Cheers, nice project. Two things, that immediately came to mind after watching: 1. You might want to have a look at the StealthChanger project to steal some ideas, like the brass bushings, they use with their dowel pins. 2. I would add a microswitch to the carriage, that can sense, if a hotend is present. That way you can react to a failed dock/undock maneuver and stop the carriage from picking up a hotend, while another one is still present.
Or you could even use a hall effect sensor in the tool carrier and magnets in the tool heads as a presence detect mechanism. So long as the hall effect sensor is triggered the machine can assume a tool is loaded into the carrier and it should undock before trying to retrieve another tool. This would be similarly cheap compared to a microswitch but also comes at the advantage of being easier to embed into the tool carrier since hall effect sensors are tiny.
Maybe use some heat-set inserts as your bushings. You can drill them out to the exact size you need for the pins to locate. They'd be easy to install and should be long-wearing.
You're a genius, man! This may become one of the most significant advances in home 3D printing in years. Even if a few parts have to be replaced with machined metal parts for optimum results, it'll still be dirt cheap compared to all the alternatives. So it'll make tool changing accessible to the masses.
Haven't done it myself but I think you can order metal parts either machined or done with metal 3D printing from online services like JLCPCB with just an STL file. That's not more difficult than getting the right size of screws. There will always be some parts you buy online.
@@EngineersGrow The vast majority of people interested in this project will want to order a complete kit. I'd assume that in a year or two you'll start a crowdfunding campaign to sell kits. And hopefully there will be so much interest that you can't produce those kits yourself anymore. So you'll order the parts from mass production 3D printing companies. At that point it's only the desire to keep price down that motivates keeping the number of metal parts low.
@@frankbauerful "The vast majority of people interested in this project will want to order a complete kit." ... Wow, the video is only 4 days old, but you already did market research among the viewers ? Impressive ...
I´ve always found it wasteful that toolchangers are swapping the entire print head instead of just the hotend. It´s nice to see that someone else had the same idea!
Mihai design also did a single extruder motor design. It's not a new concept but it is still very uncommon. I'm not sure if I'm sold on it yet. Most people probably will be. I bought a ton of extruders on clearance recently so I'm personally set. That's not repeatable though.
Please make this available to all. I'm desperate for the ability to change nozzles and colors without spending a disgusting amount of money. This is the best option I've seen so far
Suggestion: Put a metal sleeve in the 3d print the pins can go into. That way the wear is on the pins and sleeve. A short piece of pipe just big enough for the pin the easily but firmly slip into (chamfer corners of both the inside of the sleeve entrance and leading end of the pins slightly?) should help.
@@dtylerb In the past, I took some pop rivets and pushed the ping back through the hole before the rivet was 'activated'. That gives a pin and a sleeve that match in size and are typically aluminum. Just a thought.
This multi tool head design where only the hot ends change is brilliant for saving cost and weight. I wouldn't be surprised to see a manufacturer implement this design in a future product.
This is VERY very innovative. You are pioneering the next generation of 3D printing DIY so keep up the good work! If you want to, I would HIGHLY recommend starting up a discord channel for this project so that you can get a community of innovators behind it because this is huge.
You have done some great work here. If you look at what you are trying to accomplish you will see that there are some standard design approaches for something like this. The locating pins - they should be hardened. Also, they should have a bullet or taper to help guide the alignment. Also they both do not need to be round - one can be a diamond pin which makes mating forces even easier and gives up nothing to positional accuracy. Now that you have the hardened pins you can consider using bushings in your receiving block. With the magnet locking idea - perhaps this is a weakness since that is the only thing other than friction in the mounting pins that is holding your tool in place. The back plane where the magnets are located should be a milled flat surface - then when the tool comes in it has something very precise to set that dimension. If you want the parts to be mostly extruded you could get some milled flat steel and embed it into the design. To assist wit the clamping perhaps something done in the code to open and close clamps on the toolhead. With these ideas implemented any "wobble" or positional inaccuracies should be taken care of. Hope this was helpful. I am sure that what I commented on you may have already though of. Also, I did this in just a quick minute so there could be something I missed. However I have worked in manufacturing as a degreed mechanical engineer for a very long time (think decades) and over that time I have designed and built a bunch of machines - But I don't think that i ever invented anything, mostly I learned from others - over the years I have been around some amazing engineers that were simply brilliant. Your design is very inspiring, it inspired me to make this reply and really got me thinking. Great work. Keep up the great work. Cheers.
Do you know of any good low cost sources for the hardened diamond pins you mentioned? Earlier I looked on McMaster Carr and everything I was too expensive for this project. Honestly, I feel the same way. My best designs were not my own ideas, but using the ideas my coworkers suggested. Thank you for the support and your feedback! Glad to hear this project is inspiring.
@@EngineersGrow McMaster is a good source but more general. I worked in industry so the cost was not a huge factor but you can try: Carr Lane, MSC, Fixtureworks, Reid tool/supply... I think Carr Lane could be the lowest price. However you can do this with dowel material and cut your own with a Dremel and a bit of patience. You can "hack" up a diamond shape in intent, it doesn't have to be pretty... just take your time and see what you can come up with. And FYI in case it isn't known is the reason for the diamond is that it reduces redundant locating issues which can cause binding on fixtures with dowel pins and makes the install/removal a bit easier as it does not touch the dowel in the ground off areas. When you make these really tight you can often find that just the slightest angle when mating/un-mating will jam the crap of of them. There are even cone shape pins that only engage the last millimeter or so reducing the jamming you can get with dowel location pins. See what you find. If I can think of more i will let you know. You can use receiver bushings that are a few tenths oversize that also ease the mate/unmate binding. With these installed your location will be very accurate and repeatable and it will handle thousands of cycles before it starts to really wear.
Awesome idea! Use a piece of spring steel on the dock to block the oozing and help keep the hotend primed. I have seen this on a couple of tool changers.
@@EngineersGrow While this method works on other toolchangers, it wont be as effective with yours as any pressure will simply push the filament up out of the hotend. You may need to look into a way to hold the filament in place when its docked. I can see oozing or nozzle pressure being a big issue with this type of system
Interesting, will need to keep this in mind. I would think that once the spring steel touches the nozzle it will not be able to push any more filament up.
@@EngineersGrow Why would it not push any filament up? You'll have pressure inside the heatblock which is what causes the ooze. that pressure will want to release somewhere, usually you have an extruder holding the filament, so the ooze comes out the nozzle, but in your case, you plan to block the nozzle so the pressure will want to release upwards which will push the filament up and out of the heatblock into the coldside/heatsink. Changers like the XL and the Stealthchanger block both the nozzle end and the filament end of the hotend which contains the pressure
This is incredible. This is something that makes me excited. Can’t wait to see future iterations; especially as you add sensors to do automatic calibrations, filament checks, and docking checks.
A tip to prevent hotends crashing, attach end stop switch near where the crash happens. Better yet, you could use an end stop switch to confirm that the hotend detached. If the switch doesn't turn off, stop extruder from moving and throw an error/ pause printing.
This is AWESOME! Truly shows whta 3dPrinting makes possible. Truly shows what happens when pattents and expesive manufacturing don't actively stop progress.
Impressive! I'm working on a toolchanger of my own, and the extruder problem is a big one. Speed is a huge potential benefit of this setup. The stepper is the heaviest part of most hotends, and attaching it to the x rail instead of the detachable tool (with the additional flex point of a tool mount, and further away from the x rail, exaggerating any flex in the system) should make it more capable of high acceleration. One recommendation - angle one of the filament drive gears so more pressure is put on the filament and less on the drive gears. My testing indicates that the gears cause a lot of resistance, meaning wasted stepper power and a bigger extruder motor. If you can switch to a smaller extruder stepper without speed loss, you'll be able to accelerate faster. Second recommendation: Go with a kinematic coupling. If you have a bimetal heatbreak with thermal paste, a 3010 fan will keep the v6 plenty cool, and long screws can be driven all the way to behind the v6 heatsink, touching the sides and holding everything more stiff. The kinematic coupling goes on a backplate. Stiffen the backplate with a pair of steel m3 screws. They do a lot. Each kinematic link can be made with a pair of slices of 3mm steel rod on the mount side and a small magnet wedged sideways in the tool. The magnet will sit between the rods and keep things stiff. Let the magnet sit low between them - it will hold things stiffer. You can hold the coupling together with a big magnet in the center, between the triangle of links. Definitely comment with any questions! Edit - my toolchanger maxed out around 100,000mm/s^2 accel, 600mm/s my last test. Toolchangers don't have to be slow!
Thank you! Wow, those are impressive speeds! Can you explain more what you mean by "angle one of the filament drive gears so more pressure is put on the filament and less on the drive gears"? My biggest concern with the kinematic coupling in this design was that it wouldn't hold the hotend for the sideways sliding motion due to the forces between the dock and extruder pins. This is why I went with the dowel pins for this first version. But this might not be an issue in some of the updates I'm thinking about.
