Building Your Own Printer; How Difficult?

[JoeKitchen]

So the largest commercially available printer has a print width of 5 meters.  How difficult would it be to custom design and make one that is 8 or 10 meters? 

The printer would have to be designed to print on uncoated fabric (such as a poly-cotton blend) that has been treated to be fire retardant.  So not only would the print head and inks need to work on this material, but also not pull the fire retardant salts to the surface as the ink is drying. 

I know it can be done; a company in Germany built a 12 meter printer.  Just wondering if any knows what kind of issues you need to solve and how crazy this would be. 

[Lessbones]

do you have any info on that company in germany?  i'd love to see what they did... also they'd probably be the ones to ask tbh.  Your biggest issue at that size is going to be rigidity-- do they make linear rails that long?  I'd guess someone probably does for really heavy industry, but you'd also be looking at significant potential for torsion at that size.  This being a photography forum, I don't think you're gonna get a ton of help-- If I were going to attempt something like this I'd probably start by looking into and talking to custom CNC builders (most of the companies in china that build CNCs will make them in just about any size).  That would get you closer to the important parts of the chassis construction, then you can look into controller boards like BYHX and RETC that can drive printheads and feed mechanisms.  I just hope you've got a couple extra million $$ burning a hole in your pocket...

also out of curiosity, why does this need to be done at that size?  Are there really rolls of this material that come 8-10 meters wide?

TL;DR it's quite crazy, but it would be doable with enough money and resilience...

*Edit:  answering my own question... wow:  https://de.wikipedia.org/wiki/Infinitus

[JCD]

There is a reason Big Roland’s and Mimaki use Epson heads. It’s not so easy, much less keeping software up to date.


AI Overview
Epson makes the print heads, such as the DX4 and DX7, that are used in many Roland and Mimaki printers. These are OEM (Original Equipment Manufacturer) print heads that are sourced from Epson by the printer companies, although they may be sold under a Roland or Mimaki-specific package.






quote author=Lessbones link=topic=145051.msg1267022#msg1267022 date=1757628941]
do you have any info on that company in germany?  i'd love to see what they did... also they'd probably be the ones to ask tbh.  Your biggest issue at that size is going to be rigidity-- do they make linear rails that long?  I'd guess someone probably does for really heavy industry, but you'd also be looking at significant potential for torsion at that size.  This being a photography forum, I don't think you're gonna get a ton of help-- If I were going to attempt something like this I'd probably start by looking into and talking to custom CNC builders (most of the companies in china that build CNCs will make them in just about any size).  That would get you closer to the important parts of the chassis construction, then you can look into controller boards like BYHX and RETC that can drive printheads and feed mechanisms.  I just hope you've got a couple extra million $$ burning a hole in your pocket...

also out of curiosity, why does this need to be done at that size?  Are there really rolls of this material that come 8-10 meters wide?

TL;DR it's quite crazy, but it would be doable with enough money and resilience...

*Edit:  answering my own question... wow:  https://de.wikipedia.org/wiki/Infinitus
[/quote]

[Lessbones]

There is a reason Big Roland’s and Mimaki use Epson heads. It’s not so easy, much less keeping software up to date.

this was true up until recently, there's a much more varied landscape these days-- you'll find lots of ricoh, kyocera, KM, Brother, and other heads in those printers these days.  Most of the chinese manufacturers even let you choose.  But I'm not really sure what your point is anyway--  imo driving the heads is kind of the "easy part".  I can't tell what they're using for the Infinitus, but they look like single-color heads that probably use very large dots, maybe old KM heads.  But the chassis design is really gonna be the challenge.  The Infinitus seems to use a massive hollow roller transport or something like that, it's a bit hard to tell, they're obviously not trying to share TOO much info.

[JoeKitchen]

do you have any info on that company in germany?  i'd love to see what they did... also they'd probably be the ones to ask tbh.  Your biggest issue at that size is going to be rigidity-- do they make linear rails that long?  I'd guess someone probably does for really heavy industry, but you'd also be looking at significant potential for torsion at that size.  This being a photography forum, I don't think you're gonna get a ton of help-- If I were going to attempt something like this I'd probably start by looking into and talking to custom CNC builders (most of the companies in china that build CNCs will make them in just about any size).  That would get you closer to the important parts of the chassis construction, then you can look into controller boards like BYHX and RETC that can drive printheads and feed mechanisms.  I just hope you've got a couple extra million $$ burning a hole in your pocket...

also out of curiosity, why does this need to be done at that size?  Are there really rolls of this material that come 8-10 meters wide?

TL;DR it's quite crazy, but it would be doable with enough money and resilience...

*Edit:  answering my own question... wow:  https://de.wikipedia.org/wiki/Infinitus

Thanks for the info.

The company in Germany would be a no go.  They are a competitor and currently the only one in the world that can produce seamless backings over 16 feet tall.  They would not talk to us, especially after finding out we are a competitor. 

