while a spring passively pushes back).
@norWindChannel yeah you are right but I think a air cylinder will be modelled as a damper maybe I do not understand the concept enough that IllSkillz was talking about in his comment and my english is not that good
A few ideas:
1. try and get all four to join a single point, a cylinder. Please take a look at the Maslow 4 CNC for reference.
2. instead of going for a constant force spring, connect the two loose ends of opposing corners via pullies, and a single spring between them.
3. make the filament path much longer, but shallow. make two parts with the same thread pitch, one male and the other female, with the max diameter of the male equal to the min diameter of the female thread. essentially a helical path with plenty of surface area to transfer heat.
@@VANUSUAL Connecting the ends is a very good point. The spring in between only has to compensate for the movement "side to side".
There would be a drawback of coupling the two cables into one loop: you are doubling the length of cable, possibly causing issues with accumulating stretching and harmonics . But otherwise, probably a good idea
he said it in the coments in the last video if you pull it on one side the 90° get longer
This guy is crazy and I freaking love it. Seriously though, I’m in awe of your design and machining skills. Holy crap. Teach me, sir!
Hi, I commented on your shorts about 300mm³/s. I have designed a few hotends and am currently pushing 120mm³/s with a single filament. A couple of tips:
1.What you said about "too much puddle" inside a very long hotend is true, to an extent. If the internal passage is 2mm or less, there is not much problem, retraction still works (imagine a Syringe, it can bbe as long as you want). Only issue is if you sit still for a long time the whole hotend will leak out, which is a lot of filament if your hotend is 250mm long for example. What causes problems is internal voids larger than 2mm. I experimented with 2mm entry hole opening up into a wider chamber of 3 or up to 5mm before tapering back down to 2mm and finally 0.8 or whatever nozzle size, and this is a huge fail. Completely inconsistent, you send 20mm of filament into it and nothing comes out; so you send 10mm more into it and 30mm comes out, very bad. But if you maintain 2mm or even slight taper down 1.75mm or less, there is no big issue. I have tried this up to 200mm hotend length.
2. I think your hotend has way too much thermal mass. What you made is something from industrial extrusion, that could melt KGs of filament all day at consistent rate but that is not really appropriate for 3D printing. You built a jumbo jet when what you need is a rally car. A 3D printer needs to be "light weight" or "nimble" in terms of thermal mass. It needs to be able to go from 0mm³/s to 200mm³/s in the blink of an eye, and back to 0 just as fast. That giant mass of yours will take a long time to respond to changes in heating load. My hotend currently is 3 Zones in series (single filament), each is a 115W CHC heater, very light, almost no thermal mass, they respond instantly to change in demand. Within a few mS of sending a huge volume purge command you can see the output jump from 0% to 100%. The series multi-zone concept is also how slice Engineering are powering their Prime hotend, and also how industrial extruders work. If you stick with that much mass controlled by a single output you will see latent temperature drop many seconds after demand starts, by which time it will be too late and temperature will continue dropping even though your band heater is now dumping all of its beans into the heater block (and then it will overshoot and burn the filament).
3. You can get mica sheets for cheap that make excellent insulation between your cold end and hotend. Very light weight and easy to cut to size. I get mine on amazon.
4. Since your hotend is connected to your coldend by a huge bolt, you might be able to cut through the heatbreaks. I dont think you can eliminate them entirely but cutting a slot through them with a hacksaw would sever the thermal path and unless youre going to run 60A TPU i dont think the filament will find its way out through the cut.
maybee i dont understand it correctly but doesnt more thermal mass just makes it more consistent? so you dont need very much correction at all? while printing you dont switch tempretures
@@thomaskletzl6493 right, you don't switch temperatures, but the system still has to react to changing temperatures quickly. The filament draws heat out of the hotend when it melts, and the heater has to replace it. It takes time for heat generated by the band heater to conduct toward the center. A very fast change from 0mm³/s to 100mm³/s will result in a very fast drop in temperature of the inner walls of the hotend, and the thicker the hotend, the longer it will take to replace that heat.