@@EngineersGrow That's a good point on the sideways motion. What if you had a kinematic coupling *and* a pair of dowels to hold it straight? Kinematic coupling ensures precision, dowels help it line up right and keep it from sliding off sideways when mounting. It would be a pretty dramatic change though. The two drive gears are parallel in most setups. But when your extruder squeezes against the filament, it flexes, so the gears probably end up being crushed together more than they should while you're trying to squeeze the filament. This happens with the Sherpa mini as well, and many others. There's a mod for the Sherpa mini that angles the drive gear so the gears are further apart and the part that touches the filament is closer together.
Adding dowel pins to the kinematic coupling like you're suggesting probably can work. I might try that. Maybe it will be as simple as making sure the pins have some clearance designed in so that they don't interfere with the kinematic mount.
Very cool! Great job. As some people mentioned adding a metal sleeved dowel that the pins go into would prevent premature wear. Also a metal plate maybe where the bolt heads pull on the plastic when pulling the hot end off.
My experience with TPU is, if there is a small gap between the extruder gears and the bowden hose it works for loading the filament with continous flow, but as soon as you print with retractions it finds the gap and will move that way, but lets hope for the best.
Excellent excellent work. I have been working on and off on a fila-changer where the extruder (is supposed to) pick up the filament from a row of bowden flament docks at the back (just the filament, the hotend isnt changed) which would work very similar to how you have made this. Your sucess gives me renewed hope and some more ideas that may help me make my idea work. I know not changing hotend would mean the printer would need to prime the nozzle each time a change happens, but my goal was to stay within a single el-cheapo mainboard that comes with 4 stepper drivers and one/two hotend ports. Also getting perfect nozzle alignment with a toolchanger/hotend change is a pita without complicated expensive solutions. Looking forward to more updates on your design.
Get PCB way to sponsor you, use them to get metal brackets fabbed of your design, sandwich the metal brackets maybe between the pickup hub so the holes don't wear out and are more robust, ditch magnets if possible. Copy someone else's pickup hub design. Your nailing the idea totally and the extruder design!! Chef's kiss!!on your extruder idea. keep at it your SO close! Love the idea and videos keep it coming man!
Very promising. In metal work we use bushing to achieve different properties than the surrounding material. Maybe a bushing made from copper or brass will make the trick. They are harder and more wear resistant than plastic but give also a good natural grease to steel counterparts like the pin.
I tried 3d printing and using a manual tool changer which had the fans and hotend assembly all mounted together, but removable from the gantry plate, and ran into a similar problem with the hotend wobbling because it was held up by plastic. I think thats a major design challenge: using any type of plastic for the assembly of a part that is directly tied to specific structural components that need to be exact to get quality prints. Might be best to CNC those
May i suggest looking at various space docking mechanics, and copying mainly the mechanical part of it. Using probe/drouge or guides to achieve better alignment and better tolerance. You can use different design for extruder to holder and extruder to hotend. Flat head scred driver is kind a probe/drough , use funnel to guide the screw head and push it inside the funnel tube for hard capture
I’d like to recommend copper tubing to solve your problem with wear in the alignment pin holes. Just some thin tube, put it all the way through the part and either flare the ends or cut and bend them to keep them secured. I’d say flaring them so the ends are securing the tubing in place and can still be flush with the part would be your best option. This will drastically increase the wear resistance and allow a solid fit for a long time. Also, instead of screws with a flange it may be best to get studs or threaded bar stock, cut it to size, then use some thread locker to glue some T nuts on the end. My last recommendation that may not work would be to use something like a magnetic push latch to try and hold the hot end still. It may not be strong enough or the latch may require too much force to push but it’s simple and easier to integrate than a whole mechanical latching mechanism so maybe it’s worth a try. You really have an amazing thing here and it seems like it’s fairly close to the point of being to the point of an actual first iteration. I’m really interested to see where this goes.
what if you use a solonoid on the extruder assembly to push and detach the extruder from the magnets. you should alsu use 3 magnetic balls in a triangle and pins to allign and mount the hotend
Thank you! Yes you can use it on your printer and modify the design. My files are available on BuyMeACoffee. Link in the video description. My goals for this project was to have a purely mechanical docking system to reduce cost. However, for a higher end tool changer one of the options you listed might work better. I'm making updates to the design and send occasional update emails to those that signed up for the membership waitlist. Just sent an update email a few minutes ago.
This is how a tool changer should be. Having an extruder motor for each toolhead is wasteful and not necessary. With a redesign of this to make it look better, and function better, this could be one of the best and affordable options out there.
I think I also seen someone implement a wiper the covers the nozzle opening when docked. Will try to find that design once this single extruder design matures.
This is unbelievably cool stuff!! I can't wait to see how this project changes and improves over time. Hopefully there can be ways to make it lighter and more compact without sacrificing reliability!! Oh, and it might be a smart idea to make a presence detection system using a hall effect / reed or (as wilkeiner said) a micro switch to prevent the machine trying to load a new tool while one is already loaded. Other things you might want to consider would be adding brass inserts to anywhere where a metal pin locates a printed part and imparts a load on it. This will reduce the wear on the tool heads and prevent the side to side motion you've been seeing. While ABS is strong stuff, it's not completely immune to deformation and abrasion. In your case the toolheads suffer from both problems. You have the force applied by the locating pins when the tool carrier slides the toolheads across to dock them which will compress and deform the plastic over time, and you also have a degree of abrasion from the insertion and removal of metal pins into a plastic component combined with the spring pressure from the mechanism that docks the heads. I'm fairly certain this is where your excessive play in the toolhead that was printing the black filament has come from. Again though, this is a fantastic beginning to a much needed project, and hopefully it will be adopted, improved and made into a standard so we can finally have cheap, decent quality multi-colour prints that don't waste 10 metric gigatonnes of filament!
why not for the alignment holes, use drilled out heat inset nuts, that way the holes cannot be worn out and are more stable. and my other idea would be a little dropper for the parking so that the spring in there isnt nedded could also be more reliable but idk if it would work
the spring seems way way too strong yeah, and the source of the rate of damage from undocking. countersinking the holes would also help so the dowel does not have to jump to the right location at once instead of a spring I think you could have side magnets to align the retaining screws in either the locked or unlocked position, and then you just grab the hotend, move left, leave with it attached; and then put it back, move right, pull away without it
This is an innovative concept. I'm glad you put the video up. I would suggest: You are trying to make parts needing good precision with a poor quality or poorly tuned printer. Everything needs refined with fillets and chamfers which will improve stiffness. There is nothing retaining the pin in the driven filament gear and perhaps other pins. The washer head buttons used to retain the hot end cassettes have no locking method to retain adjustment. The area under these heads is small which will lead to wear in the mating part and loss of adjustment. Cudos for not using any set screws, the sign of amateurs.
This is gosh darn brilliant! I now will be looking into this kind of mod! I want to try and make a multi material Printed in Place lock. The options are endless!
Great job!! I think there're 2 things you can do to improve longevity on the docking parts: 1.- Round off the top of the locating pins, this will reduce friction, and thus wear. 2.- Have you thougth of lining the female end of the extuder mount with brass tubbing? It's cheap and durable. This trick may also help on the latching plate, were you to cut it out of brass (any time metal meets plastic, plastic looses). Keep it up, I really want to see this through!
Thank you! I already rounded off the pins and didn't catch it in the final recording, since unfortunately it was too late. Planning to add metal bushings.
Pretty sure others pointed that out but usin acorn nutes as bearing surfaces against 2 pins and a star pattern to make the hotend only able to locate at a single repeatable point on the extruder head. Also the release mechanisms doesn't need to be actuated by the hotend it can be a pin or a screw directly on the extruder head that pushed the locking bar that way the actuation doesn't load the locating surfaces.
You have a really good point about not loading the locating surfaces with the release mechanism! In the current design, most of the loading should go into the bracket secured to the printer frame. However, I have some ideas to make the loading localized to the extruder. I have concerns that the kinematic coupling might not work with the sideways sliding motion, but will probably look into it.
Was very interested to see this approach mentioned in Teaching Techs latest Video. I really like this concept to avoid a fair bit of the wasted hardware in most tool changer designs - and also the removal of the need to cut filaments to get them to correctly feed into the extruder. (As per the BigBrain3d Swapper3d). The additional cost and electronics involved in each nozzle having its own heater is balanced out by the benefit of being able to keep/preheat each hot end at an appropriate temperature to facilitate fast tool swaps. My biggest concern is whether the size and weight could be scaled down to a) allow really fast printing and b) allow room for at least 12 hot ends to be loaded - to really complex multi colour prints - such as ship models. To me the ultimate solution would have 2-4 fully independent print heads (with extruder motors) for parallel printing, with each print head able to colour swap using something like this engineers grow approach. Perhaps backed up by some type share across multiple printers, cheap AMS type multiplexer too for access to a larger library of filaments than it would be practical to have a dedicated hot end for each one. The solution would need to have a fair number of sensors in it, plus some type of auto calibration/alignment of the nozzle positions in 3 dimensions - perhaps with some sort of load cell arrangement in an alignment pin on the bed somewhere which could be checked at the start of each print, or maybe even after every tool change.
