V large 3d build advice

Hey everyone,
I am looking for some general strategic advice and to be pointed in the direction of information to help flesh out my build idea.
I'm newish to this. Thank you in advance for any help.

I need to build a printer that has a 5' wide 1' deep by 4 ' tall print area (I know this seems crazy and no I can't print in pieces.)
the printer should be able to run for a 30 day print time.
the printer should have a pause feature...see above long print time
I would love if it had quiet steppers and drivers.

I have a little experience and have built a light object 100w laser cutter from the ground up and have spent the last 6 months 3d printing.
I have a background in Industrial design. Essentially I know just enough about things to be able to understand how stuff works but not in super depth.

My plan to achieve this was going to be to take the extra Creality cr10 I have and use it as a starting point.
1. buy more 80/20 and expand the print bed
2. swap out the motherboard inside the control box with a 32 bit board with more steps and quieter drivers.
3. alter the marlin print drivers that are standard for creality to use the new print bed size
4. lengthen the cables to match the new print bed dimensions.
5. switch from Bowden to direct extruder for ease of not having a 7 foot Bowden tube
6. have a piece of aluminum cut and drilled to be used as the new 5' long build plate.
7. find a 4' tall lead screw

So in my head, this seems like the logical thing to do, start with something you know works well and alter it. Am I oversimplifying this?
I have never done this before, I am open to any thoughts of what I am overlooking, missing completely, not considering etc..

where is a trusted source for 80/20 and pieces for the new frame?
lead screw for z axis supplier?
should all axis use lead screws?
Can nema 17/23 steppers handle being on continusly for a month?

Cheers
Ez

You can't scale up a cr10. There are so many reasons why you can't its almost not worth listing, but give a quick Google on the deflection values of the materials you want to use and you'll soon find out one of the reasons. You will redesign so many components that you may aswell start the design from scratch.

What material are you printing in? The shrinking and warping of almost any plastic at those sizes will need serious consideration. Will you heat the chamber? Or go for experimental methods to hold the printed pieces? Digital dentist has done some experimentation with using disposable foam sheets as a build surface.

Hangprinter or black belt style printers may be better alternatives?

I think cartesian printer with 5' y and 1' x axis with, supported, linear rails is your best bet. Regarding CR10 - have you considered the mass of the printing bed and how you are going to move it?

Thank you all sooooo much for the responses!!!!!!
This dialog is just what I needed. I am totally capable of researching and building a thing once I know what the known's are but here I have been completely out of my element.

I am printing PLA in thin pieces no thicker than 8mm .2/.3 mm layer height
For this reason I started out thinking I should make a cartesian printer.
Also for the base plate I figured I would do dual, belt driven nema motors.

Not sure about heat yet since I technically don't need it to stick to the base and with long print times like a month It would be nice to avoid making all that heat and using electricity if possible.

I looked at the hang printer and was all about it for 5 minutes until I saw the high mass and torsion of the print head will not allow it to be detailed enough to do what I need it to do.

Black belt looks like it could do exactly what I need but holy smokes looks very very expensive.

Quote
chromaglow
Thank you all sooooo much for the responses!!!!!!
This dialog is just what I needed. I am totally capable of researching and building a thing once I know what the known's are but here I have been completely out of my element.

I am printing PLA in thin pieces no thicker than 8mm .2/.3 mm layer height
For this reason I started out thinking I should make a cartesian printer.
Also for the base plate I figured I would do dual, belt driven nema motors.

Not sure about heat yet since I technically don't need it to stick to the base and with long print times like a month It would be nice to avoid making all that heat and using electricity if possible.

I looked at the hang printer and was all about it for 5 minutes until I saw the high mass and torsion of the print head will not allow it to be detailed enough to do what I need it to do.

Black belt looks like it could do exactly what I need but holy smokes looks very very expensive.

If you think the blackbelt is expensive, you may want to reconsider your project.

My current coreXY has set me back over £1000 easily and it has a standard-ish build area of 300x200x500mm build area. make sure to cost this project well as it will quickly get expensive. Larger frame members, custom parts, nema 23 motors all come at a premium.

When I say shrinkage and warpage, it is a serious consideration. If you have a standard style cartesian, your longest continuous lengths of extruded plastic will be 150cm. If this has just 5% shrinkage it will shrink by 7.5cm. When this shrinks, it will either rip away from the build plate, rip the layers apart or bend the bed. Even PLA at room temp shrinks after being extruded. We control this primarily with heated beds and chambers. The strength of warping plastic is not something to scoff at..... The other side is that you simply don't care about these warping factors as you are printing so large anyway. This is a legitimate train of thought -- extremely large injection moulded parts like bins are allowed to warp as it is within spec of what a bin does.