I like to use analogies but I struggle to find a good one for this. The best I can think of is like a line of boats being towed on the ocean. You have a course plotted, a line, that is your temperature setpoint, and ideally you want all of your boats to stay on that line, but most importantly the last boat in line (the nozzle) needs to stay on the line. Your tug boat out front (your heater) needs to keep an eye on the line of boats and make adjustments to keep the last boat on the line. Temperature drops from rapid extrusions are like side-winds that blow the boats off course, and the angle/degree of drift, the difference between the plotted course line and the actual line formed by the boats, is like the temperature gradient caused by the delayed reaction between the heater and the nozzle. As long as there are no winds (no extrusion) the tug boat has no problem keeping every boat (every square mm of the heater block) perfectly on the line. As soon as stong side winds (rapid extrusions) come, the tugboat must veer off the line to bring the last boat back onto the line. The longer the line of boats (the more thermal mass of the hotend), the longer it takes for actions of the tugboat to affect the position (temperature) of the final boat, and the further the final boat can be blown off course before corrections are effective. If you have a long line of boats and cross-winds are strong but consistent (high volume constant speed extrusion like on an industrial extruder line), there is no problem. There will be a steep angle of divergence between the plotted course and the boat line (a dramatic gradient between heater temperature and bore temperature) but the final boat will track consistently on the line regardless. The problem comes with rapid start/stop and increase/decrease of the side winds. The tugboat just cannot react fast enough if there are a lot of boats in the line. A dramatic side wind comes and the tugboat starts (late, as usual) to react in an equally dramatic way, but then suddenly the side-wind dies off and the tugboat's reaction proves (late, as usual) to have been an overreaction as now the hotend is too hot and burning the plastic.
I made the best of these particularly inconvenient physics on my hotend by making my boat line as short as possible (low thermal mass) so that it can react quickly _and_ splitting the heating load across 3 separately controlled series heating zones, which is like having two extra tugboats evenly distributed in the boat line, so that pretty much every boat (every mm of the bore) can be held tightly to the course, controlled from both ends and not just from a distant tugboat at the lead.
@@charlesstaton8104 okay thank you for the explanatoin i get what you mean now. i did not know that it needed so much energy too melt it
@@charlesstaton8104 MPC in Bleeding Edge Danger Klipper may be able to assist in the issues you described
Thank you for your comment and effort. I never answered but I read and understood what you said. Be prepared for this weekends video, I got something for you ;)
This is a great project! Very interesting to see the failures, dont shy away from showing them as it allows us to learn from them together.
Keep going man, I subscribed and looking forward to the next one.
The fact it printed anything on first try is super impressive. I built a coreXY based on standard plans, and it was probably 20 tries before i got a print that made it to 80% finished 😂
Constant force spring you can try....quill spring from a drill press. Cheap and easy to find. Mcmaster Carr has lots of constant force springs. You can also find them in tape measures
@@ukulelefatman I think they are called power springs
Stumbled upon this video in my feed. Glad I did. I like your style and your passion. You, sir, have earned yourself a new subscriber. Keep up the great investigative work. The community is a better place with the kind of skills and ingenuity of people like you.
As someone who’s more of a casual viewer of the speed printing community, and as an engineer who dabbles in machining, it’s always incredible seeing these custom motion systems and wild hot end designs; I really hope I can see some of these developing tools in a package I can use in my work when I’m looking for a faster FDM printer someday!
This project is turning out really impressive! Super excited to keep watching this series.
failures are the building blocks of success. The way you challenge and provoke, the way you think out of the box, impressive. I look forward to the many twists and turns this project will provoke. Love your den.
For the tubes going between the hot end and cold end, you can buy medical thin wall tube in stainless for sure, maybe also in titanium from the right supplier. Saves you having to turn tiny tubes
Hey. Awesome Project. Regarding the constant-force "spring" system, I know it would be too much trouble but you could use a counterweight system through a series of pulleys and hang the weights at the edge of the table. This system would also be easily adjustable but definitely hard to implement.
Also, one reason for plastic oozing out of the hotend from heartbreak connections could be the different thermal expansion rates of Aluminium and SS causing a gap in those connections resulting in material oozing out due to nozzle backpressure. Since the hotend isn't mechanically constrained relative to the heatsink, It is allowed to freely expand and I think this results in a gap. One possible solution would be to heat-press them together.
To adjust the filament feedrate, one easy hack could be to use 2*1.75mm = 3.5mm filament Diameter in the slicer which would give you 4 times the surface area similar to running 4 extruders at once.
I also think that using a material like ABS with a higher melting range (around 100C) would allow you to increase the temperature of your heatsink without melting the material in there which would result in the filament requiring less heat to get to a printable state in the hotend.
Best of luck and thanks for sharing this process.
Yes, was going to suggest that too, using a pulley with a weight seems like an almost perfect option..though at those speeds the counterweight would need to be restricted in some what to not jump around, and if the acceleration is faster than gravity it could causes problems too maybe
@@dinissantiago6729 Yes. I completely forgot that we're working in very high acceleration regimes. Although you could still use a moving pulley system to increase the maximum acceleration but it would be so much work. you would need 4 pulleys for each axis to get 160k (mm/s^2) acceleration which I think is needed for a project like this.
One other good suggestion that I saw in another comment was to use another motor as a constant force spring for each axis. afaik TMC drivers have a constant force mode (which isn't utilized in Klipper or Marlin since there has been no use for it) but I think it requires some calibration.
It would be prone to oscillation...