@@EngineersGrow Thanks - the small 1:1000 scale up to 11 colour ones take about 15 hrs to print, and the bigger 1:500's can take over 40hrs. I would estimate that a nozzle swapping approach such as yours could reduce the print time by up to 4 times, and also reduce most of the waste. Plus could open up the possibility of moving to a smaller nozzle size.
@@dwuk99 Do you have any models that only need 3-5 colors? Would you be interested in a collaboration? Your cruise ships seem like a great case study. Being able to switch between small nozzles for perimeters and large nozzles for infill is high on my list.
@ yes, happy to collaborate. I did downmix one of my models to 4 colours so that it would work on an AMS lite - p&O Azura on Makerworld. Mixing nozzle sizes in the same print is an interesting idea, perhaps using the 0.2mm nozzle for highly detailed parts like deck furniture, railings and text, with the bigger nozzle used for the rest. I think prusaslicer might support mixed nozzle sizes - will do some tests on the latest model I am working on - Cunard Queen Victoria 1:1000. I've had to remove some of the detailed parts that work ok at 1:500 scale - but could try re-introducing them with a smaller nozzle size just for the detailed parts to see how that affects overall print times. Mixing nozzle sizes would of course increase the total number of tools required further - but this could be partly mitigated by splitting the models into separate parts. Are you on teaching tech's patreon forum?- as there is a thread on there relating to your video that would be a fairly convenient place to post images of slicer tests.
I have reworked my latest Cruise ship model today to split out some of the more detailed parts for potential smaller nozzle printing, plus also down coloured it to 5 colours - I have done a post on the TT Patreon with some stats, plus have also put a summary few of them under my 'Cruise Ship Print' thread on the BambuLab Community Forum - but this is more BL 1Q25 new printer focussed than nozzle changer focussed. My findings so far though do bear out your suggestion that mixed nozzle capability may end up being more important in terms of print time savings vs the smaller than I expected savings you would get from tool changing vs colour multiplexing.
This is one of the rare times where a recommended video is pure GOLD! Liked, subbed, hit and bell and all that, I would love to see where you take this!
@EngineersGrow you're very welcome! I've been wanting to try add a 2nd head to my Anet A8, I love this single extruder approach. Only having X axis movement it would need to be open on both sides, this would be quite a challenge but I mat try it down the line.
Would be interesting to have a printer that can change not only filament color/type, but size as well! I can imagine that'd be a pain to get a slicer to account for, but would be cool to do highly detailed areas with 0.2 mm nozzles / 1.75 mm filament and then switch to a 1.2 mm nozzle / 2.85 mm filament to fill in larger areas.
I suggest to use one round hole and one slot to mate with the dowel pins for locating the hotend. This might be easier than having to meet the two round holes. The hole is now locating in XZ, while the slot is constraining the rotation.
I'd love a tool changer to print supports out of different materials for better results e.g. PETG supports on a PLA model. But as you say the commercially available ones are prohibitively expensive. Great to see more innovation!
Very cool project, enjoying watching your approach very much. So many design considerations for hardware and software. Impressive to see your endeavors. I love to watch folks work on difficult projects from start to finish. This type of enthusiasm and effort always leads to great improvements and offerings for the end users. Very impressive to watch how you make changes on the fly. Good stuff!
Ideas that came to mind from your wear issues: 1) Lining the female holes for the alignment pins with bushings, to give a hard-wearing surface 2) A slight re-think of the lock/unlock system - instead of torquing the hotend to the side each time to lock and unlock, perhaps have a sliding "locking plate" that grabs/releases the button screws? Actuate the plate using a mechanical toggle, like that used on latching pushbutton switches. That way it's unlocks when you run you go to pick up the hot end, but hitting the same mechanism a second time relocks it? Not as simple, but it would remove a lot of the stresses induced during the unlock procedure...
That's orders of magnitude better idea than the enraged rabbit's, Prusa's, and Bambu's takes on the problem! Would love to see that come to life to a production grade level
bro you are a boss. im a newbie and i love learning all these new things. im an inventor. i have easily lost 40k dealing with china. now im going to do everything on my own. great video. you get a sub brother. keep it up. thank you. this is above my pay grade right now but sweet to see someone so innovative and creative. thats what my business is made from
It depends what you're trying to save money on. This design saves a bunch on hardware, but during operation it will have to clear lots more plastic to purge every time you switch plastics. So this might be cheaper to build but more expensive and slower to operate. Still beautiful design and quality work. Bravo 👏👏👏
Thank you. Can you elaborate on why the plastic would need to be purged when switching filaments? Maybe I'm missing something. The idea is that there will be minimum material waste because there is still a dedicated hotend for each filament.
Nice job putting in the work. I'd thought of this years back to simplify the parts list, but wasn't sure it would work, and didn't have the time to invest in trying it out. The other thing I'd thought of was having the gears dock with the extruder motor. It would mean a set of gears on each hot-end, but wouldn't have to catch the filament each time. Neat to know that the idea works though. I think you'll have to do something to lock the hot-end onto the extruder mount, otherwise you'll keep having play that will mess with offsets, but I could be wrong. Keep up the good work!
Thank you! Yeah, getting the gear teeth to align when picking up the hotend would probably be tricky without using encoders. The teeth on the filament gears don't need to align to each other, so it's more simple from that perspective.
I guess I'll be spending more money on another project now LOL! Jokes aside, this is such an ingenious design, top to bottom! Every design choice was brilliant and I loved the used of PTFE as a centering spring! I've been thinking about this idea of 1 extruder motor for 5 hotends for a long time and never had the time to tackle it. I think you'll also be interested in Mihai Designs' " pitstop " project which has a similar approach! Cheers
For the XY offset calibration, maybe a technique from pick and place machines can be borrowed. They pick up a part and then move it over a camera to figure out its exact offset and orientation to let it precisely place it, similar could be done with a camera and looking at the nozzle. Then for Z offset calibration a technique similar to CNC machines could be used, likely even using the same parts, where they have a physical probe pad that they touch off on that makes electrical contact with the tool. Another way would be to just use nozzle bed probing but that requires nozzle bed probing. All this could be bundled into a startup calibration routine where it picks up each toolhead in turn and measures their offsets.
Great ideas! It will be great to get to a point when all of the calibration is automated. I think I've seen some videos of people implementing some of these ideas.
genius work! genius idea! This is how thinking should work! Not accept the things you can buy, but think about how to make it even cheaper, and better! Beautiful! Even tho i am really confused by your totally bad 3d prints (party for your changer) :D they look so bad ... ^^ BUT as long as it works, all fine! I will follow this .... please keep it up! things i would improve: 1. small sensor that triggers IF an Toolhead is loaded (so to be sure if its even unloaded properly too) 2. Proper shaft / inserts so the slignment is perfect ( you can get this sooo cheap) 3. maybe sth to geht the z-height for each head properly...
Man! Thank you so much, I had a similar idea, because why do we need to have so many damn motors eh?! Your design is way more elegant than what I was thinking, but the concept of a grab extruder is the same. I had to no time to work on it, so thank for doing all the work and bring us into a more affordable future. Subbed.
The idea is great. I think long term you need to think about CNC machined parts to increase durability and reliability. ABS might not be the right material choice in my opinion. Thanks for your effort and sharing everything with the community
One thing you'll want to look into that the other tool changers have and yours doesn't is a kinematic mount. You should have all six degrees of freedom constrained (x,y,z translation and rotation). A kinematic mount lets you get away with less accuracy in your prints while still grabbing the hotend in exactly the same way every time.
I watched this on the side, not really paying attention, wondering why this any better than tap-changer and stuff like it are already out there. But doing tool changing with one extruder is really damn cool! Very cool project! Love to see this develop further!
Great idea. For making it more reliable and cheaper I recommend David Malawey’s video or his shorts on “Borrow a Tolerance: Mindset for Designers”. He has some good ideas for functional 3d prints that get their reliability or durability from precise things like dowels, ball bearings, premade materials. I recommend watching his shorts (I never thought I’d recommend shorts) on it because they go over more relevant ideas than the longer video. The longer video can help with understanding the theory from an engineering pov.
I feel like a significant cost that has been missed is the electronics. Each hotend is going to end up needing at least one heater, thermistor and fan. This doesn't tend to be an issue for 2 tools, especially if using a mains powered bed, most big control boards have enough spare ports for that much. But, for more tools that won't be enough and you'll effectively need to have a toolhead board per tool, adding ~$30/tool. On the bright side, that'll also allow for all sorts of additional reliability improvements like sensors to ensure that tools have successfully docked/undocked.
That's a great point! My hope is that board manufacturers will provide lower cost boards for single extruder tool changers in the future. Each hotend will still need a heater, thermistor, and fan like you mentioned. But they can save money by only having one extruder stepper motor driver
@EngineersGrow yep, a custom board for that can be much more cost effective! I was also thinking that fan ports needed can also be reduced, there could be one fan on the toolhead for the active tool, and the ones not in use could be cooled by a single more powerful fan instead. Then you only need thermistor and heater power, so a multichannel ADC board and a bunch of mosfets, which can be very affordable for a large number of tools.