We toured a stage this week with two sets.  One had our backing and the other from the German company.  Ours was 18 x 75 and the other 25 x 360.  We came away thinking there must be a way to figure this out. 

Most backings fall into the 18 to 25 foot tall range, so we need to seam almost everything.  The seams are the Achilles heel to a backing since they are done by hand and need to be perfect every time.  Eliminating the seams would be ideal.  We constantly get asked by DPs about our seams, and some refuse to use any backings with seams. 

Yes, we can get fabric up to 12 meters in seamless width, although I think that is over kill.  Yesterday I received fabric samples for four different fabrics we can get up to 10 meters.  They are raw and uncoated, and would more then likely come in loose rolls or bales, which adds to the problem. 

Yes, it is a huge undertaking building a printer that big.  But, we are one of only five companies in all of N America who can do this work, and being the only one doing seamless up to 32 feet with full production in the USA (the largest film/TV market in the world) would be monumental for business. 

My business partner is also a metal fabricator, so I do not think the construction would be "too difficult."  You would need interesting design solutions for something that wide.  My main issue right now is the print heads, getting custom inks (already found a company for that), the software, and whether or not we need to coat the fabric.   

[Lessbones]

the drive hardware isn't that crazy-- look into those companies i was talking about BHYX and RETC systems can be found pretty easily on alibaba for not very expensive-- I have a printer that uses RETC, it's not a very good system, and the biggest issue you'll face is having proper waveforms for whatever type of ink you're trying to jet.  But in terms of constructing the actual movement portion, both systems can use a variety of linear optical scales (raster strips).  Mine has a 180dpi encoder strip that is measured by infrared LEDs in quadrature, so it ultimately has a true resolution of 720dpi.  I was thinking about upgrading this to one that is 360dpi, and it's easy enough to swap them out, then you can have the software simply measure how many ticks it is seeing for a given distance (reported by the servo motors i believe) in order to calculate the resolution of your positioning sensing system.

The main boards for these systems come separate from the print head driver boards, so you'll have a choice on what to use for that.  I have ricoh gen6 heads that are (unfortunately) being driven by ricoh gen 5 drivers, but you can also choose kyocera drivers, epson drivers, and so on.  A lot of integrators of cheap machines use the Epson XP600 heads because they are very cheap (~$300) and as such can be treated as a true consumable.  That's kind of the direction I'd think you should start with.  All of these systems also provide utilities that you can use to define multiple rows of heads etc. so you can print a bigger swath at a time.  I can show you more of what I'm talking about for my system (the RETC aka Royal Shanghai) but I would suggest that you look into BHYX as they seem to be less hack-y (ive heard that a lot of the code is based on stolen stuff from other systems) and you'd ideally have access to the waveform editor.

[JoeKitchen]

Thank you for all of the information.  There appears to be a printer's expo in Orlando end of October I think I will be going to as well to see what is available. 

Anyway, so after a week of research, I have come to the conclusion that designing and building your own printer is feasible, just expensive. 

The really complicated part is deciding on which fabric and ink to use. 

Not going to lie, jumping from paper to fabric really makes things complicated. 

[narikin]

There is a company in Madrid: 'Factum Arte' (remarkable fine art company) who had their Epson fine art pigment wide format taken apart and rebuilt by some tech nerd guy from Wales, UK (iirc) who completely rebuilt it to work as a flatbed, with computerised alignment on a massive print table. a bit like Swiss Q or Epson's table version (but before either of those existed)

It's not what you want, but clearly there are people out there who take the gubbins of a machine apart and rebuild/ reprogram them to work how they want - manufacturer be dammed!

Factum print their XLs on smooth gesso'd fine belgian linen for ultimate resolution and flexibility.

what is your current biggest seamless, and at what resolution/ inkset?

[JoeKitchen]

Our biggest seamless is 5 meters (16 feet) and we are using an EFI UV printer. 

For an UV printer, you need fabric that has been coated to receive UV ink otherwise you will run into adhesion issues.  And since all commercially available printers max out at 5 meters, so do the fabrics.  Unfortunately you can not coat fabric with UV coatings outside of a mill with very specialized equipment that keeps track of particle size and solid to water ratios during the application and curing process, or at least according to someone in the coatings manufacturing industry I spoke to this week. 

So, UV printing is out due to the unavailability of appropriate fabrics. 

The only real combination that would work, it seams, are fast drying water based inks printed on raw bleached cotton and then fire treating the print after heat curing.  This would also create a highly flexible backings, which is a real plus too. 

Oh yeah, the other thing to consider, they need to have day|night (aka front|back lighting) capabilities.  You get this by either using an image|white|image printing method (like with UV printers) or double sided printing (like with DTG printing). 

Anyway, I keep coming back to the same starting point, getting or creating the right water based inks first.