@@at0mic282 Yes. I've been an idiot. This is basically adding mass to the bed but on the other side
What he need is basically a constant force that isn't a function of length, speed, or acceleration. all I can think of at this point is either a constant force spring with a large load value and long life cycle (typically in 10k extension range) or an extra motor with active feedback loop to maintain a constant force on the cord wich will have it's own challenges.
yeah, a counterweight system would be good. As for the inertia, using a gear ratio squares it, so if you have a 5 to 1 mechanical advantage, it'll only add 1/25 the inertial of the counterweight mass to the system. And, using pulleys, there's no reason it has to hang off the table. They can just live off to the side.
What about leaning into the molten puddle issue by using a heated pressure vessel above nozzle that contains a molten filament pool that is supplied by a regular spool as filament could be introduced through air tight seal higher up before melting. Then use air pressure to extrude as fast as you like. Think of it like the bowl on a carburetor maintaining a fuel supply. Retraction could be handled via negative air pressure or a pressure release valve below molten filament line than channels out as waste or reintroduced like an overflow tank on a car radiator. Kind of like injection molding but dynamic. Could use a bladder system like a well water expansion tank to avoid air bypassing molten filament. As a real estate broker with zero engineering experience, I’m 100% sure this will work. I did receive extensive training early on from Nickelodeon’s Mr. Wizard. :)
15:23 Dude. I am so impressed. What a beautiful design and build. I bow.
No way! You retrofitted an old CMM. I always dreamed of doing that. Was a CMM programmer for a decade before switching careers. Way cool!
The heat break tubes have terrible thermal conductivity. So you should end them as soon as possible instead of having them run all the way to the top. You want the filament path to be surrounded by aluminum to keep it cool at the top. The SS tube is preventing the heat sink from working well.
love this about the 3d printing community, keep going!! 🎉
When I used to work in extrusion, we would have gearboxes metering the material through the die. These were very useful to keep the back pressure at a constant amount and to get a very precise extrusion.
Measuring hotend performance like a coke dealer, how much "bang per gram".
Amazing engineering and machining, you got a new sub. Cheers.
This is the first video I've seen of yours and I've gotta say I am really impressed. I am also SO jealous of that workshop. I subbed because I am very excited to see where this goes. Also it would seem to me (a layman compared to you so take it with a grain of salt) like liquid cooling would be the way to go on a printer like this that isn't so heavily constrained by the weight of the tool head
Thoroughly impressed. Very cool idea for the whole printer.
When it comes to part cooling, seeing as the hotend only moves up and down have you thought about a wind tunnel around the whole contraption. That way you can suck most of the heat out. Hey a prints a print when everything is built from scratch. I just can't wait for the evolution of this.
You already have a source of "constant force" springs: The stepper motors. If you can make some clever way of controlling the motor torque, you can use them to provide constant tension on the cable regardless of its position, which eliminates the need for attaching a spring (or anything) on the other end of the motor at all. It would mean that the entire load of the bed motion is carried by the motors. With some precise device to get feedback on the stepper's feedback (like a high-resolution encoder) you can do it.
We do that in theatrical rigging installations using large multipole servo motors (which are basically giant steppers) all the time.
Constant force spring = gravity? Just a thought. This project is fascinating! Love your work!
Totally awesome, can’t wait to see what quality you will be able to get from this thing
I understand that this is a video about3d printing technologies, but that CNC tool change is fantastic
Its so satisfying to see this level of engeneering and using maschines. Stunning job, thank for sharing! Greetings from Austria!
Nice innovating. Also I'm from Stratsys and you'll be hearing from our solicitors.
Haha go ahead :D
Brilliant! Under a minute! Almost there - all that’s left is shape! Keep it going
I love how so much of this system is new knowledge to me. Very interesting to follow! Also don’t mind the length as its chill to watch while I am tuning my RatRig 3.1 500x500
Love your projects man. This is some really neat stuff you’re doing and wish I had the tools to try something like this myself. Thanks for sharing and cannot wait to see what you manage to pull off with this.
Another idea…. You can use such a setup along with CMYK filament at different ratios, and basically create any color, if it mixes well enough in such a short area (not sure it will though)
Honestly on a bed slinger with Bowden extruder this wouldn’t be a terrible product. With programming I’d be willing to bet full color CMYK would be plausible
Hi, i have some thoughts about this project i wanna share with all of you. Maybe some of my ideas are usefull, others might be not, but a discussion is always good.
1. For cooling the hotend, you could use an old water cooling from a PC. These are cheap and in combination with a costum made coolingblock, you will have a capeable system. Also you can than change your cooling effectiveness, by simply changing the speed of the fan that is sitting on the radiator.