@@dotnet97 Great idea about reducing fan ports! Now that you mentioned it, the hotend fans are small, so it's possible to even have all of the hotend fans driven by a single port. This is how I added stepper motor cooling fans on my MakerGear M2 that I modified with a heated chamber.
Being a designer of my own toolchanger as well, its a very nice concept and idea of a toolchanger. Appreciate the effort put into this design. Wouldnt this design requires u to have 2 umbillical at all times? 1 on the carriage and 1 on the toolhead?
Nice design! When I saw your design with only one extruder for multiple toolheads I thought about MihaiDesigns Pitstop (2) toolhead. Maybe you could get some inspiration there. Amazing work!
Metal bushings press fit into your hotend brackets for locating the tool changer together will work with the metal pin design. Metal on metal movements should solve your durability and precision problems while keeping parts easily home-built. Everything will get a lot more rigid too.
@@EngineersGrow Cant wait to see V 2.0! If you need someone to demo it/torture test a future version, I'd love to help out and give some more feedback. I've been meaning to build a fully home-built clipper printer for a while now.
Love the idea! You could try to use a hall sensor to see if sucessfull pickup. Also you could use 2 moves like the prusa, 2 directions could be easier for the Integration:)
Maybe some sort of metal insert for the pins to slide into. Like a heat set insert drilled out to be smooth with a flanged opening. Genius idea. I look forward to seeing you progress. Great video. Keep up the good work.
In addition to metal sleeves and detecting tool presence (there might be 3rd, I already forgot, sorry), how about: - improve tool stability by locking hotend plate to base plate with some kind of latch mechanism actuated by extruder opening arm? Locks hotend in place when moving away from the dock, but allows releasing the hotend to the dock. If possible, grabs hotend plate from the bottom? - reduce pin and hole wear by moving X to release location against dock spring, pulling bolt heads just out in Y direction, and moving X back to release spring tension before moving completely away in Y. This way dock spring does not work against pins and spring-loaded "snap" does not try to carve holes during release Prusa XL has also locking based on movement IIRC, maybe they have drawings or models released? Inspiration for parts that could work better :)
A simple latch mechanism would be great to use instead of magnets. Interesting idea about the dock spring. It's a pretty light spring, almost too light for this application. I don't expect much wear to occur, especially with metal inserts. However, I also don't like it because it takes up a lot of space in the X direction. So I'm actually thinking about maybe how to get rid of it in the future.
Hello. Thank you. I think that price of multi extruders is ok. We can use different extruders for different materials (abs, soft tpu and wax). So you shouldn't dance with only 1 extruder. The main problems are changing mechanisms and reliability.
I have been going down a different rabbit hole to make a tool changer for my core zy machine. Stumbled across this video and I like it. Your approach makes a lot of sense and sure would like to give it a go.
@@EngineersGrow We are working one it. Taking it one piece at a time to make my prototype. Started at the linear rail end and right now working on the docking stage! Love it so far.
Hi. I am a automotive mechanical engineer from Austria and I love your base concept and I would love to be part of the design team - some ideas about improvements and new ideas are already coming to my mind as I am writing this message.
This is an awesome project. It's an idea so simple, it's crazy that it hasn't been brought up more often (if at all - this is the first I've seen of this type of tool changer). One thing I'm thinking for the docking, could you add some sort of button (like an old endstop) that when triggered, confirms that the tool head is docked? And if not docked after tool changing command, home the printer and try again? At the very least, I think it would prevent potential damage from crashing the tool heads.
Ever considerd a split-extruder design, where the gears are at the part that will be changed and just the motor is shared ? Makes the mechanical part easier
I was thinking about that earlier. However, getting the teeth on the gears between the motor and extruder to align and mesh together seemed challenging. There's probably a different coupling that might work, but would need to do more research. So I moved on from this idea.
I want to see this work! Especially because there is a big challenge, that is because the extruder and the hotend are not rigidly mounted as one piece, there is the introduction of flexing/bending when extruder is pushing the filament to the hotend. I would love to see how that issue will be resolved
Great work, and very clever! Even if this proof of concept wore out quickly, it was still worthwhile for what you learned. If you can reduce the docking force required and improve the robustness and wear resistance of the interacting surfaces, the next iteration should work a lot better. Subscribed so I can follow along!
Good to see some alternative thinking - I think this might have legs. As I've mentioned before in YT comments, we need to standardise a tool change mount and this could be part of that. The mounts need to be based on a Maxwell kinematic couplings to ensure accuracy with a latch to secure. All doable with your concept. Would also be good to evolve the design for CNC as it would be much stiffer. Would be happy to chat/help if it's open source.
Thanks! I think keeping the hotend on the extruder while sliding sideways might be challenging with the kinematic mount. However, it might be less of an issue with future improvements to the extruder.
Join the Tool Changer Membership Waitlist!
stats.sender.net/forms/dJ6ZAK/view
@@EngineersGrow have you seen this tool change? th-cam.com/users/shortsaycVfRJ6zDg?si=3LYd0LFGG-yB-1V9
@@--3D Yes, very interesting approach
Cheers, nice project.
Two things, that immediately came to mind after watching:
1. You might want to have a look at the StealthChanger project to steal some ideas, like the brass bushings, they use with their dowel pins.
2. I would add a microswitch to the carriage, that can sense, if a hotend is present. That way you can react to a failed dock/undock maneuver and stop the carriage from picking up a hotend, while another one is still present.
Metal bushings for the pins was the first thing I thought watching this too.
I came here to say the exact same thing, combine your idea with the stealthchangers proven docking solution and you have a real winner.
Or you could even use a hall effect sensor in the tool carrier and magnets in the tool heads as a presence detect mechanism. So long as the hall effect sensor is triggered the machine can assume a tool is loaded into the carrier and it should undock before trying to retrieve another tool. This would be similarly cheap compared to a microswitch but also comes at the advantage of being easier to embed into the tool carrier since hall effect sensors are tiny.
Thank you! Will take a look. Adding a way to detect if a hotend is on or not is definitely on the list for future upgrades.
Maybe use some heat-set inserts as your bushings. You can drill them out to the exact size you need for the pins to locate. They'd be easy to install and should be long-wearing.
I think TeachingTech would love to work with you on this, especially for an SVO8 Tool Changer mod like he's already doing
As someone with an SV08 I really REALLY hope he sees this and impliments it! I love not having to use multiple extruders. Freaking BRILLIANT!
Yeah it could be more cheaper to use only one extruders
Ooo, yeah that'd be awesome!
What Ill say on this is that they use a different motion system. The sovol has a flying gantry while this has a fixed gantry.
@@WooHooSum I don't really see a problem with that other than it might make it a little bit more difficult
You're a genius, man! This may become one of the most significant advances in home 3D printing in years. Even if a few parts have to be replaced with machined metal parts for optimum results, it'll still be dirt cheap compared to all the alternatives. So it'll make tool changing accessible to the masses.
I agree that a few machined metal parts might be necessary. I'm hoping to find a way to get around machined metal parts for DIY setups.
Haven't done it myself but I think you can order metal parts either machined or done with metal 3D printing from online services like JLCPCB with just an STL file. That's not more difficult than getting the right size of screws. There will always be some parts you buy online.
@@EngineersGrow The vast majority of people interested in this project will want to order a complete kit. I'd assume that in a year or two you'll start a crowdfunding campaign to sell kits. And hopefully there will be so much interest that you can't produce those kits yourself anymore. So you'll order the parts from mass production 3D printing companies. At that point it's only the desire to keep price down that motivates keeping the number of metal parts low.
lol they already have this its called the stealtchanger for voron
@@frankbauerful "The vast majority of people interested in this project will want to order a complete kit." ... Wow, the video is only 4 days old, but you already did market research among the viewers ? Impressive ...
this is why we love this hobby, and pioneer individual like you sir is why we shine as a community.
Yes, looking forward to see what the community does with this project.
I´ve always found it wasteful that toolchangers are swapping the entire print head instead of just the hotend. It´s nice to see that someone else had the same idea!
Looking forward to see where the community takes this idea in the next couple of years.
Mihai design also did a single extruder motor design. It's not a new concept but it is still very uncommon. I'm not sure if I'm sold on it yet. Most people probably will be. I bought a ton of extruders on clearance recently so I'm personally set. That's not repeatable though.
Please make this available to all. I'm desperate for the ability to change nozzles and colors without spending a disgusting amount of money. This is the best option I've seen so far
Watch the last 5 minutes. He tells you how to get the files.
I will have 3D models and a membership available in the future for others to try out.
Suggestion: Put a metal sleeve in the 3d print the pins can go into. That way the wear is on the pins and sleeve. A short piece of pipe just big enough for the pin the easily but firmly slip into (chamfer corners of both the inside of the sleeve entrance and leading end of the pins slightly?) should help.
yes, Borrow durability from metal!
