My 6 input (5+1) mixing hot end



  • I've made a bit of progress on this if anyone is interested - details on my blog here

    https://somei3deas.wordpress.com/2020/02/24/my-6-input-51-mixing-hot-end-version-2/



  • Hi Ian,
    I like the design! I'll be following the progress to see how well it works.

    For my opinion, as you know they are like butts, everyone has one and they usually stink 😊 .

    The comment about possibly switching to flat head screws would in my opinion be the best choice. Not only for more room to tighten the heat breaks, but flat head screws would keep a more consistent alignment between the 2 plates as opposed to socket head cap screws. Maybe the alignment isn't that critical though?

    Otherwise a couple of tubular dowels on the screws would also insure repeatable alignment as well if necessary.

    Anyway great job so far!



  • @timcurtis67 Hi Tim,

    Thanks for the feedback. Alignment of the heat break inlet block and the combining block isn't ultra critical. The filament holes have a countersunk "lead in" and they step down in diameter as well - aligning the mixing blocks is more critical. I did think about dowels but everything is a bit tight and there is nowhere to put them unless I make everything a bit bigger. It might come to that in the end - we'll see. For now, I just made the bolt holes a really tight clearance so there isn't much freedom for the plates and blocks to move relative to each other once the bolts are in the holes.



  • WOW!!!

    I love the design. This was a beautiful read, thanks!

    Wrt thermal camera, I am very satisfied with SEEK https://www.thermal.com/
    low price (200$) and high resolution (higher resolution than FLIR sensor on first next available camera that cost ~50% more). It has "annoying" click (it self calibrates around once a second using a solenoid driven plastic cover to block the sensor, recalibrate and then open the sensor back so you hear that clicking sound of the solenoid while ti works) but the output, resolution, fps is rather good. Not comparable to $20,000 professional flir but usable enough that I use it to detect faults on a pcb.
    I tried also (don't own) a FLIR ONE, it has, IMHO, much better software than SEEK but much lower resolution so not that useful but FLIR ONE Pro comes with high resolution and great software but 200% the price (it is iirc $400)

    One question, you said you used stock alu plates, did you face them or? I'm just interested in how tight the few plates can be with only 2 screws to prevent leak between plates? Have you tested for leaks with air/water? Will aluminium stock hold and not bend?



  • @arhi said in My 6 input (5+1) mixing hot end:

    One question, you said you used stock alu plates, did you face them or? I'm just interested in how tight the few plates can be with only 2 screws to prevent leak between plates? Have you tested for leaks with air/water? Will aluminium stock hold and not bend?

    It's a long post so I'll forgive you for not reading it properly ☺ Here is what I actually wrote (copied and pasted)

    "The plates were all made from flat stock but then hand lapped to make them as near perfectly flat as I could. Initial assembly was β€œdry” but if there is any leakage, I could try using Boron Nitride paste between them."

    Oh, and if you look again at the pictures, you'll see that there are 4 screws, not 2.



  • I read it twice actually as I'm really interested in this topic, and I'm sure to read it few more times. Looks like this faulty SD card affected me more than I realize. I think I missed it because I don't recognize the "hand lapped" (had to check the Internet for it now, learn new things every day).

    Screws, I see it now, sorry, I was looking at that "grove" along and assumed it was screw after the whole post, I remember the 4 holes for screws earlier on the blanked images of the plates.

    Thanks for pointing me in the right direction. When you started talking about mosquito at the beginning I had a completely different idea (that I think can't work) but what you came up with I love!! Aaaaaand is waaay smaller than I anticipated. kudos.



  • @arhi No worries. I should have taken more care when I wrote the blog post. I tend to forget that some readers may not be native English speakers, so terms such as "hand lapped" may not be easily understood or may not translate well into other languages.



  • @deckingman Internet searc showed me that it is a pretty common term in machine shops, but I somehow missed it. Our local vocabulary is mostly influenced by german words with regards to the machining industry but TBH I don't know what would even be a local term for smoothening of the surface :D. I will recognise it if I heard it for sure but nothing comes to mind.

    More and more I look at this hotend I like it more. I'm sure you have seen those mixing nozzles for different 2 part and 3 part compounds. I never looked closely, I remember they usually have zigzag path visible from the outside, but now I got interested after reading your post, both about the "toothpaste" effect and about proper mixing :).



  • The next step along my journey is now live on my blog for anyone who might be interested https://somei3deas.wordpress.com/2020/02/28/my-6-input-mixing-hot-end-v2-part-2/


  • administrators

    @deckingman

    So I can switch between my working printer and the β€œtest rig” simply by changing the configuration file. That process only takes a few seconds using Duet firmware.