2. I have never seen this before, but in theory you only use retract in a normal FDM printer, to decrease the pressure inside the nozzle. Therefore the friction between the molten plastic and the walls of the nozzle is strong enough to prevent a leakage of filament. So in order to make a fast extruder, with a long heating zone for the incomeing filament (this would have benefits you have mentioned in the video) we need to make a design that can rapidly decrease the pressure inside the nozzle. Furthermore i would suggest useing a small stepper motor and a moveing "wall", so you can adjust the volume of the nozzle really quickly. This comes with a lot of challenges like: sealing the entire nozzle setup, so no filament can clock up the motion of the wall; building a system that is light weight enough to not disturb the motion of the z-axis; create a mechanism that can withstand a constant pressure of up to 300 bar in inside the nozzle. So it is hard, but might be a good idea for this use case.
3. As some other comments suggested already, you could use a vacuum system for your motion system. However i would recommend a magnetic system instead. A combination of permanent and electronical magnets could solve this issue while also being more reliable and faster than the vacuum system. Let me explain:
3.1 The permanent magents would be used to create a constant tension on the ropes so they dont slip.
3.2 The electro magnets would be turned on and off individually depending on the current motion of the build plate. This would result in a stronger force and an even faster accelleration. In my opinion it is also easier to build such a system in comparison to a vacuum based one.
4. Consider useing a laser 3d printig method for manufactoring the hotend. By drilling in a solid block, you will always find some sharp edges inside, that you cant get rid off. But these increase your friction. A good laser based 3d printer can print this part with a higher precision due to lower tolerances.
Nun nochmal auf Deutsch, da du ja augenscheinlich der Deutschen Sprache mächtig bist ^^
1. Für die Kühlung der Heatbreak würde ich an deiner Stelle eine alte Wasserkülung mit 120 oder 240mm Radiator verwenden. Damit hast du definitiv genügend Kühlleistung. Auch kann durch die Regulierung der Lüfterdrehzahl auf dem Radiator die Kühlleistung angepasst werden. Wenn du kleine Schläuche, mit jedoch einer hohen Durchflussgeschwindigkeit wählst, sollte auch das Gewicht des Teils nicht alzu große Auswirkungen haben.
2. Um das Problem mit dem Filamentfluss zu lösen. Ich habe überlegt, wie man es schaffen könnte ein langes Hotend zu verbauen und gleichzeitig gute Retractwerte zu erhalten. Da wäre ein System das das Volumen innerhalb der Düse anpasst natrülich ideal. Ein Steppermotor oder Servo könnte ja einen kleinen Teil der Wand bewegen und so das Volumen vergrößern, wodurch der Druck in der Düse abfallen würde. Dann würde auch kein FIlament mehr aus der DÜse gedrückt werden und du hättest ein sauberes Ergebnis. Das System habe ich so noch nie gesehen und schon in der Theorie ergeben sich für mich da ein paar Probleme. 1. Wie schafft man es das ganze richtig abzudichten, sodass diese bewegliche Wand nicht verklemmt. 2. Wie schafft man es den Druck der (ich meine es mal gehört zu haben, aber sicher bin ich mir bei dem Wert nicht) bis zu 300 Bar innerhalb der Düse erreichen kann aufrecht zu erhalten (also die Mechanik so stabil und gleichzeitig leicht zu bauen). 3. Was passiert wenn der die Extrusion wieder fortgesetzt werden soll? Man könnte den Stepper Motoren die das FIlament einführen zwar den Befehl geben kurz anzuhalten und gleichzeitg die bewegliche Wand wieder zu schließen, sodass das "überschüssige" Filament in der Ausweichkammer aufgebraucht wird, aber das könnte auch in der Praxis zu Problemen führen. Das müsste man testen. Trotz alledem wäre es sicherlich das spannenste Teil am ganzen Drucker!
3. Wie andere schon beschrieben haben könnte man ein System nutzen das mit Vakuumpumpen funktioniert. Jedoch bin ich kein Fan davon. Es ist sehr schwer 4 Zylinder zu bauen, die Luftdicht sind, aber in denen die Reibung gleichzeitig gering genug ist, um wahnsinnig schnelle Bewegungen auszuführen. Auch ist es meiner Meinung nach nicht möglich eine gleichmäßige Kraft in jedem Zylinder in ultra kurzer Zeit aufzubauen, um so die Seile im richtigen Moment zu spannen. Klar, man kann die Zylinder permanent unter Vakuum setzen, jedoch arbeiten dann die Motoren ständig gegen diese hohe Kraft an, was den Strombedarf erhöht und unnötig Leistung zieht (erhöht dadurch auch die Temperatur der Motoren, die dann zustätzlich abgeführt werden muss). Stattdessen wäre es meiner Meinung nach sinnvoll ein System mit Magneten zu bauen. Wenn man Permanentmagneten verwendet, kann man denselben Effekt wie jetzt mit den Feder erreichen. Jedoch bin ich der Meinung das es am besten wäre zusätzlich, kleine Elektromagnete zu verbauen, die im richtigen Moment (wenn Spannung im Seil zu niedrig ist, eine lange Wegstrecke zu fahren ist, oder einfach um noch schneller die Platte bewegen zu können) zusätzlich die Kraft auf das Seil erhöhen und so die Motoren unterstützen.
