Well it sucks the weekend is over and I have to go back to work tomorrow. But we are starting to look like a printer again..https://i1050.photobucket.com/albums...oWiz/115_4.jpg
Printable View
Well it sucks the weekend is over and I have to go back to work tomorrow. But we are starting to look like a printer again..https://i1050.photobucket.com/albums...oWiz/115_4.jpg
I have finally got the tops of the towers situated and figured out. One thought that I have struggled with is the weight I am adding to the gantry and the fact that i replaced the lower metal frame ends with printed plastic parts. I am or was afraid that the weight of the gantry might crush the plastic and change z height on 1 or both sides. Here is the lower bearing and mount I am talking about..
https://i1050.photobucket.com/albums...oWiz/117_4.jpg
So I have the metal upper plates flipped upside down so the bearing goes in from the top and I have the collet on top of that so the top of the frame is what is supporting the weight of the gantry. In raw form here is what that looks like..
https://i1050.photobucket.com/albums...oWiz/116_4.jpg
So i made some pieces to cover the metal plate so there is just one shade of red on my printer. The only exposed piece of the metal plate is the stop for the gantry so it doesn't climb up and out of the printer..
https://i1050.photobucket.com/albums...oWiz/118_3.jpg
https://i1050.photobucket.com/albums...oWiz/119_4.jpg
And here are the tops completed and together hopefully for the last time..
https://i1050.photobucket.com/albums...oWiz/120_5.jpg
Here is the backside of my printer.. https://i1050.photobucket.com/albums...oWiz/121_3.jpgI need to create some type of mount for the SKR 1.3 board, the mosfet, the 2 stepper driver expansion boards the tft35 v3.0 screen and the buck lm2596 step down converter. And it all has to look clean.
So I had to do some backtracking. The parts I made to attach the gantry to the linear rails on Z were insufficient and would just slide up and down and off of the metal plates they were supposed to be holding. So I have redesigned these parts. each took just shy of 7 hours to print and the machine uses 4. But that is all behind me now and my machine is really starting to look the part..https://i1050.photobucket.com/albums...oWiz/122_3.jpgAnd from the back..https://i1050.photobucket.com/albums...oWiz/123_4.jpg
I am gonna get a delta CA. When this is done I am gonna get a Tevo Little Monster and rework it to be a 3 color mixing printer with the diamond hotend. And I will have 2 completely different ways of chasing large color prints. I just gotta get through all this first.
So today i have the belts on and adjusted and i have the extruder mounted. From here i plan to move the drag chain from on top of the gantry to underneath the gantry. This is because the Prusa MMU2S unit now sits where the drag chain used to be. And it will look cleaner anyways to bring the wires down and under the gantry..
https://i1050.photobucket.com/albums...oWiz/125_3.jpg
So I have been busy with reworking my printer albeit with a gross lack of design skillz. But here is my progress anyways..
https://i1050.photobucket.com/albums...oWiz/126_3.jpg
Since I am now going to be running a 32 bit controller I am going to swap the drivers for X, Y, and E with LV8729 and chase the 1/128 microstepping. I am going to leave Z with the TMC2208 for now.
Ok so I am trying to cope with the linear guide costs but I already have my correct parts for Y. So I am working on completing that axis. I am turning the bed sideways so instead of 370 or 380 X 250 I can get the full 400 X 250 build volume. Also I am applying linear guides so I need a very specific piece to attach the mount/ base plate for the heat bed to the linear guides. And here is my design.. https://i1050.photobucket.com/albums...0mount%202.png https://i1050.photobucket.com/albums...0mount%201.png
My new stepper drivers showed up. I got 5 LV8729 drivers that I have every intention of running in 1/128 micro stepping. 2 for X, 2 for Y, And 1 for E. Z already has so many steps per MM that I do not know if it will be beneficial for the extra cpu effort required so I am going to run 2 TMC 2208 drivers for the Z motors. Because they are silent at 1/16 micro stepping.. https://i1050.photobucket.com/albums...oWiz/133_3.jpg
I learned a hard and expensive lesson on linear guides. The right answer is expensive and to be perfectly honest anything less and you are better off saving your money. I bought the cheap ones first. And then I bought the genuine HIWIN. And while my butt hurts immensely, life has never been so good.. https://i1050.photobucket.com/albums...il%20color.jpg
So after I got the guides sorted out I finally made a real mount for the Y axis drag chain. For the longest time I just had the drag chain zip tied to the bed screw/spring. But now I have a part I can attach the drag chain to with 3 screws like it was meant to be.. https://i1050.photobucket.com/albums...oWiz/135_3.jpg
So my bed is going to move more now that it is turned sideways. 400mm travel instead of 250mm. And with the MMU2S unit hanging off the gantry it is clear that the ptfe tubes will have to make sharp bends to avoid the bed at the start of prints. So I am thinking about making a bracket and hanging the MMU2S off the gantry still but just at 90 degrees so the PTFE tubes are off to the side.. https://i1050.photobucket.com/albums...oWiz/136_2.jpg
My last order of HIWINS is on the fedex truck out for delivery right now. I am literally standing on my porch staring at the watch I dont have on my wrist and tapping my foot. lol. And then I got some backtracking to do to get the gantry off and the z motors off so I can change the 4 rails for Z. Hopefully at some point here today I will produce a picture of my Black Widow with 8 HIWIN rails screwed to it.