    πŸ˜€ πŸ‘



  • @T3P3Tony said in My 6 input (5+1) mixing hot end:

    @deckingman

    So I can switch between my working printer and the β€œtest rig” simply by changing the configuration file. That process only takes a few seconds using Duet firmware.

    πŸ˜€ πŸ‘

    I try to get you guys a mention in whenever I can. ☺



  • Have you noticed that I've changed my profile picture?



  • how do you plan to control this little 'monster' - what controller board, what software are you going to use?



  • @spllg said in My 6 input (5+1) mixing hot end:

    how do you plan to control this little 'monster' - what controller board, what software are you going to use?

    My existing Duet 3 that is currently running my 5 input Diamond will do just fine. I've has this in mind for quite some time so the machine has 6 extruders fitted, even thought I currently only use 5.

    Edit. This hot end itself only has a single heater, a single thermistor and a pair of very small fans that I can connect in series. So nothing special needed.



  • This might be of help, E3D made the heater that has wires that come out at 90 degree angle: https://e3d-online.com/high-temperature-heater-cartridges



  • @arhi said in My 6 input (5+1) mixing hot end:

    This might be of help, E3D made the heater that has wires that come out at 90 degree angle: https://e3d-online.com/high-temperature-heater-cartridges

    Yes I've seen heaters with wires at 90 degrees thanks (also, leads welded to pins and oversleaved, leads with stainless steel conduit, leads with armoured braid, and leads exiting at 90 degrees through a welded tube).



  • Hi Ian,

    I am keeping track of your blog for quite a while, but remaining quiet up to now. It is a pleasure to see what you have achieved with your printer.

    I am also using a multi color printer with 2 e3d cyclops+, so 2 nozzles with 2 color mixing each. From this, I have a rough idea on what might make sense to consider in your full color hotend/coldend design.

    Some ideas:

    1. Heat conduction. A 6 color hotend does not necessarily dissipate 6 times the heat, but each of your heatbreaks need to stay cool on the coldend side. And as there are 2 colors sandwiched in between, they do get less airflow. The mosquito heatbreaks help, but I am not sure this is enough.
      Consider water-cooling. I sandwiched a water-cooled aluminum block with my 2 cyclops coldends and that really made a big improvement. Especially when swapping filaments completely - when pulling back the filament, the cold side is cleanly separated from the hot(melted) portion and it does not lead to jams.

    2. Heat radiation. Make a silicone sock for the hotend. It is huge and it radiates a lot of heat to the outside. I would guess this contributes a lot to not being able to reach high temperatures with your 40W heater. Design a cast and get some silicone to do it. Did the same for the cyclops, should put the design online for the community...

    3. The 6th color input. I think this is the most important part of your hotend. What I really miss from the Cyclops is an effective way to retract. Doing a lot of retracts may kill the input which is not printing currently.
      Your idea to use transparent filament is a good shot a mitigating the issue (maybe the best), or you use an approach that does not involve any filament at all. E.g. use a rod of aluminum that you pull back within a channel close to the nozzle. But for this you need to somehow seal the construction and have tight tolerances. Don't know if that's feasible...
      The hotend is very long (like a volcano) and I am guessing this is to make the colors mix nicely, which is fine. For the 6th color, however, you should not attach the input to the top of the hotend. I am guessing that you have a separate channel to the nozzle only for this color, but this is not as effective as it could be for retraction. Consider putting a side entry (or mill a part of the block down) and attach a heatbreak as close as possible to the nozzle. Of course, this decreases the throughput of that input, but it improves retraction due to the reduced friction between filament and channel.

    This became a pretty long post, but maybe you find some useful input.

    All the best!
    Tommy



  • @Wurstkarton

    Hi Tommy,

    Thanks for the kind words and I'm always open to hearing other people's ideas. Ref points 1 and 2 - keep an eye on my blog. I think the next post (part 3) which is due to be published shortly will have some interesting data which might surprise you (it did me).

    The silicone sock is something that I have always planned - I made one to fit the Diamond 3 colour back in June 2017 - details here https://somei3deas.wordpress.com/2017/06/15/making-an-insulating-sock-for-the-diamond-hot-end/. The reason that I haven't yet made one for this hot end is that I might need to modify the mixing chamber which in turn, might mean that the outer size and/or shape might change. So if I made a silicone sock now, I might have to repeat that exercise at a later date.