3.1 In dem Zusammenhang fällt mir gerade ein, dass es sein könnte das das Druckbett minimal in der Höhe vibriert. Dadurch könnten zukünftig auch Druckfehler auftreten. Das könnte man jedoch lösen, indem man die Motoren mit einem leichten Gefälle auf allen Seiten anordnet, sodass quasi zusätzlich von jeder Seite leicht nach unten gezogen wird.
4. Das Hotend mit der Düse konnte ggf. mithilfe eines Micro SLS Druckers gedruckt werden. Damit lassen sich sehr genaue Toleranzen erreichen. Somit entstehen anders als beim bohren keine scharfen Kanten und noch glattere Oberflächen innerhalb der Düse, wodurch der FIlamentfluss verbessert wird. Auch kannst du so das Teil direkt aus Titanpulver herstellen lassen. Ob das am Ende preislich sinnvoll, und überhaupt notwendig ist ist abzuwägen.
Generell gilt, ich wollte dies Ideen einfach teilen. Vielleicht sind sie doof, aber vielleicht auch nicht. Zumindest drüber nachdenken kann man und vielleicht bringt es ja einen dann auf die richtige Idee für das jeweilige Problem. In jedem Fall bin ich gespannt wie diese Projekt weitergeht und was so die Lösungsversuche sein werden :)
Very cool project! Looking forward to seeing a sub minute video being uploaded here soon ;)
I haven't watched the two follow-up videos yet, but I have an idea for balancing the spring force on opposing sides of each axis! You could hydraulically or pneumatically link them. Basically, instead of a spring, use a syringe on each side with a tube that connects both of them. Anytime one axis is moving, it will transfer the substance between syringes. With a little air in a hydraulic system it may act as a slight spring buffer, but shouldn't favor either side. Not sure how well this would work for sure, but it's an idea.
Looking forward to see how this adventure continues!
10:47 That tool changer was so satysfying.
Thats a serious challange, I wish you every success. Nice job on the machining.
You need a sock to protect the print from radiant hot end heat. Very interesting project.
What a crazy endeavor. I love it!
I was thinking about a hot end design like this for color prints, but without wasting material, but I have no tools to prototype it. After you succeed with the speed benchy project I would love to see you try it with different colors alternating the extruders based on the needed colors. Side note, I think you could make it compatible with existing nozzle designs.
I just found this video, if it was planned I can't wait to see it print in colors :)
Really Cool Project 👍🏻
A couple maybe crazy suggestions but who knows.
Cooling the cold end, have you considered liquid coolant, maybe something from the PC CPU/GPU market you could re-purpose.
Constant force springs, tension, slack etc seems like a real headache. Could you go with a closed cable system for each axis and do away with the springs entirely? If you route the cable from the bed to the stepper then under the bed assembly and back up to the opposite stepper then back to the bed giving you a closed system. Then you could adjust the tension and have a very thought cable system. Depending on the loads you might be able to eliminate two of the steppers.
awesome video! I just found your channel and plan to watch more! you might want to add thermal epoxy to the contact area between your heat breaks and cold end. Then, maybe add an 'insulating epoxy (is that a thing?)' between the heat break and the hot end. that should kills the heat creep and oozing issues respectively. Unless there is something about that that wouldn't work? I'm certainly not the most knowledgeable person out there.
Hey, awesome work!! Im surprised you already have done it to a first print 👏 just give yourself enough of resting time 👍 also love the community standpoint !!
Really love your project and your videos!
Love this project!!
Another way of having constant force is just gravity, it’s really simple and robust and also fine controllable
With springs, if you apply pre-tension on the springs you can end up in the better end of the spring curve that is not as steep in Force / distance as an untensioned spring.
also pressure based spring might help here.
For the constant force, can't you just put another stepper with a spool on each of the xy cables to replace the springs? I don't know about the code but if a gimbal can keep its position, then stepper should be able to hold a position on the cable/constant force. 4 more steppers is annoying to deal with, but it might be worth a try.
You would also be able to vary the force/tension on the fly if needed.
Keep up the great work, I can't wait for the next update.
Super cool project excited for the next episode
Gas struts instead of springs, like the type that open the boot on a car. They are almost 100% constant force. Maybe, thread the four tubes with a fine tapered thread into the hot end. This would allow you to remove the centre bolt and hold the tubes with grub screws on the top.
Never seen any of your videos...30 sec in I see the craziest motion system I've ever seen!!! I immediately Like and Subscribe! Popcorn ready binging on your channel for the rest of the night!
It is so cool to follow this project! Can’t wait to see the next video!