You should be able to get bronze bushings designed to fit the pins since they are a standard size.
Came here to suggest the same.
@@dtylerb In the past, I took some pop rivets and pushed the ping back through the hole before the rivet was 'activated'. That gives a pin and a sleeve that match in size and are typically aluminum. Just a thought.
Thank you! I'll take a look at the pipe idea. I looked at bushings on McMaster Carr to match the pin size, but they were expensive per bushing.
This multi tool head design where only the hot ends change is brilliant for saving cost and weight. I wouldn't be surprised to see a manufacturer implement this design in a future product.
That's the goal.
This is VERY very innovative. You are pioneering the next generation of 3D printing DIY so keep up the good work! If you want to, I would HIGHLY recommend starting up a discord channel for this project so that you can get a community of innovators behind it because this is huge.
Thank you! I agree, it will take teamwork to make this project successful. Will be starting a membership community soon.
This project needs to be picked up and run with! This is awesome! Tool changers should absolutely be cheap and wide spread!
Agree and that's the goal!
You have done some great work here. If you look at what you are trying to accomplish you will see that there are some standard design approaches for something like this.
The locating pins - they should be hardened. Also, they should have a bullet or taper to help guide the alignment. Also they both do not need to be round - one can be a diamond pin which makes mating forces even easier and gives up nothing to positional accuracy.
Now that you have the hardened pins you can consider using bushings in your receiving block.
With the magnet locking idea - perhaps this is a weakness since that is the only thing other than friction in the mounting pins that is holding your tool in place. The back plane where the magnets are located should be a milled flat surface - then when the tool comes in it has something very precise to set that dimension. If you want the parts to be mostly extruded you could get some milled flat steel and embed it into the design.
To assist wit the clamping perhaps something done in the code to open and close clamps on the toolhead. With these ideas implemented any "wobble" or positional inaccuracies should be taken care of.
Hope this was helpful. I am sure that what I commented on you may have already though of. Also, I did this in just a quick minute so there could be something I missed. However I have worked in manufacturing as a degreed mechanical engineer for a very long time (think decades) and over that time I have designed and built a bunch of machines - But I don't think that i ever invented anything, mostly I learned from others - over the years I have been around some amazing engineers that were simply brilliant.
Your design is very inspiring, it inspired me to make this reply and really got me thinking. Great work. Keep up the great work. Cheers.
Do you know of any good low cost sources for the hardened diamond pins you mentioned?
Earlier I looked on McMaster Carr and everything I was too expensive for this project.
Honestly, I feel the same way. My best designs were not my own ideas, but using the ideas my coworkers suggested.
Thank you for the support and your feedback! Glad to hear this project is inspiring.
@@EngineersGrow McMaster is a good source but more general. I worked in industry so the cost was not a huge factor but you can try: Carr Lane, MSC, Fixtureworks, Reid tool/supply... I think Carr Lane could be the lowest price. However you can do this with dowel material and cut your own with a Dremel and a bit of patience. You can "hack" up a diamond shape in intent, it doesn't have to be pretty... just take your time and see what you can come up with.
And FYI in case it isn't known is the reason for the diamond is that it reduces redundant locating issues which can cause binding on fixtures with dowel pins and makes the install/removal a bit easier as it does not touch the dowel in the ground off areas. When you make these really tight you can often find that just the slightest angle when mating/un-mating will jam the crap of of them.
There are even cone shape pins that only engage the last millimeter or so reducing the jamming you can get with dowel location pins. See what you find. If I can think of more i will let you know. You can use receiver bushings that are a few tenths oversize that also ease the mate/unmate binding. With these installed your location will be very accurate and repeatable and it will handle thousands of cycles before it starts to really wear.
@@genjitsu7448 Great, thank you!
Awesome idea! Use a piece of spring steel on the dock to block the oozing and help keep the hotend primed. I have seen this on a couple of tool changers.
Thank you! That's on the to do list once I get the mechanism more reliable.
@@EngineersGrow While this method works on other toolchangers, it wont be as effective with yours as any pressure will simply push the filament up out of the hotend. You may need to look into a way to hold the filament in place when its docked. I can see oozing or nozzle pressure being a big issue with this type of system
Interesting, will need to keep this in mind. I would think that once the spring steel touches the nozzle it will not be able to push any more filament up.
@@EngineersGrow just watched Teaching Tech’s video on the Voron docks and they use a silicon pad to stop oozing.
@@EngineersGrow Why would it not push any filament up? You'll have pressure inside the heatblock which is what causes the ooze. that pressure will want to release somewhere, usually you have an extruder holding the filament, so the ooze comes out the nozzle, but in your case, you plan to block the nozzle so the pressure will want to release upwards which will push the filament up and out of the heatblock into the coldside/heatsink. Changers like the XL and the Stealthchanger block both the nozzle end and the filament end of the hotend which contains the pressure
This is incredible. This is something that makes me excited. Can’t wait to see future iterations; especially as you add sensors to do automatic calibrations, filament checks, and docking checks.
Looking forward to it as well.
A tip to prevent hotends crashing, attach end stop switch near where the crash happens. Better yet, you could use an end stop switch to confirm that the hotend detached. If the switch doesn't turn off, stop extruder from moving and throw an error/ pause printing.
Great idea!
A small hall effect sensor will do the trick.
This is AWESOME!
Truly shows whta 3dPrinting makes possible.
Truly shows what happens when pattents and expesive manufacturing don't actively stop progress.
Thank you! This is why I love 3D printing, designing machines is way faster.
I have no need for multi coloured parts, I make parts to suit a job, & to work. Colour is immaterial. An amazingly well worked project. 👍
Thank you!
Impressive! I'm working on a toolchanger of my own, and the extruder problem is a big one. Speed is a huge potential benefit of this setup. The stepper is the heaviest part of most hotends, and attaching it to the x rail instead of the detachable tool (with the additional flex point of a tool mount, and further away from the x rail, exaggerating any flex in the system) should make it more capable of high acceleration.
One recommendation - angle one of the filament drive gears so more pressure is put on the filament and less on the drive gears. My testing indicates that the gears cause a lot of resistance, meaning wasted stepper power and a bigger extruder motor. If you can switch to a smaller extruder stepper without speed loss, you'll be able to accelerate faster.
Second recommendation: Go with a kinematic coupling. If you have a bimetal heatbreak with thermal paste, a 3010 fan will keep the v6 plenty cool, and long screws can be driven all the way to behind the v6 heatsink, touching the sides and holding everything more stiff. The kinematic coupling goes on a backplate. Stiffen the backplate with a pair of steel m3 screws. They do a lot. Each kinematic link can be made with a pair of slices of 3mm steel rod on the mount side and a small magnet wedged sideways in the tool. The magnet will sit between the rods and keep things stiff. Let the magnet sit low between them - it will hold things stiffer. You can hold the coupling together with a big magnet in the center, between the triangle of links.
Definitely comment with any questions!
Edit - my toolchanger maxed out around 100,000mm/s^2 accel, 600mm/s my last test. Toolchangers don't have to be slow!
Thank you! Wow, those are impressive speeds!
Can you explain more what you mean by "angle one of the filament drive gears so more pressure is put on the filament and less on the drive gears"?
My biggest concern with the kinematic coupling in this design was that it wouldn't hold the hotend for the sideways sliding motion due to the forces between the dock and extruder pins. This is why I went with the dowel pins for this first version. But this might not be an issue in some of the updates I'm thinking about.
@@EngineersGrow That's a good point on the sideways motion. What if you had a kinematic coupling *and* a pair of dowels to hold it straight? Kinematic coupling ensures precision, dowels help it line up right and keep it from sliding off sideways when mounting. It would be a pretty dramatic change though.
The two drive gears are parallel in most setups. But when your extruder squeezes against the filament, it flexes, so the gears probably end up being crushed together more than they should while you're trying to squeeze the filament. This happens with the Sherpa mini as well, and many others. There's a mod for the Sherpa mini that angles the drive gear so the gears are further apart and the part that touches the filament is closer together.
Adding dowel pins to the kinematic coupling like you're suggesting probably can work. I might try that. Maybe it will be as simple as making sure the pins have some clearance designed in so that they don't interfere with the kinematic mount.
@@EngineersGrow hope it works out! Looking forward to an update!
Very cool! Great job.
As some people mentioned adding a metal sleeved dowel that the pins go into would prevent premature wear.
Also a metal plate maybe where the bolt heads pull on the plastic when pulling the hot end off.
Thank you. Yes, I'm hoping to find a metal keyhole thats available online so that I won't have to make a custom one.
Wow another cool tool changer! Awesome. Thanks for sharing ❤
You're welcome!
Thats great, I can see this is going to the be a huge success.
Thank you, looking forward to see this project mature and grow.
My experience with TPU is, if there is a small gap between the extruder gears and the bowden hose it works for loading the filament with continous flow, but as soon as you print with retractions it finds the gap and will move that way, but lets hope for the best.