so what made you go for the mmu2 over the mosaic palette 2 ?
On balance the mosaic would seem the better option as it will blend colours As well as do multi filament selection, rather than just select single filament types like the mmu2. Also only a couple hundred bucks more. Or about 6 inches of linear rail :-)
I am going after multi material from 2 different angles. As far as multi material goes, the MMU2S takes in 5 filaments and it costs $300. For twice as much your mosaic takes in 4. This will be my printer to print in many different filaments. When I am done with this build I will most likely hunt down a little monster and apply a diamond hotend to it with my knowledge from today and get a mixing printer with 3 colors. I mean, lets just be real for a second about mixing. All colors come from 3 primary colors. Don't you remember this?
https://i1050.photobucket.com/albums...toWiz/1_55.jpg
3 primary colors. So then if we are mixing what is the 4th filament for? If we are mixing we need 3. If we are chasing multi material then we can have more than 4. And for hundreds of dollars less I might add. And from a big name like Prusa Research.
I respect you, your thoughts, and your input CA. And as such I have done some further research on the palette 2 and I still feel the same. But for a new reason I want to share with you. For it to take in 4 filaments and feed one out to the hotend and be able to handle mixing can only mean that this machine has it's own hotend built in and it is melting and squeezing out new 1.75mm diameter filament on the fly. While this is a cool feature on the surface my 24v printer is already up to a 600watt power supply. We must consider power consumption when we are dealing with machines that can run for literally days at a time. Of course this is all my thoughts and ideas on the subject this side of a working model or prototype.
the 4th element is for white. Think about inkjet printers.
4 colours because the white is already present in the paper and you need black to adjust for shades.
Although - it would actually need 5
In life, the white is represented by - well light itself.
In 3d printing you need white and black to make shades of colours. The three primary colours themselves will mix any colour - but the lightness and darkness of shades will be fixed. The white and black are there to change the 'brightness'.
So ideally for a full colour mixing fdm setup you would have black, white, red, yellow and blue.
Quite how you would blend actual filaments together - no clue.
The best option at the moment are the inkjet solutions like the xyz davinci colour uses. A white filament that's coloured with the three primary colours and a black.
That said, looking at the mosaic, it doesn't do blending after all. Pretty much the same as the mmu2s, one colour at a time.
But I do like the way it cuts and shunts the filament into the system as it actually connects each piece of filament to the next with a proper melted joint.
I guess the advantages over the mmu2s would be mainly in speed as you just get a continuous length of filament rather than all the stopping and starting and retracting with the mmu2s.
Each length is precision cut and joined to the next in line.
So no it isn't actually extruding new filament - but it is heat joining the different lengths.
That makes no sense on how we get gradient colors from a section of one color than a section of another. There must be a melt zone inside the palette 2 to create these gradual colors. The nozzle does not mix UNLESS multiple filaments go into the hot zone together. That is the only way we can achieve new colors is by mixing the filaments. not a binary red than yellow then red. Because we will see red and yellow and red in our prints rather than orange. Because nozzles hold a very tiny amount of filament and surely not enough to mix sections of filament. Not even the diamond hotend has a melt zone that big.
So I decided to mount the MMU2S on the gantry so the PTFE tube going to the extruder can be as short or shorter than it is on the Prusa MK3 with the MMU2S. But I do not want the 5 PTFE tubes to exit the rear of the printer. This can cause interference issues with the heat bed. So I designed this bracket to orient the mmu perpendicular to the gantry.. https://i1050.photobucket.com/albums...mount%20v3.png https://i1050.photobucket.com/albums...oWiz/151_1.jpg https://i1050.photobucket.com/albums...oWiz/152_1.jpg
palette 2 doesn't mix colours.