    Your point 3 may be valid. But then again, it might not be. It's an idea but without any data, it can only be speculative. The 6th filament does indeed have a straight, vertical path to the nozzle so it might work well. The Diamond hot end works well with respect to retraction (one of the few things about it that does work well) and all the filament inputs have that straight path to the nozzle. So I'll test this design and see what happens. The 6th filament may also prove to have a secondary use for single colour, high speed printing (due to it's long melt chamber). So I prefer to stick with this design for now. Only by doing tests will I know if there is a problem that needs to be solved.

    Ian



  • @Wurstkarton With regard to your pair of cyclops hot ends, are you using firmware retraction? It's important that you do when using hot ends with multiple inlets and a single nozzle because retracting just one filament simply pulls filament from the other input, rather than the nozzle tip. And what extruders are you using? If they are the spring loaded tensioner design such as E3D Titans, they tend to wear away any filament that isn't moving forward for part or all of a print but still needs to be retracted. The fixed tension type such as Bondtech BMGs don't suffer from this problem. https://somei3deas.wordpress.com/2018/05/11/bondtech-bmg-vs-e3d-titan-extreme-retraction-torture-test/



  • Your explanation makes sense. Start off with your approach and check the result before going for something else.

    Yes, I use firmware retraction for it. 4 BMGs doing the labour. It still grinds through the filament sometimes, but mainly with some soft wood pla filament. Well, I guess I could revisit the topic and try a retraction torture test like you did to find out the limits. πŸ™‚

    I am generally a big friend of letting the firmware do the work. Firmware retraction, setting temperatures for all tools in firmware (not in the slicer, slicer just says T3, T2, etc., Duet does the actual work), extrusion multipliers, etc. Not only can you print the same gcode on arbitrary tools (if it is a single color print), but also do testing for many different settings without reslicing.

    For me, the whole multicolor printing topic is kind of bottomless, meaning that it is not ready for the final consumer. Especially your shot at fullcolor is challenging - both hardware and software(slicer) side. I am wondering if there is a slicer that lets you put full color on your object (like a bitmap)?
    Quality wise you somewhat degrade by having to use a Bowden setup, which you countered by using a second xy axis. This is something that I am not willing to do and hence I am thinking if it might not have made more sense to use a prusa MMU with direct drive extrusion in my case.
    But I have invested a lot in the current setup and it's working and so I think I should stick with it. In the other hand knowing you may perform better is nagging. πŸ˜‰



  • @Wurstkarton said in My 6 input (5+1) mixing hot end:

    For me, the whole multicolor printing topic is kind of bottomless,

    That just about sums it up - having masochistic tendencies helps too ☺

    I am wondering if there is a slicer that lets you put full color on your object (like a bitmap)?

    Not that I know of but this https://hackaday.com/2019/08/20/lithophanes-ditch-the-monochrome-with-a-color-layer/ is on my list of things to investigate

    Quality wise you somewhat degrade by having to use a Bowden setup, which you countered by using a second xy axis.

    This new hot end doesn't have the huge great cooling fan mounted on the top like the Diamond, which makes it quite a bit shorter. So maybe I will be able to drop the height of the upper "UV" gantry which will reduce the length of my Bowden tubes even more. I might be able to get them down to 100 - 150mm or so - we'll see.... The other thing is that the Bowden tubes on the Diamond exit the heat sinks at an angle of about 20 degrees, which means that they have to be curved to line up with the extruders. With my new hot end, the tubes should be very much straighter which should also help.

    This is something that I am not willing to do

    Have you considered using "remote" flex drive extruders like the Zesty Nimble ?

    and hence I am thinking if it might not have made more sense to use a prusa MMU with direct drive extrusion in my case.

    AFAIK, the Prusa MMU does not mix. It just allows you to automatically load different filaments but you can only print one at a time. And have you seen the size of the purge tower it produces?

    But I have invested a lot in the current setup and it's working and so I think I should stick with it. In the other hand knowing you may perform better is nagging. πŸ˜‰

    ☺ Which brings us back to your first sentence "..... the whole multicolor printing topic is kind of bottomless".



  • @deckingman Kiss slicer supports applying colour via a bitmap image now and supports unto 12 extruders now

    Kiss features



  • @Dougal1957 interesting. I was originally on slic3r&cura, but now switched to iceSL for having the possibility of coding parts of the slicing myself easily in lua.

    How about kisslicer in this regard?



  • @Dougal1957 said in My 6 input (5+1) mixing hot end:

    @deckingman Kiss slicer supports applying colour via a bitmap image now and supports unto 12 extruders now

    Kiss features

    Wow, that's interesting! I really must try to find the time to start playing with it.



  • Part 3 is out now for anyone who might be interested https://somei3deas.wordpress.com/2020/03/09/my-6-input-mixing-hot-end-v2-part-3/


Log in to reply