Thanks for the heads up!
its a beautiful idea, so much finesse. Cant wait to see this in future
Thank you.
This is a very good implementation, cutting the material waste and redundant extruders. And it's so DIY, love it.
Glad you like it!
Excellent excellent work. I have been working on and off on a fila-changer where the extruder (is supposed to) pick up the filament from a row of bowden flament docks at the back (just the filament, the hotend isnt changed) which would work very similar to how you have made this. Your sucess gives me renewed hope and some more ideas that may help me make my idea work. I know not changing hotend would mean the printer would need to prime the nozzle each time a change happens, but my goal was to stay within a single el-cheapo mainboard that comes with 4 stepper drivers and one/two hotend ports. Also getting perfect nozzle alignment with a toolchanger/hotend change is a pita without complicated expensive solutions. Looking forward to more updates on your design.
Great to hear, hope you get your idea to work soon!
Wow that's really clever! And very well executed proof of concept!
Thank you!
Get PCB way to sponsor you, use them to get metal brackets fabbed of your design, sandwich the metal brackets maybe between the pickup hub so the holes don't wear out and are more robust, ditch magnets if possible. Copy someone else's pickup hub design. Your nailing the idea totally and the extruder design!! Chef's kiss!!on your extruder idea. keep at it your SO close! Love the idea and videos keep it coming man!
Thank you, will keep working on it!
Very promising. In metal work we use bushing to achieve different properties than the surrounding material. Maybe a bushing made from copper or brass will make the trick. They are harder and more wear resistant than plastic but give also a good natural grease to steel counterparts like the pin.
Great idea.
I tried 3d printing and using a manual tool changer which had the fans and hotend assembly all mounted together, but removable from the gantry plate, and ran into a similar problem with the hotend wobbling because it was held up by plastic. I think thats a major design challenge: using any type of plastic for the assembly of a part that is directly tied to specific structural components that need to be exact to get quality prints. Might be best to CNC those
May i suggest looking at various space docking mechanics, and copying mainly the mechanical part of it.
Using probe/drouge or guides to achieve better alignment and better tolerance.
You can use different design for extruder to holder and extruder to hotend.
Flat head scred driver is kind a probe/drough , use funnel to guide the screw head and push it inside the funnel tube for hard capture
Thanks! I'll take a look.
I’d like to recommend copper tubing to solve your problem with wear in the alignment pin holes. Just some thin tube, put it all the way through the part and either flare the ends or cut and bend them to keep them secured. I’d say flaring them so the ends are securing the tubing in place and can still be flush with the part would be your best option. This will drastically increase the wear resistance and allow a solid fit for a long time. Also, instead of screws with a flange it may be best to get studs or threaded bar stock, cut it to size, then use some thread locker to glue some T nuts on the end. My last recommendation that may not work would be to use something like a magnetic push latch to try and hold the hot end still. It may not be strong enough or the latch may require too much force to push but it’s simple and easier to integrate than a whole mechanical latching mechanism so maybe it’s worth a try. You really have an amazing thing here and it seems like it’s fairly close to the point of being to the point of an actual first iteration. I’m really interested to see where this goes.
Lots of great points! Some sort of push latch that is simple and compact would be a great solution.
amazing work, please don't stop
Thank you, planning to keep going!
i would love to use this design on my own machine if im allowed. what a lovely design! good job
im pretty sure i can help you with the docking mechanism aswell as the attachment points to the extruder!
what if you use a solonoid on the extruder assembly to push and detach the extruder from the magnets. you should alsu use 3 magnetic balls in a triangle and pins to allign and mount the hotend
hell, you could use an electromagnet to mount the hotend to the extruder
Thank you! Yes you can use it on your printer and modify the design. My files are available on BuyMeACoffee. Link in the video description.
My goals for this project was to have a purely mechanical docking system to reduce cost. However, for a higher end tool changer one of the options you listed might work better.
I'm making updates to the design and send occasional update emails to those that signed up for the membership waitlist. Just sent an update email a few minutes ago.
@EngineersGrow got it!
You are doing an excellent job. Thanks for inspiring us all
You're welcome!
This is an incredibly different approach! I am excited about the next version. Metal pins and bushings can help with alignments and positioning.
Thank you!
Amazing, gotta love simple machines
Yes indeed.
Thank you for developing that and sharing it with us
You're welcome!
Great effort! Congratulations
Thanks a lot!
Needs a springed metallic sheet under each hotend dock, touching nozzle tip directly and blocking plastic from oozing.
Great point!
Project of the year!!! 🤯🤯 Amazing!!
Thank you!
This is how a tool changer should be. Having an extruder motor for each toolhead is wasteful and not necessary. With a redesign of this to make it look better, and function better, this could be one of the best and affordable options out there.
Thank you!
pretty damn slick bro!
Thank you!
Very ingenious. I like the thought of sharing the extruder and not needing any purge or poop.
You’d probably still want some sort of priming of the hotend after each swap, since there will be a bit of ooze or retraction in the hotend.
@@Leclaron True - but maybe just a little "turd" instead of a proper "poop"
I think I also seen someone implement a wiper the covers the nozzle opening when docked. Will try to find that design once this single extruder design matures.
This is unbelievably cool stuff!!
I can't wait to see how this project changes and improves over time. Hopefully there can be ways to make it lighter and more compact without sacrificing reliability!!
Oh, and it might be a smart idea to make a presence detection system using a hall effect / reed or (as wilkeiner said) a micro switch to prevent the machine trying to load a new tool while one is already loaded.
Other things you might want to consider would be adding brass inserts to anywhere where a metal pin locates a printed part and imparts a load on it. This will reduce the wear on the tool heads and prevent the side to side motion you've been seeing. While ABS is strong stuff, it's not completely immune to deformation and abrasion. In your case the toolheads suffer from both problems. You have the force applied by the locating pins when the tool carrier slides the toolheads across to dock them which will compress and deform the plastic over time, and you also have a degree of abrasion from the insertion and removal of metal pins into a plastic component combined with the spring pressure from the mechanism that docks the heads. I'm fairly certain this is where your excessive play in the toolhead that was printing the black filament has come from.
Again though, this is a fantastic beginning to a much needed project, and hopefully it will be adopted, improved and made into a standard so we can finally have cheap, decent quality multi-colour prints that don't waste 10 metric gigatonnes of filament!
Lots of great points! Thank you!
I am so glad that this works! I had this idea but I didn’t have the skills required. Prusa, hire this man!!!!😊
Don't be afraid to try! Many times the best ideas come after you're forced to learn skills you didn't have when you started.
why not for the alignment holes, use drilled out heat inset nuts, that way the holes cannot be worn out and are more stable. and my other idea would be a little dropper for the parking so that the spring in there isnt nedded could also be more reliable but idk if it would work
the spring seems way way too strong yeah, and the source of the rate of damage from undocking. countersinking the holes would also help so the dowel does not have to jump to the right location at once
instead of a spring I think you could have side magnets to align the retaining screws in either the locked or unlocked position, and then you just grab the hotend, move left, leave with it attached; and then put it back, move right, pull away without it
Agree, it will probably need to be a semi-custom solution. But inserts will definitely work.
This is an innovative concept. I'm glad you put the video up. I would suggest: You are trying to make parts needing good precision with a poor quality or poorly tuned printer. Everything needs refined with fillets and chamfers which will improve stiffness. There is nothing retaining the pin in the driven filament gear and perhaps other pins. The washer head buttons used to retain the hot end cassettes have no locking method to retain adjustment. The area under these heads is small which will lead to wear in the mating part and loss of adjustment. Cudos for not using any set screws, the sign of amateurs.
Thank you for all of these good points. Yes, lots of refinement work ahead.
This is gosh darn brilliant!
I now will be looking into this kind of mod!
I want to try and make a multi material Printed in Place lock. The options are endless!
Pretty cool. Hope you can make it soon!
Great job!! I think there're 2 things you can do to improve longevity on the docking parts:
1.- Round off the top of the locating pins, this will reduce friction, and thus wear.
2.- Have you thougth of lining the female end of the extuder mount with brass tubbing? It's cheap and durable. This trick may also help on the latching plate, were you to cut it out of brass (any time metal meets plastic, plastic looses).
Keep it up, I really want to see this through!
Thank you! I already rounded off the pins and didn't catch it in the final recording, since unfortunately it was too late. Planning to add metal bushings.
Pretty sure others pointed that out but usin acorn nutes as bearing surfaces against 2 pins and a star pattern to make the hotend only able to locate at a single repeatable point on the extruder head.
Also the release mechanisms doesn't need to be actuated by the hotend it can be a pin or a screw directly on the extruder head that pushed the locking bar that way the actuation doesn't load the locating surfaces.
You have a really good point about not loading the locating surfaces with the release mechanism!
In the current design, most of the loading should go into the bracket secured to the printer frame. However, I have some ideas to make the loading localized to the extruder.
I have concerns that the kinematic coupling might not work with the sideways sliding motion, but will probably look into it.
Was very interested to see this approach mentioned in Teaching Techs latest Video.
