@oliof here are some pictures and videos on google drive. Google you probably have 
https://drive.google.com/drive/folders/17GNqAortkIXxc3fdv3P3tbFUaRGGbXFE
and this is somewhat more ambitious build, quad markforged mockup XY plane, but awesome life happened to me so making it proper and metal is on hold:
https://www.youtube.com/watch?v=dbukhT5yJ8k
@dc42 This is an update to an old topic. I have waited very, very long to get the multiplexer chips on breakout boards. So, yeah, a delay.
But, just want to report that i have success with multiplexing only enable pin. External drivers are daisychained for step and dir and En is muxed.
Mux is controlled with 100-103 GPIOs.
And it works! Such a cool little thing this mux modules.
Hi all.
I have rewired everything. That was it. I had 105% belief in Duet that it’s not faulty. I do hope it fixed the issue for good and i think i know exact reason: me not puting everithing on a mount and due to using breakout boards not being fixed and cables constantly moved all over during “makeover”, it left those little hairs doing nasty business to thermistor cable (and maybe other places).
Everything looked ok for a freshly assembled proto-mod of chinese base machine, yet it is/was everything but a nice and tidy thing. The box is printing
As i redid the crimping, & replaced AC power switch, all is fine. I didn’t suspect the breakup board of vga cables to printheads early on..
How does it come to a state, when you can make testing, heating, moving etc, but it fails to operate on gcode?
And coincidentaly, couldnt it be a sort of check-up test?
i edited/deleted because i thought you won't get it 
i'll call my coding brain friend and will share if anything comes out of this.
Thank you for a very constructive debate.
@Phaedrux as promissed, the pictures of a build in progress
https://ibb.co/9qDPTWn
https://ibb.co/CsGbTPb
https://ibb.co/6b5DQxL
It got pretty compact.
Needs a lot of work still to get the flat round cabling routed to both sides of X gantries from the bottom, then flat cables from there to each head, make breakout boards on toolheads for ribbon cables etc.
But all in all. There is progress
RESOLVED
For anyone, that might experience such headache and stumbles upon this:
one maps the axes, not the motors to the tools. and you count the axes, not specify motors...
@droftarts i did not understand what you meant that first time. Now i get it.
Thank you.
@dc42 & @Phaedrux : issue has been resolved. One maps the axes not the motors. I get it now.
@gringo you set the tool 0, tool 1 and tool 2. Tool 2 is having components from tool 0 and 1.
You set the origin to bed center.
And you only set one piece in your slicer. The machine will print it with both heads.
Should you print mirrored, you’ll just control that in the slicer.
What kind of printer are you using? If it’s i3 design, and running Duet, i have such and have a repository to share for you to build on. I use Tenlog D3 base, design evolved but IDEX stuff is still there in shared map.
There also is an S3d profile..
Anyhow, if people here don’t mind, repo si here:
https://drive.google.com/drive/folders/14S6cTYI0-bUS4PPHw44QKlxj1BM217J-
And facebook group with a bunch of people is here:
https://www.facebook.com/groups/237163797566846/
@wdenker i don’t think that is how it works... unless you want to do parallel printing, but that adds a whole new level of complexity to the slicer side, daisychaining control boards via SBC, probably syncing Z, or go with Autodesk project Escher kind of way... complicated stuff..
As current setups go, you do mirror or duplication, and while it looks individual, it really isn’t. Without a user error, they will not collide.
If you’d like to just add redundancy and eliminate the potential user slicing error, you could put an extra endstop (preferently optical) on one of the gantries and before they collide, they stop.
@kru96 when selecting T2 you are selecting ditto tool (t0 is left probably, t1 is right and T2 is both...)
They move each on aproximate center of their bedhalves because in Tool definition you are setting the offsets and that positions them accordingly.
The bed center is in the middle of the bed for ditto/mirror mode to function.
@oliof thanks .
Wont lie: A lot of thinking went into concepting and construction on both of those.
@oliof here are some pictures and videos on google drive. Google you probably have
https://drive.google.com/drive/folders/17GNqAortkIXxc3fdv3P3tbFUaRGGbXFE
and this is somewhat more ambitious build, quad markforged mockup XY plane, but awesome life happened to me so making it proper and metal is on hold:
https://www.youtube.com/watch?v=dbukhT5yJ8k
Hey @Pyper
I guess this is too late of a response but i will put my 5 cents in anyways regarding this topic.