I had a good look at their website and it's just a multimaterial machine, like the mmu2 but cleverer (and I suppose, potentially more to go wrong) :-)
The machine is starting to look impressive !
What's it going to be called ?
i don't know yet. But it is so not a Tevo Black Widow anymore. I will think of something. So all this talk about the palette 2 and the mmu has given me an idea. But one that i am not capable of completing by myself. I am no master at writing code or programming. And i struggle with it. But i can physically build one large format printer with a palette 2 being fed by 4 MMU2S units for a total of 20 filaments into one hotend. But i wouldn't know where to start for controlling software. But i sure can design build and wire it.
I need financial backing to flesh out some of my bigger ideas. Can you imagine the palette 2 + 4 MMU2S units? that is almost $2k right there.
So after the afternoon with this thought I am left wondering if it is possible to just daisy chain MMU2S units. like 6 MMU2S units. 5 feeding one that feeds the extruder. 25 filament potential. All controlled by one mainboard. This setup would require careful planning and building to keep all PTFE tubes as sort as possible. But how awesome would that be? all different types and colors of filaments. That should open the playing field on what we can make with our printers.
well, it would probably make more sense to build a spool handling machine instead. Plus can you imagine how slow a 6 mmu setup would be ? Also a lot to go wrong.
I guess if you had enough space and a large cabinet to put everything in, it might work. But I reckon I could probably change a spool faster than 6 mmus could get filament to the actual hotend :-)
The man to talk to is the guy who builds all the bizarre machines for duet 3d. That guy is an engineering genius.
He's also proved that the duet board family can handle just about any printer configuration conceivable by man. Plus iof you used a duet board, they'd probably help with all the programming :-)
You'd need to be using amf files and some kind of specially customised slicer.
Good luck :-)
If you really want to kix things uop - why not go with an idex system, and link mul;ti mmus to each head :-)
Actually I'm surprised you haven't gone down the idex route yet. would seem the natural next step up in the evolution of the Brazillian Huntsman (worlds most venomous spider, and one of the most aggresive).
Or 'Brian' as I can't help thinking of it ;-)
I think or at the very least hope I am going to seriously impress you when i am done with this printer. The MMU2S is slow when it is printing out a wipe tower behind the item it is printing. But when it is using wipe to infill it is notably faster. The high quality linear guides are also all about turning up the print speed. As with the 2 stepper motors for each axis. And the shorter PTFE tube from hanging the unit off the gantry will also help but not as much as connecting everything to a 32 bit controller and running the MMU2s off of marlin 2.0. It will also run on 24v. This MMU2S should behave a little different than the rest. But time will tell. I will post videos as I chase print speeds with it. I have been against Idex because it is only 2. Just a single step above the most basic of 3d printers. 3 with mixing or 5 in my mind are the minimum standard we should demand here at the end of 2019. But the thought of an idex with each extruder being linked to an MMU2S is a powerful idea. And thank you for that, CA. But I have to see this MMU up and running first. I still have absolutely no experience with this Prusa MMU2S. It just looked real cool and like something I wanted to try. I need to assess for myself whether the MMU2S is good or bad for me.
...Or an idex with 2 diamond hotends. 2 independent sets of mixing extruders. That might lead to some cool creations.
I will only play with open source. Custom slicer to me sounds very much like a software that will stop receiving updates at some point. But open source promises I can just go find another software that is properly maintained. Even as businesses and trends rise and fall. We should all be very cautious about setups that run 'custom' firmware or slicers. Typically printers built with this tech resell on ebay for dirt cheap. There is Absolute Beauty and Perfection in Mainstream Open Source. Can it run repetier? Can it run simplify3d? Can it run Prusa Slicer? Can it be run by octoprint/Cura?
Thank You CA for your creativity in suggesting a name for my printer. I have been struggling with this and maybe 'Wandering Spider' shall be the new name of my Black Widow after this round of extreme reworking.
or Brian ;-)
The thing with idex is not that you can use 2 materials, but more that you can use 2 completely different meterials with different print speeds and temperatures.
So you could have one head hooked up to tpu, nylon, polycarbonate, soluble material etc and the other to a squillion different colours of pla.
With a one nozzle multi material machine you'll always either be waiting for the nozzle to heat up or cool down. Just inefficient.