I really like this concept to avoid a fair bit of the wasted hardware in most tool changer designs - and also the removal of the need to cut filaments to get them to correctly feed into the extruder. (As per the BigBrain3d Swapper3d).
The additional cost and electronics involved in each nozzle having its own heater is balanced out by the benefit of being able to keep/preheat each hot end at an appropriate temperature to facilitate fast tool swaps.
My biggest concern is whether the size and weight could be scaled down to a) allow really fast printing and b) allow room for at least 12 hot ends to be loaded - to really complex multi colour prints - such as ship models.
To me the ultimate solution would have 2-4 fully independent print heads (with extruder motors) for parallel printing, with each print head able to colour swap using something like this engineers grow approach.
Perhaps backed up by some type share across multiple printers, cheap AMS type multiplexer too for access to a larger library of filaments than it would be practical to have a dedicated hot end for each one.
The solution would need to have a fair number of sensors in it, plus some type of auto calibration/alignment of the nozzle positions in 3 dimensions - perhaps with some sort of load cell arrangement in an alignment pin on the bed somewhere which could be checked at the start of each print, or maybe even after every tool change.
Very impressive 3D printed cruise fleet!
Faster printing speed will become the focus in the future.
@@EngineersGrow Thanks - the small 1:1000 scale up to 11 colour ones take about 15 hrs to print, and the bigger 1:500's can take over 40hrs. I would estimate that a nozzle swapping approach such as yours could reduce the print time by up to 4 times, and also reduce most of the waste. Plus could open up the possibility of moving to a smaller nozzle size.
@@dwuk99 Do you have any models that only need 3-5 colors?
Would you be interested in a collaboration?
Your cruise ships seem like a great case study.
Being able to switch between small nozzles for perimeters and large nozzles for infill is high on my list.
@ yes, happy to collaborate. I did downmix one of my models to 4 colours so that it would work on an AMS lite - p&O Azura on Makerworld.
Mixing nozzle sizes in the same print is an interesting idea, perhaps using the 0.2mm nozzle for highly detailed parts like deck furniture, railings and text, with the bigger nozzle used for the rest.
I think prusaslicer might support mixed nozzle sizes - will do some tests on the latest model I am working on - Cunard Queen Victoria 1:1000. I've had to remove some of the detailed parts that work ok at 1:500 scale - but could try re-introducing them with a smaller nozzle size just for the detailed parts to see how that affects overall print times.
Mixing nozzle sizes would of course increase the total number of tools required further - but this could be partly mitigated by splitting the models into separate parts.
Are you on teaching tech's patreon forum?- as there is a thread on there relating to your video that would be a fairly convenient place to post images of slicer tests.
I have reworked my latest Cruise ship model today to split out some of the more detailed parts for potential smaller nozzle printing, plus also down coloured it to 5 colours -
I have done a post on the TT Patreon with some stats, plus have also put a summary few of them under my 'Cruise Ship Print' thread on the BambuLab Community Forum - but this is more BL 1Q25 new printer focussed than nozzle changer focussed.
My findings so far though do bear out your suggestion that mixed nozzle capability may end up being more important in terms of print time savings vs the smaller than I expected savings you would get from tool changing vs colour multiplexing.
This is one of the rare times where a recommended video is pure GOLD!
Liked, subbed, hit and bell and all that, I would love to see where you take this!
Thank you and I appreciate your support!
@EngineersGrow you're very welcome!
I've been wanting to try add a 2nd head to my Anet A8, I love this single extruder approach.
Only having X axis movement it would need to be open on both sides, this would be quite a challenge but I mat try it down the line.
This is a fantastic project, I look forward to seeing it evolve :)
Thank you!
Would be interesting to have a printer that can change not only filament color/type, but size as well! I can imagine that'd be a pain to get a slicer to account for, but would be cool to do highly detailed areas with 0.2 mm nozzles / 1.75 mm filament and then switch to a 1.2 mm nozzle / 2.85 mm filament to fill in larger areas.
Interesting point about switching to 2.85mm filament!
I suggest to use one round hole and one slot to mate with the dowel pins for locating the hotend.
This might be easier than having to meet the two round holes.
The hole is now locating in XZ, while the slot is constraining the rotation.
Great idea! This is actually what I have right now, just didn't go into that.
I'd love a tool changer to print supports out of different materials for better results e.g. PETG supports on a PLA model. But as you say the commercially available ones are prohibitively expensive. Great to see more innovation!
Great, thank you.
Very cool project, enjoying watching your approach very much. So many design considerations for hardware and software. Impressive to see your endeavors. I love to watch folks work on difficult projects from start to finish. This type of enthusiasm and effort always leads to great improvements and offerings for the end users. Very impressive to watch how you make changes on the fly. Good stuff!
Thank you very much!
Ideas that came to mind from your wear issues:
1) Lining the female holes for the alignment pins with bushings, to give a hard-wearing surface
2) A slight re-think of the lock/unlock system - instead of torquing the hotend to the side each time to lock and unlock, perhaps have a sliding "locking plate" that grabs/releases the button screws? Actuate the plate using a mechanical toggle, like that used on latching pushbutton switches. That way it's unlocks when you run you go to pick up the hot end, but hitting the same mechanism a second time relocks it?
Not as simple, but it would remove a lot of the stresses induced during the unlock procedure...
Great points! Yeah, a toggle style mechanism would be pretty cool. I thought about it awhile back but it started getting complicated quickly.
That's orders of magnitude better idea than the enraged rabbit's, Prusa's, and Bambu's takes on the problem! Would love to see that come to life to a production grade level
Thank you! Looking forward to this concept going into production.
I REALLY hope Teaching Tech gets in contact with you. This on his SV08 project would be AMAZING
Sounds like that would be a cool project.
bro you are a boss. im a newbie and i love learning all these new things. im an inventor. i have easily lost 40k dealing with china. now im going to do everything on my own. great video. you get a sub brother. keep it up. thank you. this is above my pay grade right now but sweet to see someone so innovative and creative. thats what my business is made from
Thank you, I appreciate that!
It depends what you're trying to save money on. This design saves a bunch on hardware, but during operation it will have to clear lots more plastic to purge every time you switch plastics. So this might be cheaper to build but more expensive and slower to operate. Still beautiful design and quality work.
Bravo 👏👏👏
Thank you. Can you elaborate on why the plastic would need to be purged when switching filaments? Maybe I'm missing something. The idea is that there will be minimum material waste because there is still a dedicated hotend for each filament.
This shouldn't have more waste that Prusa XL. Filament never leaves the hotend.
Nice work. Good to see people still innovating.
Thank you.
Nice job putting in the work. I'd thought of this years back to simplify the parts list, but wasn't sure it would work, and didn't have the time to invest in trying it out. The other thing I'd thought of was having the gears dock with the extruder motor. It would mean a set of gears on each hot-end, but wouldn't have to catch the filament each time. Neat to know that the idea works though. I think you'll have to do something to lock the hot-end onto the extruder mount, otherwise you'll keep having play that will mess with offsets, but I could be wrong. Keep up the good work!
Thank you! Yeah, getting the gear teeth to align when picking up the hotend would probably be tricky without using encoders. The teeth on the filament gears don't need to align to each other, so it's more simple from that perspective.
I guess I'll be spending more money on another project now LOL! Jokes aside, this is such an ingenious design, top to bottom! Every design choice was brilliant and I loved the used of PTFE as a centering spring! I've been thinking about this idea of 1 extruder motor for 5 hotends for a long time and never had the time to tackle it. I think you'll also be interested in Mihai Designs' " pitstop " project which has a similar approach! Cheers
haha. Thank you! Yes, I did look at it and it's a unique approach.
That was amazing. Can't wait for tool changer to be cheaper thanks to your innovation.
Thank you! Me too.
That's awesome !
I'm curious to see if you can make a same system with a lower entry price 3D printer
What would you use an entry level tool changer for?
How important is fast printing?
I'm curious to know what the requirements should be
For the XY offset calibration, maybe a technique from pick and place machines can be borrowed. They pick up a part and then move it over a camera to figure out its exact offset and orientation to let it precisely place it, similar could be done with a camera and looking at the nozzle.
Then for Z offset calibration a technique similar to CNC machines could be used, likely even using the same parts, where they have a physical probe pad that they touch off on that makes electrical contact with the tool. Another way would be to just use nozzle bed probing but that requires nozzle bed probing.
All this could be bundled into a startup calibration routine where it picks up each toolhead in turn and measures their offsets.
Great ideas! It will be great to get to a point when all of the calibration is automated.
I think I've seen some videos of people implementing some of these ideas.
genius work! genius idea! This is how thinking should work! Not accept the things you can buy, but think about how to make it even cheaper, and better! Beautiful! Even tho i am really confused by your totally bad 3d prints (party for your changer) :D they look so bad ... ^^ BUT as long as it works, all fine! I will follow this .... please keep it up!
things i would improve:
1. small sensor that triggers IF an Toolhead is loaded (so to be sure if its even unloaded properly too)
2. Proper shaft / inserts so the slignment is perfect ( you can get this sooo cheap)
3. maybe sth to geht the z-height for each head properly...