Respect from a mechanic engineer 😊 keep going 😁💪
You should use weaker springs (lower spring rate) stretched more. So practically just use longer springs stretched until they offer the required force. This way any given movement will have a smaller effect on the spring force.
i'm so excited to see it work. wouldn't something like water cooling perform better than using shop air?
Water will easily sap heat from the plastic and the injection head causing uncontrolled hardening of the plastic during operation as the molten and solidified plastic are moving at very high speeds. There's a fine line between plastic that flows and plastic that locks in place and there needs to be a physical connection between soft and hardened plastic as each layer touches during extrusion for a good bond. You won't achieve this bond if one layer is too cold.
@@btwbrandI don’t think anyone thought of using the water PART cooling. But for cooling the hotend after the heat break, or cooling the NEMA steppers it could for sure work out well.
I guess shopair was already on hand, and water cooling will require purchasing equipment just for this project.
It should be noted that that path the filament takes in the hot end is not optimal, as by traveling a path with a sudden angle change, energy is lost, resulting in less pressure. I dont know of a great way (if any) to achieve a more efficient path with cnc, but still a great video and interested to see where it goes!
Angle 4x cold ends so they're inline with the drilled path through the hot end would be my preferred solution. More work but would look and work cool.
The leaking filament on the stainless steel tubes is due to those tubes not having a proper tight seal.
It looks like the tubes slide in easily. The aluminum block expands when heating up more than the stainless steel tubes. Even is there is a press fit of the stainless steel tubes in the heater block you risk getting a loose fitting that allows filament to leak out. I think this mainly is the reason for having the heat break and nozzle both screwing in the heater block as tight as possible on the E3D style hotend.
You also want to have those heat break tubes as thin as possible. Both to have the smallest possible heat transfer between hot and cold sections but also to have the fastest cool-down zone (you don't want the filament to be in a soft state around glass transition because this may result in a blockage of the filament transport)
My heat break is a 6mm stainless steel bolt with a 2mm hole and a 2mm long section thinned down to a 0.4 mm wall thickness. This gives a 3 mm² area in the thin section. Your 4mm tube has a surface area of 9.4 mm² and with 4 of those, this is 12 times as much as on my standard hotend.
Thank you. Next iteration is already on its way!
Use syringes with their ends plugged up to make a constant force spring, the vacuum will result in a constant force that you can adjust by using bigger/smaller diameters
*Alloys are ALWAYS worse in transferring Thermal Energy*
The same goes for Electrical resistance! Purity of a Metal makes HUGE differences. That's why we bother with oxygen free 99.99% Copper aka Cu-OFE (in applications where it counts)
I have some thoughts for the spring tension issue. Firstly, constant force springs are not too hard to come by, on McMAster Carr anyhow. But one with exactly the tension needed might be harder.
I think that one way around this would be to make it so that your X axis and Y axis are each a pair of steppers (or a single larger stepper) connected to a 'single cable' which loops with pulleys from one corner of the build plate to the opposite corner, with a single spring in the middle of the cable to set the tension. This does not help at all with the rotation issue though.
Ok, so hear me out. I think we can bring it down to a sub-15 second print.
If we scrap the standard build plate and replace it with two or three build plates that are at opposing angles to each other, then we go ahead and adjust the geometry of those plates to match the geometry of the benchy. Then we can add a few more extruder nozzles, and have them sticking straight into the build plates from the outside. . . .
Using 3 filament tubes in sted of 4 will make sure all of the tubes get the same pressure against the face seal making them less likely to leak.
really interesting and intriguing - especially the motion system.
Awesome work. You have two major issues the base and the extruder.
The base has numbers of excellent suggestions and other projects that can be informative - though the shear acceleration/speed that you will be working at, even though short distances, will stress any system.
So to the extruder. You need to extrude 12g/min you also need to cool that filament BEFORE the next pass occurs over the same spot - easier for the lower half of the benchy but progressively harder as you move up.
I know there are ways of calculating the pressure needed to extrude through a 0.5mm nozzle and achieve the volume flow required, but my mind is fuzzy at the moment. But directionally with the 4 feeds you are going the right way as you need to keep the speed as low as possible for as long as possible. Also you may need a change of material r, to perhaps teflon or something else, to reduce the friction. If it wears fast who cares it only has to deliver 1 benchy, the job is to meet the extruder nozzle dimension for the print.
The cooling has to occur as soon as it exits the extruder and hits the previous layer. So this sounds like an extended nozzle so that you can make a colinear nozzle that has an air gap to the nozzle for insulation. The air is then directed directly down onto the extrusion. The pressure/volume of flow needs to be worked so it doesn't blow the extrusion away. The nozzle may need to be quite wide to allow sufficient mass flow.
Just a thought, you want the cooling traveling behind the extruder - so following it like a slicing knife. Perhaps a rotary outlet that is spun insync with the travel direction so that most of whatever is being used for cooling is directed straight onto the newly laid bead and avoids the area about to be extruded onto.
