Been working on the design again.  I'm going to use twin Bowden feeds to hotend with nozzles for 3mm and 1.75mm filament.  These will be fed from two E3D Titan extruders mounted on the XY frame (Z carriage).  To equalise tensions in the support cords the extruders with their stepper motors will go on the opposite side from the XY motors.  Now since I shall want to get to the extruders easily they will need to be at the front - so the XY motors will have to go at the back.  I can't see this being a problem.  I can simply attach the frame the other way round with the cords crossed over across the front (unless I can arrange both the drive motors and pulleys at the back).

Another idea (maybe more usual) would be to have the extruders at the top of the printer near the middle.  That would remove the weight of extruder motors from the Z carriage and also make the Bowden run straighter.  Maybe I could rearrange the XY motors and cord runs to balance the carriage better.  The water cooling pump and reservoir will also probably go on the top so with all that going at the top I think I shall make use of a piece of plywood.

It sounds sensible to me to remove as much weight as possible from the moving bits

Absolutely Dave   I would like to remove the XY motors from the carriage but I think that would be too complicated - that's if it's possible.

I can balance the Z carriage by moving one of the XY drive motors.  Two diagrams - first showing standard CoreXY layout with drive motors at the front and second showing one motor moved to the back to balance the carriage.

Now seeing what wants to go on the X carriage.  A lot less since getting rid of extruders and replacing with Bowden feed.  Heater block will be bigger to accommodate two 24v 120W cartridge heaters.  Then there's the Z probe and the part cooler fan and duct.  I'm now about to print an increased size part cooler air duct.

The increased size might be enough (120%) but I'm now printing an even bigger one at 140%.  The 120% one will go on my Titan printer.

What size air duct I need will depend on whether I go for a single or double hotend.  Haven't quite decided yet.  I guess a single would be easier to start with.  Also, not sure whether to go for a dual hotend or interchangeable hotends.

I've decided to go for interchangeable hotends rather than a double nozzle jobbie.  This will provide for upgrades/experiments without a major rebuild of the X carriage every time.

Here is a diagram of the heatsink and water blocks.

Now with heatblock and nozzle added.  Bowden coupler, heater cartridges and clamping screws plus water cooling tubing omitted for simplicity.

A clip will hold the heatsink against the water block with thermal grease between to aid conduction.  The clip will attach to the X carriage.  To change hotends only the clip will need undoing.  Then the heatsink, heatbreak, heat block and nozzle assembly is simply swapped for the alternative assembly.

Been trying to make the heatsink block only to break an M6 tap   I'll try it another day...

Designing the X Carriage around the hotend.  A bar across the front screwed onto the X carriage main part will hold the hotend system in place and the heatsink against the water block.  The round part to the left of the hotend will be glued to the air duct for part cooling.

Screenshot of model for first test print of X carriage.

The heater block is being produced by another member for me so as an interim I shall print this plus the heatsink (which I haven't made yet) as a model in plastic to check how everything fits round it.

That was very instructive   The air duct for part cooling needs to be 12mm lower and 8mm further forward than I first though.  I find that however well you try to model things in CAD, there's nothing like real 3D parts

More photos.

Moved the air duct.

X carriage mounted on X rail.  The limit on printing size due to range in X and Y directions plus probe offset seems to be 460mm x 460mm on the 500mm square aluminium bed.  With the heater pad being 400mm square, this seems adequate.  Printing area of 18" square in Imperial measure.

Having a moving XY carriage is so awkward that I decided to compare the weight of this and the print bed.  Frame complete with X carriage and rail = 2.9Kg.  Print bed aluminium panel with heater = 4Kg so over a Kg heavier.  However, breaking strain of cord = 80lb = 36.3Kg so four of them could carry 145Kg.  So a 4Kg print bed plus a bit extra for the wheels and brackets is hardly heavy in comparison.  Even adding a couple of Kg for a huge print it would still not be significant.

Another point - the most accuracy wanted in Z is when the printing level is low.  Using a moving print bed the amount of cord carrying the weight is at a minimum at the beginning of a print - cord is unwound off the horizontal bar to lower the bed.  OTOH with the XY frame moving the cord is fully unwound when the printing starts.  OK the bed levelling corrects for height errors but why make things more difficult than needed?

A moving print bed would mean that all motors are on the fixed frame.

Having looked into how to arrange a moving bed I've found it's not as simple as I thought and would involve quite a lot of extra work so I've decided to stick with a moving XY frame and leave the bed stationary on the floor.

Designed and printed new Y carriages with three wheels - two on top and one underneath.  Plus the LHS Y carriage has been designed to take the X endstop.

Now I know the amount of XY movement, I can calculate the amount of wrap wanted on each of the CoreXY drums.  Starting off with 20mm diameter drums and a print area of 460mm square, the maximum movement of either A or B cord is twice 460mm = 920mm (eg. from X=0, Y=0 to X=460, Y =460mm, for the X+Y drum).  The circumference of a 20mm drum is 20 x Pi = 62.83mm which gives 920/62.83 revolutions = 14.64.  Cord is 0.4mm in diameter so 14.64 turns would occupy 14.64x0.4 = 5.9mm.   That's fine.

NOTE - Above edited due to silly mistake.

The power tapping taps and drills have arrived this morning so I shall be having another go at tapping the M6 hole through the aluminium heatsink block.  They look very good, as usual from Axminster Tools & Machinery.  Everything I've bought from them has been first class.