And yeah fascinated to see just what kind of speeds you can get from Brian :-)
The time for temp change can be helped. There is a dwell setting or I think marlin calls it residency where you tell marlin how long you need to see that extruder holding a temp before it is ready to print. We can also adjust how lax this temp setting is. Whether you want the extruder to hold 235c for 10sec within 2 degrees of target might take a lot longer to achieve than 2 seconds within 5 degrees of target before resuming the print. And turning on the layer fan can help bring down extruder temps a little quicker. So I would just need a way to command the layer fan on at 100% while waiting for temp to decrease. Or maybe I will setup another fan just for that. I have to get to a starting point first, but I look forward to bumping into obstacles. Otherwise this would all get real boring real quick.
OK Let's talk electronics. The Black Widow came with this Control box and all these wires with connections and it was just a mess. One goal of this rework is to do away with the control box and mount all the electronics directly to the frame. I already have the power supply mounted and I chose the space between it's support and a frame rail to mount the Mosfet.. https://i1050.photobucket.com/albums...oWiz/138_2.jpg https://i1050.photobucket.com/albums...oWiz/141_3.jpg https://i1050.photobucket.com/albums...oWiz/156_1.jpg
On the other side of the rear half of the frame is where I am mounting the SKR 1.3 mainboard and the 2 stepper driver expansion boards.. https://i1050.photobucket.com/albums...oWiz/153_1.jpg https://i1050.photobucket.com/albums...oWiz/154_1.jpg https://i1050.photobucket.com/albums...oWiz/157_1.jpg
So, given all the other expense - why go with a dirt cheap controller board ?
I mean I've no idea what the top end boards do for you, but I would think it's kinda like spending 3 weeks pimping a car and then fitting it with lawnmower engine.
Also on the 'pick an arachnid' front - the australian Huntsman spider, while not venomous is frigging huge !
https://www.youtube.com/watch?v=wRQucp31n0c
The thing is eating mice in his house and he says: 'that is so cool'.
I'd be going for the shotgun !
Okay this video is not for anyone who's nervous of spiders - you were warned !
https://www.youtube.com/watch?v=u93jSjd2sX8
So as i understand it all of these control boards are pretty much the same. With the difference being whether it has an 8 bit or a 32 bit controller. Past that there are features on the board that we all like. like integrated or changeable steppers or 5 or 6 stepper driver sockets. But in reality with expansion boards this is totally a non issue. And even with the best board i would still need an expansion board, sooooo. Why not save money on the same part? All 8 bit boards are the same and as I move to 32 bit I seriously doubt I will see any difference with a duet or a skr pro. If there are other differences or advantages that I am not aware of I would be interested in learning about them.
I am sure you have already seen this but this could be helpful https://youtu.be/W6U-u0wBdfQ
No I had not seen that video. Thank you for sharing. If you guys would believe there are 8 bit boards that are selling for over $100 today. Boards like the Panucatt Azteeg X3 Pro. I am using extra stepper drivers but I will never use most of the features that set that duet 2 and 3 at their price point. Tool changers are for corexy and do not translate so well to reprap. But if there is one thing we can all be absolutely sure of it is that this will NOT be the last configuration of this printer. And if I am anything less than utterly impressed with the SKR 1.3 I will change it. But I choose my parts by specs. And the ~$20 SKR 1.3 has very similar specs to all the more expensive controllers. I think we all just need to be able to better understand what specs are and how they should apply to our selections. Ya know? The best specs isn't always the best specs. Sometimes the best specs are what is right for what you are doing. Even thou there are better specs out there if you will never take advantage of them then you have built a mismatched printer. And that is just not what I am all about. For me and all that I do and build Form Follows Function. I strictly adhere to this golden rule of building and I think my results have shown as much. The $1000 rails started as cheap ebay rails. They met the specs but when I got them they were trash. I never ran them and tried some others and then went with the genuine HIWINS. Hopefully the control solution wont go the same way, but I am not against 3,4 or even $500 spent on a control solution if that is what it takes. I am already somewhere between $2 and $3k into this machine since and including it's purchase at the start of 2017. I have no problems putting money into it. But it needs to be the right part. Not just the most expensive or the part with the most features that do not apply to what I am building. With that said, you guys are scaring me. The SKR 1.3 is a 32 bit board for $22 in a sea of other like boards for over $100. Well this STARTED as a rework on a budget, anyways.
Have you guys seen the MMU2.0 auto deplete function? https://www.youtube.com/watch?v=pyHUdQ-sSGI