TLDR
The following was written and as I developed the idea I realised that air is AWESOME and is very difficult to beat - if you can't use a liquid (hence liquid nitrogen gets a mention).
So if you want to know why air is so good here is some more reading. - You have been warned. TLDR is OK.
Can a normal cooling fluid be used? I can't think how as you need to keep the area to be printed on clear so that it can be printed on......
So if no fluid then it becomes gas.
Moving away from air the only gases that have significantly better heat transfer are Hydrogen & Helium. Hydrogen wants to go bang so lets look at helium. Helium has ~5X the specific heat kJ/kgK (what it can absorb) compared to air, its relatively safe (dilutes in the air and escapes UPWARDS) and not to bad to be around - except for high voices....
Helium has a specific heat of 5.2 J/gK (13X better than copper with ~0.4 ) & and a conductivity of 0.156 W/mK (~400 for copper or 2500X) (also a Joule (J) is one watt per sec (W/s)).
What this shows is that you want copper between your heater and your filament as it transfers the heat VERY well - 2500 times better.
However, you want helium to cool it down, as you are wanting the heat to be absorbed and removed by the coolant - and its13 times better than copper.
FYI Helium is ~20% better than water by MASS, so needs a lot more volume.
Now we come to the next unfun fact. Air weighs 1.2 kg/m3 and helium weighs ~0.2. So for the same flow rate (m3) air flows 6X more weight AND helium has 5X the specific heat - which is not enough to overcome the mass advantage that air has. Air is AWESOME
Another way of going is cryogenics, or rocket science..., Nitrogen has ~the same thermal mass as air. BUT you can get liquid Nitrogen and that obviously is WAY cold, AND the latent heat of vaporization (or getting it to boil) will provide a MASSIVE cooling effect even before we get the Nitrogen gas thermal mass and the massive temperature difference. This would be more likely something like having 4+ nozzles that can be programmed to emit nitrogen liquid at the right spot at the right time. I have NO idea how to actually do that - but the whole internet and its denizens are out maybe that can help? Maybe even some rocket scientists????
Yeah, Helium cooling would be pretty dope 😀
@6:20 You have 4 time the lenght of a volcano, but filament will move 4 time slover for the same volumetric speed, because you feed 4 filaments in one nozzle. In my opinion there is a maximum flow limit what can be usefull before the presure turns the nozzle from a plastic depositing device into a cutting device, especialy becouse at that speed and that short layer time you will not be able to cool down the printed part and you will squirt moldet plastic on moldet plastic... This type of speed can be usefull (maybe, depends on the precision of the motion system) only if the layer time is long enough to cool down the previous layer.
love this project, cant wait to see it finished
If rotation is an issue, The simple four bar parallelogram linkage with control all rotational movement
What about looping the cable from one side to the other?
Then you can remove the springs, and tension the loop by moving the motors?
Keep up the good work. 😊
How about “home running” your cables so when one cable is extending the other cable is retracting? Think about an airplane autopilot servo where when the pilot pulls / pushes on the yoke the control cable goes from the cockpit back to the aircraft control surface and another cable runs from the control surface back to the yoke. In the middle of all of this is an autopilot servo with one run of the cable wrapped several times around the hub of the servo. Run the cables under the table so the cables are a continuous loop under constant tension. Use pulleys for the cables to reverse direction. I’m thinking you would just use one drive (two total) with the cables wrapped around the hub of the drive. Search - - aircraft cable tension meter - - to get it dialed in. Best of luck.
………perhaps the only cable would be a short cable just millimeters longer then the movement of the printer table that connects a rod on the top of the table down under the table via pulleys to a 12 inch Stroke Linear Motion Guideway & Ballscrew Combined on the bottom of the table that exits out the other side to connect the opposite side of the printer table. By using solid rods it takes out the slop that would magnify as the cable length gets longer and longer.
I'll be here for the sub 1 minute speed boat if you keep going for it.
The good old german budget beer - Oettinger 😄
Maybe instead of finding a constant force spring you can use gravity to tension the strings by attaching a weight to them via a pulley? It would definitely give you a constant force but am not sure if it would solve more problems than it causes but that was my first thought on the issue.
Maybe you could pump molten plastic instead of heating it at the nozzle? It looks like melting and moving the solid filament fast enough would require a lot of energy and effort.
I was thinking. Just treat the plastic as a liquid and pump it out. Then it could be melted of the head and brought there in a heated tub that only got sufficient heating to keep it warm
Four different coloured filaments might offer some hot end flow insight.
Wonderful project!
Question: Why PLA? ABS has lower specific density and should be easier to heat up. And as you want to go >250°C anyways...
Richtig gutes und interessantes Projekt ....
Ich denke irgendwo wird Schluss sein ..