I made my daily workhorse merging 2 cheap IDEX machines. Just 2 X-gantries on a bedslinger, one on the front and one on the back, shared Z but 2 independant motors driving it to assure gantry level. X gantries both have linear rails. When built right (parallel linear rails!), it’s easy to use. Tool 0 has a fixed mount while other 3 have a height adjustable mount i came up with.
There are many that say this is crazy and unreliable. I made it and it works. Probably people that dis this approach should first build a machine like that
My main goal was to have one machine for big output of mostly jigs and fixtures, adapters, etc for what FDM is great.
While it does work for multimaterial, it’s not my gig. If at all, i use secondary supports material.
So yeah: Very doable & very price performing results.
I am not that active here last months so joining a “Marx group” on facebook ( https://www.facebook.com/groups/237163797566846/ ) will give you access to my repository and a more direct help if needed.
@aitor very sorry to jumping in here… but…a 10 individual head machine? 10 ?! what kind of hellfire is this?
I am walking this path for a while now - an IQEX… in process of building a new one, based on what i’ve learned in past 2 years with it.
To start I used cheap i3 style IDEX (Tenlog D3 was my choice), threw out crappy boards and did height adjustable toolmounts (very important). Used Duet2 + Duex5. Then just doubled the IDEX axis and viola, got IQEX… at this point, i use one as IDEX and one as IQEX. Daily.
So this is totally doable and huge advantage when you want high output.
What you want is take time and assemble the thing VERY precisely. Everything square and BOTH X rails EXACTLY parallel and square to Y and Z. Not a huge challenge, but I for example didn’t think of this when building first time.. fixed that, learned something. Making a jig on which you assemble the X gantry so that both X linear rails sit flat and then you torque everything up helps a bunch.
Whoever relies on meshbed levelling is doing stuff wrong. Yeah it’s nice, but you tram the machine and level the damn flat bed. Then you print on it with absolutely no problems day in and day out.
Disadvantage: with i3 style, Y axis is only one so while you can do multiplication, you can only mirror in X direction. Not huge disadvantage & still has better output at the end of the day than any “i printed a benchy in 10 minutes” gimmick. Machine is not realistically encloseable tho.
Works absolutely fantastic for about 1000 bucks printer.
What i am doing now is a box design - so XY plane is moving and bed is on the Z. It uses (well it will eventually) 2 stacked double markforged kinematics. Something similar to Muldex printer, but doubled.
Is it worth it? Not really sure. Should be fast as hell compared to my current i3 IQEX. More silent for sure. Encloseable. Insane output. Etc. Still, if concept works, then i will have to scale the design up to like at least 600x600 mm buildplate so this monsters electronics and build price are justified.
If this triggers your heart, i do posts from time to time on my FB group (“Marx group”. As in Marks’ group, not in marxistic way )
@dc42 DOLD mechatronics sent another 3HC board. It now connected and updated instantly. Under a minute job after i got replacement. 
Now to figure out this M669 for quad markforged kinematics matrix stuff.. not sure if excited or scared down to the bones. 
Maybe see ya in another thread I think i'll make something like this for my signature when 1s, -1s and 0s are in place 
M669 K1 X1:1:0:0:0:0:0:0:0:0:0 Y1:-1:0:-1:0:0:0:0:0:0:0 Z0:0:1:0:0:0:0:0:0:0:0 U0:0:0:1:0:0:0:0:0:0:0 V0:0:0:1:0:0:0:0:0:0:0 W0:0:0:1:0:0:0:0:0:0:0 A0:0:0:1:0:0:0:0:0:0:0
Thanks for warranty. 
@dc42 thank you. I've let it sit there on 12 volts for 2 hours. then 2 hours on 10 volts. it was 2 hours on 24 volts before so... sadly, no luck with booting. The other one with a newest bootloader is ticking like expected. Thanks again.
@dc42 i ask for warranty approval.
@ja cabels are well made. Swapped them. No change. Reordered and also used only that one board. No change. Shuffled addresses. No change.
Since DIAG led is not lighting up even a bit, yes, i would have to load a bootloader. That can be done via CAN bus, or, if for some strange reason you get your bootloader *ucked up out of the box, theoretically with Atmel ICE. Which is a 125 eur device that you then use one time on a 125 eur board.
But yeah. Agree 100%
@jay_s_uk run it on 24 but i can try 12v... i could revert, but the DIAG led should at least do something whatever firmware it has. or not? funny thing is, the board that works has serial ending 71, and the one that does not has 69.
Anyways, not doing this because it should work. does not. I'll RMA it. This is why i bought original not cloned.