The blackbelt has a terrific idea as the longest lengths of extruded filament are now in the smallest plane, rather then the longest.

What are you making? Have you considered other styles of production? Vacuum forming, casting, Foam CNC or automated hot wire cutters could be a better solution.

Well I should say I'm not opposed to spend money to build a project, I have built laser cutters from kit and understand the costs Involved with projects like these. With that said I Feel like there is a difference between building a custom machine you that does what you need for $4k vs a store bought machine for $11k.

I'm on board to get this done with what it costs, I'm more interested in building this machine rather than buying it ready to go because I love the learning as well as saving money.

hmmm, I have seen several companies online that sell larger cartesian machines like 3' squared perhaps a good test would be to find someone who has a larger machine and have them print part of a file and see how it goes with the warping.

My cr10 has a 300 mm print bed and I have made parts that span that whole length without issue.

Edited 1 time(s). Last edit at 02/09/2018 04:07PM by chromaglow.

I mean when I think of it building a giant coreXY seems like the best solution. less problems with base plate mass, easy to box in for heat, if I had to. Anyone? does this make sense ?

I'm not sure coreXY would work over larger distances. Would belt stretch be an issue? At the very least getting the pulleys all on the same plane would be difficult.

With such a large machine though the weight savings of CoreXY will be very minimal (your XY gantry may weigh a few kgs, so adding a 300g motor is not a big deal. Perhaps a rack and pinion or belt and pinion design would be easier to implement? Less issues with belt stretch.

Would the bed move in Z? The bed could become insanely heavy. Perhaps the hot end should move in Z? Maybe a CNC style Z axis?

Edited 1 time(s). Last edit at 02/08/2018 07:08PM by Origamib.

Here's what I would do. This is by no means the only way, but feel free to use the ideas:

Standard Cartesian motion, nothing fancy. My gut feeling is that trying to compensate for the huge belt distances in CoreXY, and then trying to align the pulleys so that the belts don't ride up will be your worst nightmare.

To that end, I'd create a rack and pinion design for the longest axis, the shortest axis will be your gantry, this should not be difficult as its only 1 foot wide.

The Z axis will be the next problem. Will the bed rise in Z? How heavy will the bed become? A lightweight bed will almost certainly flex over a 5foot distance, a heavy one will require extra engineering to lift. So perhaps the bed should be completely stationary? Buy a huge slab of aluminium, and shim it until its completely flat. It will never move and you will likely never have to re level it ever again. in this case, your Z axis will now have to be the X gantry (to make a CNC style XZ gantry) or the entire XY assemblage will lift in Z. This is perhaps not an easy task, as an XYZ gantry would have the same problems of rigidity weight etc, although it would certainly still be lighter then the bed.

The CNC style XZ gantry is not without its problems either, as when it is fully extended the machine will be 8foot high or more. Do you need a 4foot high build area?

I received your PM about what you are making and I assume you don't want me to blurt it out but your product could also be made with a CNC engraver using thin sheets of plastic.

Origami,
I think everything you said sounds like as close as we can get to a KIS design.

Ive helped build a long axis cnc for a friend who is a ski builder and needed an 8 foot long bed.
This should be similar and I can understand that.

I can build an 80/20 base set a piece of aluminum or polished marble in it and shim it reasonably and then print a substantial 5/6 layer thick raft before each print to even things out a bit further.

At work now, more research to do.
I can not convey to you how much I appreciate everyones help thus far just wrapping my head around this.

Sounds good I'm happy to help - - always interesting to wrap your head around new problems.

It may be worth looking into the disposable foam bed idea developed by digitaldentist. The nozzle plunges into the foam for the first few layers. The plastic makes a bond with the foam and the straightness of the gantry dictates how flat the print is, so no need to level the foam bed. The first layer is of course ruined as the bond between foam and plastic is not pretty, but this is easily sorted with a raft. I've heard that this technique is used by some high end stratasys machines when printing large parts.

The next consideration will be protecting your printer from common faults that normally cause a print to fail. Power outage, filament run out, jams, filament stripping etc. These are not huge issues on normal prints but all could be extremely likely on a print taking days or weeks.