Das gedruckte Objekt bzw die Höhe und die Masse von diesem werden die Geschwindigkeit drastisch beschränken...
bin gespannt auf die weiteren Teile
If you want a constant force from your springs at all lengths the fastest and most effective, for the investment, solution is to use a Very long spring. The longer spring will have much less rate change as it is stretched from minimal to maximum operational lengths in your builds required movement. If it's long enough the rate of change will seem constant in comparison to the current lengths.
Injection molding nozzels in industrial settings take the feed stock ( pellets ) and induce lots of turbulence in that plastic as it travels to the injection site using screens and labyrinth nozzle geometry. This thoroughly mixes the plastic and brings the entire injected mass to a consistent temperature for smooth operation that is also fast and repeatable. Your heated head dealing with a small volume of material could likely benefit from more thermal homogeneity despite its size.
The feed stock seems to have a tendency to twist as it is unloaded. If this was hand wound then your wrist will induce twist in the loading process and can cause tangles when unloaded.
I'd bet it'd be possible to make a print bed support structure using compliant mechanisms that would prevent rotation while also not limiting movement with excess friction. Imagine an upside down pendulum, with the fulcrum being a 300mm below the print surface. Since the print bed is so small and the pendulum radius so large, the nonplanar movements of the bed would be insignificant, but you'd get cleaner cartesian movement. The support between the fulcrum and the print bed could be a carbon fiber rod to limit mass and air resistance, but still prevent rotation. Though in thinking about it, if the fulcrum was mounted above the print bed and offset from the print head, compliant u-joints could connect the carbon fiber rod(s) to the corner(s) of the print bed and prevent rotation, while the print bed itself is still resting on the table as you have it now, or even completely hanging from the pendulum. With a bit more thought, there's probably a way to flatten the pendulum so the whole compliant mechanism could sit under the print bed and above the tabletop it's sitting on.
I did something similar but with a 5 into 1 design. Looks neat tho :)
But my goal was not to speedrun, i wanted to mix the colors ^^
For the cooling: As weight is not a problem, just go watercooling or CO² - for a sub min use it would be worth it.
Also part cooling will be a problem as u have seen yourself. This could also be solved with a shield to the nozzle and a sourrounding cold gas shroud to cool the part down immidiately.
And yep.. to attach the cables all in mid with cableshoes by just one bolt would be better i guess.
The ultimate bed slinger
Instead of a resistive heating element it may be an option to use an inductive heater. Thermal conductivity virtually goes entirely out of the equation then as the heat can be made virtually uniform right across the heating element.
Under "hot mess" in the dictionary there is a link to 31:20 .
Overall I think this project is going awesome but I still think that moving the molten object is going to be problematic when it's at super high G forces.
I have heard of people using needles for heat breaks since they are designed to be very thin with good tolerances and very smooth walls. They may be a better option for you. You would have to watch out for filtered ones though.
Thanks, I'll look into that!
Try cooling down also the tubes with air, or made the tubes in other metal. You should use a peltier to cool down the aluminium.
this is insane, keep going!
Love the motion system.
I vote water cooled cold end!
Bro is 1000 IQ Ryan Reynolds.
Edit: 36:55 couldn't you use a weight on a pulley as opposed to a constant force spring?
That would beat the purpose as it would add mass.
I wonder how much heat from the heater block is transferring to the heatsink and the print below it just from radiation. Leaving the surface of the heater block shiny instead of matte would reduce that effect. Alternatively adding a silicone sock would might have a greater effect and then you could blast that area with part cooling and not have to worry about maintaining temperature.
Regarding the heat breaks - if you can't source titanium, I wonder if you could find someone with a high temp printer (intamsys, vision miner, etc) who could print it from PEEK or CF-PEEK. they can withstand something like 280C and have low thermal conductivity.
Springs With Constant Force? Power springs - like the kind found in tape measures? Feels like a ThisOldTony Teachable Moment (TM) let me just film my hands... Hmmm... Gas cylinders ;)
Using water in lieu of air for cooling the heat break should solve some of your issues. Cooling the part could be a bigger problem. An alcohol sprayer/fogger might work.
How about returning the cable over to the opposite side? I feel like using springs for tension is always a bad idea as they vary depending on the load. Better to make something that have a constant tension and goes around to the motor on the other side
If I'm understanding crrectly, you're suggestign a closed-rope system for each two opposing motors. That's also what I thought of at first but since the bed has to move perpendicular to each axis, the total length of rope for each two opposing motors isn't always constant.
As someone else suggested you could connect the ends via a spring. That spring has to compensate for the little bit of movement when you move in both axis.
Now that I think about it, isn't there a setup that would work without springs? Maybe using only three ropes which are connected in one point on both ends.
@@sobhanabedi3363 Yea true. Need to think some more on this. I still believe springs are a bad idea
Could you join oposire sides of the strings together by routing them around through rollers? And have one of the rollers adjustable and spring loaded. Sorta how its done on cartesian printers =)