Great ideas! What is "sth"?
Man! Thank you so much, I had a similar idea, because why do we need to have so many damn motors eh?! Your design is way more elegant than what I was thinking, but the concept of a grab extruder is the same. I had to no time to work on it, so thank for doing all the work and bring us into a more affordable future. Subbed.
You're welcome and thank you!
The idea is great. I think long term you need to think about CNC machined parts to increase durability and reliability. ABS might not be the right material choice in my opinion.
Thanks for your effort and sharing everything with the community
You're welcome. I agree metal parts might be needed. I'm still hoping to get around machined parts.
One thing you'll want to look into that the other tool changers have and yours doesn't is a kinematic mount. You should have all six degrees of freedom constrained (x,y,z translation and rotation). A kinematic mount lets you get away with less accuracy in your prints while still grabbing the hotend in exactly the same way every time.
Good point. I'll be looking at whether a kinematic mount will work with this design.
I watched this on the side, not really paying attention, wondering why this any better than tap-changer and stuff like it are already out there. But doing tool changing with one extruder is really damn cool! Very cool project! Love to see this develop further!
Thank you! Will do!
Great idea. For making it more reliable and cheaper I recommend David Malawey’s video or his shorts on “Borrow a Tolerance: Mindset for Designers”. He has some good ideas for functional 3d prints that get their reliability or durability from precise things like dowels, ball bearings, premade materials. I recommend watching his shorts (I never thought I’d recommend shorts) on it because they go over more relevant ideas than the longer video. The longer video can help with understanding the theory from an engineering pov.
Thank you for bringing up David's video. Lot's of amazing information in there!
I feel like a significant cost that has been missed is the electronics. Each hotend is going to end up needing at least one heater, thermistor and fan. This doesn't tend to be an issue for 2 tools, especially if using a mains powered bed, most big control boards have enough spare ports for that much. But, for more tools that won't be enough and you'll effectively need to have a toolhead board per tool, adding ~$30/tool. On the bright side, that'll also allow for all sorts of additional reliability improvements like sensors to ensure that tools have successfully docked/undocked.
That's a great point! My hope is that board manufacturers will provide lower cost boards for single extruder tool changers in the future. Each hotend will still need a heater, thermistor, and fan like you mentioned. But they can save money by only having one extruder stepper motor driver
@EngineersGrow yep, a custom board for that can be much more cost effective! I was also thinking that fan ports needed can also be reduced, there could be one fan on the toolhead for the active tool, and the ones not in use could be cooled by a single more powerful fan instead. Then you only need thermistor and heater power, so a multichannel ADC board and a bunch of mosfets, which can be very affordable for a large number of tools.
@@dotnet97 Great idea about reducing fan ports! Now that you mentioned it, the hotend fans are small, so it's possible to even have all of the hotend fans driven by a single port. This is how I added stepper motor cooling fans on my MakerGear M2 that I modified with a heated chamber.
Being a designer of my own toolchanger as well, its a very nice concept and idea of a toolchanger. Appreciate the effort put into this design. Wouldnt this design requires u to have 2 umbillical at all times? 1 on the carriage and 1 on the toolhead?
Yes. The carriage has it own cables in addition to the cables on each of the hotends.
This is really cool! Well done! I hope this matures into something more efficient and consistent.
Thank you!
Nice design! When I saw your design with only one extruder for multiple toolheads I thought about MihaiDesigns Pitstop (2) toolhead. Maybe you could get some inspiration there. Amazing work!
Thank you! I haven't seen it yet so will check it out.
Metal bushings press fit into your hotend brackets for locating the tool changer together will work with the metal pin design. Metal on metal movements should solve your durability and precision problems while keeping parts easily home-built. Everything will get a lot more rigid too.
Great idea! I will be trying that out.
@@EngineersGrow Cant wait to see V 2.0! If you need someone to demo it/torture test a future version, I'd love to help out and give some more feedback. I've been meaning to build a fully home-built clipper printer for a while now.
Sounds good! Remind me once I release the next video with the updated version.
wow what a cool diy mod! appreciate a lot! keep up the amazing work man
Thanks, will do!
Love the idea! You could try to use a hall sensor to see if sucessfull pickup. Also you could use 2 moves like the prusa, 2 directions could be easier for the Integration:)
Great points! Thank you.
Maybe some sort of metal insert for the pins to slide into. Like a heat set insert drilled out to be smooth with a flanged opening.
Genius idea. I look forward to seeing you progress.
Great video. Keep up the good work.
Thank you! Planning to try that out.
Wow I’m going to modify one of my printers to do this!!! Awesome!!
Great, go for it!
Hi A.S.! Great to see you have reached an interesting working solution. Keep up the great work : )
Thank you! will do!
In addition to metal sleeves and detecting tool presence (there might be 3rd, I already forgot, sorry), how about:
- improve tool stability by locking hotend plate to base plate with some kind of latch mechanism actuated by extruder opening arm? Locks hotend in place when moving away from the dock, but allows releasing the hotend to the dock. If possible, grabs hotend plate from the bottom?
- reduce pin and hole wear by moving X to release location against dock spring, pulling bolt heads just out in Y direction, and moving X back to release spring tension before moving completely away in Y. This way dock spring does not work against pins and spring-loaded "snap" does not try to carve holes during release
Prusa XL has also locking based on movement IIRC, maybe they have drawings or models released? Inspiration for parts that could work better :)
A simple latch mechanism would be great to use instead of magnets.
Interesting idea about the dock spring. It's a pretty light spring, almost too light for this application. I don't expect much wear to occur, especially with metal inserts. However, I also don't like it because it takes up a lot of space in the X direction. So I'm actually thinking about maybe how to get rid of it in the future.
Great work. Love it.
Thank you.
Hello.
Thank you.
I think that price of multi extruders is ok. We can use different extruders for different materials (abs, soft tpu and wax). So you shouldn't dance with only 1 extruder.
The main problems are changing mechanisms and reliability.
I do see your point about having an extruder for every hotend. I agree that reliability is the most important.
This looks like it's working really smoothly! Impressive af.
Thank you!
this is just so awesome
Thank you.
I have been going down a different rabbit hole to make a tool changer for my core zy machine. Stumbled across this video and I like it. Your approach makes a lot of sense and sure would like to give it a go.
Nice, go for it!
@@EngineersGrow We are working one it. Taking it one piece at a time to make my prototype. Started at the linear rail end and right now working on the docking stage!
Love it so far.
Hi. I am a automotive mechanical engineer from Austria and I love your base concept and I would love to be part of the design team - some ideas about improvements and new ideas are already coming to my mind as I am writing this message.
Awesome, feel free to comment with your ideas to improve the design!
I'll be launching a membership community for conversations like these as well.
This is an awesome project. It's an idea so simple, it's crazy that it hasn't been brought up more often (if at all - this is the first I've seen of this type of tool changer).
One thing I'm thinking for the docking, could you add some sort of button (like an old endstop) that when triggered, confirms that the tool head is docked? And if not docked after tool changing command, home the printer and try again? At the very least, I think it would prevent potential damage from crashing the tool heads.
Great idea!
Ever considerd a split-extruder design, where the gears are at the part that will be changed and just the motor is shared ? Makes the mechanical part easier
No open-close mechanism neccessary
I was thinking about that earlier. However, getting the teeth on the gears between the motor and extruder to align and mesh together seemed challenging. There's probably a different coupling that might work, but would need to do more research. So I moved on from this idea.
I want to see this work! Especially because there is a big challenge, that is because the extruder and the hotend are not rigidly mounted as one piece, there is the introduction of flexing/bending when extruder is pushing the filament to the hotend. I would love to see how that issue will be resolved
Hang in there. With some time and hard work it will work!
That’s a neat approach…kudos buddy. Well done, I’m sure this will inspire others too.
Thank you. Looking forward to see what others come up.
Great work, and very clever! Even if this proof of concept wore out quickly, it was still worthwhile for what you learned. If you can reduce the docking force required and improve the robustness and wear resistance of the interacting surfaces, the next iteration should work a lot better. Subscribed so I can follow along!
Thank you!
Excellent work
Thank you!
Thats pretty clever, i think you're onto a winning idea here.
Thank you! Looking forward to see where the community takes this idea in the next few years.
Good to see some alternative thinking - I think this might have legs. As I've mentioned before in YT comments, we need to standardise a tool change mount and this could be part of that. The mounts need to be based on a Maxwell kinematic couplings to ensure accuracy with a latch to secure. All doable with your concept. Would also be good to evolve the design for CNC as it would be much stiffer. Would be happy to chat/help if it's open source.
Thanks! I think keeping the hotend on the extruder while sliding sideways might be challenging with the kinematic mount. However, it might be less of an issue with future improvements to the extruder.
Such a cool design! Maybe a PCB way metal part is in your future for the docking piece.
Thank you! It's nice to have options where we can order custom parts.