VERY interesting Andrew! I think we'll all be curious to see the edits in Marlin. Very innovative use of the Core concept. You not only thought outside the box, you got Core outside of a box too! Good use the new C-Beam too. What plate are you using for the z axis wheels? It doesn't seem to be an Open Builds plate. What advantage does it have over the standard OB Mini V Wheel Plate? When designing belt driven z axis I've had to set Marlin to keep the z motor(s) powered when not in use to keep the x axis from falling until power is turned off. Maybe in this core design that wont be a problem.
Well I gave up on Marlin after getting it somewhat working (with some help from @bearwood), the stepper.cpp file needs some re-work to get limit switches working with on a CoreXZ setup. The motion works fine though. Good new is that the Repetier firmware has support for CoreXZ and works great. The plate for the z-axiz is the OpenBuilds 20mm gantry plate. Haven't seen it used in many builds but it is half the cost of the full size gantry plate. Its a little cheaper than the mini-v plate and I can use the same wheels throughout the build. I reduced the cost of the wheels by using 625zz bearing for a lower cost over the 625-2RS that come with the wheel kits (not worried about a high dust environment). On ebay you can get 50 bearings for about $18, which puts it at about 1/3 of the cost per bearing. With the CoreXZ design I have no problem with the gantry moving without power, must have to with the two belts working opposite of each other. I was worried about it a bit but after getting the guts together tonight I can say with certainty that it is not a problem as long as you don't have too much weight on the gantry. I placed a Wade's reload extruder with a nema 17 on the gantry and it didn't move until I placed a small piece of extrusion on it. I was going to use the Wade's direct to hotend but rethought that about a week ago due to concerns on gantry weight and am going with a Bowden setup, so I should be fine. Here's my progress so far, decided I couldn't wait for my missing F625ZZ to use as pulleys and ended up getting some fender washers (3/16 x 7/8") for $0.17/each that fit perfectly on M5s and used some regular 625zz ball bearings I had let over from making the wheels sandwiching 25mm m5 screw, fender washer, m5 washer, 2 bearings, m5 washer, fender washer, 6mm spacer then drop in tee nut. Works great, but I will probably switch to the F625ZZ once I get them since I'll have them. Anyways here is some pictures of my progress. Also you'll notice in the pictures I used 20X60 vrail on the uprights, but I think 20X40 would work fine, thats whats in the sketchup file.
Huh, somehow I missed that there is a 20mm Gantry Plate. Cool! F625ZZs from Robotdigg take forever (the proverbial "slow boat from China"). I got some F695ZZs for about the same price out of California in 5 days. Two F695ZZs have only 6mm between flanges so things really need to line up, but they are only 13mm flange dia instead of 16mm and also have 5mm holes. I hadn't even thought about endstops being a problem, but - YEAH! Good to know Repetier supports CoreXZ. I wonder if that's one of those things where the programmer just went, "Well why wouldn't I add it?" or if they actually had a Core XZ design needing it. A good test for this would be to take a skp of the full size gantry plate, raise it to 100mm high, set the slicer to Z lift .5 on non-print moves, and run the print with no filament. At 100mm/20% infill it takes about 24 hrs to run. With that plate's 71 holes to work around it will give your z axis the same workout as hundreds of hours of normal printing. Be curious to see how level the extruder is with the 4 corners of the bed afterwards. In fact, leaving the extruder off but hooked up so it runs, and running the print with a dial indicator mounted on the x axis plate is how I do that test (I have a totally different view of so-called "auto-bed-leveling" after doing that test). Keep us up to date Andrew. This is a pretty radical concept (radically cheap too). Hope it all works out!
Can you show a few more pics from the other side? It looks like each motor is only pulling in one direction, lol.