Duet WiFi supports resume after power outage and filament monitor is in the works as well I hear which will solve most issues

Sooooo where we last left off...

It's looking like I will build a v short 8/20 torsion table, think 6" tall and 5' in length.
Inside that table I will end up using 1/2 inch polished marble for the print bed in 1' squares so each square can be leveled via shimming underneath.

The long axis will be rack and pinion cnc style and use dual steppers to make moving the XZ bridge mass easier over long periods of time.

It looks like using the duet wifi will indeed be best option for many reasons including the print resume feature, nice trinamic drivers that are 2.8 amp to pair with nema 23 1.8 degree steppers and remote control.

I will use a titan direct extruder above a e3v6 volcano hot end.

All this will be powered by either
[www.amazon.com]
or
[www.amazon.com]

This leads me to a question about motors and choosing a power supply.
I was told to calculate total power needed you should multiply the amp x watts and then add 50% for peak usage. sooooo 24W x 2.8 = 67.2 so lets call it 68 then add half of that for a total of 102.
with 5 total steppers 2 for Y 1 X 1Z and 1 Extruder equals a total of 510W needed?

The rest of the system such as fans and the duet board will take v small amounts of power to the tune of 15 more watts maybe.

The above math and spec is the furthest thing from my area of expertise so IF i have the entirety of the power calculations and amperage above wrong please correct me and let me know where I went wrong.
Electricity is not my strongest suite.

What else am I missing?

Cheers

There is a Duet3D wiki page on calculating total power needed from the PSU at [duet3d.com]. The stepper motors take much less power than you have calculated.

Edited 2 time(s). Last edit at 02/11/2018 03:31PM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Related question.
If I slice files in Simplfy3d will the Reprap software on the duet read and execute it properly?

Quote
chromaglow
Related question.
If I slice files in Simplfy3d will the Reprap software on the duet read and execute it properly?

Yes, lots of users including me use S3D to slice models for Duet powered printers. S3D does have some limitations, in particular it doesn't support firmware retraction, and its support for multiple dissimilar printers is abysmal.

Cura has come on a lot recently.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

What do you mean by "multiple dissimilar printers" ?

I mean printers that have different bed sizes and shapes. Other slicers make it simple for me to configure several printers and then choose which one I want to slice a model for.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

One thing that will need careful consideration is the 30 day print time\pause and resume feature. For sure you are likely to need to pause the print to fit new rolls of filament unless you can do it on the fly (I do but it can be hit and miss). It would probably be wise to fit a filament monitor which can be configured to initiate a pause (Duet can do this).

There are a couple of caveats in the event of a pausing due to a power outage though. Firstly on resuming, it is necessary to re-home the printer. For Cartesian or CoreXY homing Z with a print still on the bed can be problematic unless you home at X max rather than the more usual Z min. Secondly, and more importantly, if power is lost and you rely on a heated bed (which most people do) then once the bed cools, the part will become detached and the print will be lost. So maybe a cold bed and something like blue tape may be the way to go? Not ideal though. Alternatively some sort of serious UPS or even backup generator might be needed. I guess it's a case of weighing up the likelihood of a sustained power outage against the cost, in both time and money, of starting again 29 days into a 30 day print.

---

[somei3deas.wordpress.com]
[www.youtube.com]

Something else has occurred to me. If I was building such a printer, I'd using a Diamond 5 colour hot end (5 in, 1 out) but not for it's multi colour capabilities. I'd load the same filament into each input (which is something I also do now). Coupled with 5 off E3D Titan extruders. This means that you can load 5 times the filament so will reduce the need to pause, change filament, then resume by a factor of 5. I'd also buy bulk reels of filament. If you used 2kg reels which are fairly commonly available, you could then have 10 kg of filament loaded and in use.

What this also gives you is 5 melt chambers which means that you can melt filament very much faster than even an E3D volcano. This may be of benefit if the parts you intend to print can stand the coarser resolution that a (say) 1mm or 1.2mm nozzle will give you. By employing 5 melt chambers there is a very much larger hot surface area coupled with the fact that with the mixing ratio set to 20% for each input, each filament moves at 1/5 the speed of a single hot end. So not only can you use larger layer heights and widths, but because you have 5 melt chambers, you won't have to slow the speed down anything like as much as you would have to with single input hot end.

See the blog in my sig for various posts I've done with using a Diamond at high speed and printing largish parts such as a coffee table.

---

[somei3deas.wordpress.com]
[www.youtube.com]