So after my frustrations last weekend with part 1, I ordered some new bolts with allen key heads, and some nylon insert nuts. The idea being that I'd get more torque on my tightening, higher locking force and better resistance to coming loose.Ordering from screwfix was slightly frustrating insofar as the postage was 1/3rd the cost. But for 50 16mm and 50 20mm bolts + 100 nuts, it was just another £15 to the cost. Pushing me up to £85 so far. Though I may be able to take some of the b&q stuff back, but I shall count it against me ,if only as a reminder to use screwfix in the future.
This time round I clamped a length of the square section vertically in my bench vice. Then arranged the metal plates at 45 degrees and clamped that in place, so that I could position the bearings and get a good tight lock on everything.
This approach worked pretty well. I still had to loosen and adjust again when actually between the rails, but the tighter starting position, and additional stiffness of the connections seemed to make life a little easier. I used some blocks between the front wood board (the one screwed to front facing brackets) and the metal plates. This helped to keep things even and in place as I adjusted the bearings.
Another tactic that worked well was to wrap rubber bands around to hold the bearings against the bars. Here you can see I've used screwdrivers to hold the bands in place whilst I tighten.
The result is I now have a z and y axis with zero rattle in them as they slide. Whether they'll stand up to vibration and use without slipping is something only time and use will tell.
Having got that in place I moved on to the x-axis. This has to bear the weight of everything else, whilst still smoothly sliding along rails. I drew a few designs, and toyed with the idea of suspended table versus not. Basically the question is where do you attach the drive that will push the axis forward/backward. The obvious thing is to run a rod central to the axis, as with the others. This means connecting the two sides via something with a threaded hole in the middle. But that means suspending the work table above that. The alternate is that you provide driving force at one, or both ends, so the only connection is where the Y axis runs between each side. This feels like it might cause a twisting torque (in the case of pushing at one side) or be more complicated (trying to drive two sides at the same time) however I felt that keeping the height down, working against a simple flat work surface, was better than the complexity of building a suspended work surface that is flat and stiff enough not to cause problems later.
Having decided this, measured the available travel in my z-axis (100mm) and allowing for some clearance, I calculated how tall to make the x-axis 'legs'.
I chose a width of 200mm, nothing terribly scientific, I just felt that any more and I'd lose working area, and less risked an unstable platform. I decided to make up 4 metal plates, each 100mm x 70mm. Each bracket would get 3 bearings (because I only bought 30, and after z/y axis have 14 left, so this uses 12) The bearings are arranged as 2 above the bar, and 1 central to those, but beneath the bar.
To give best stability each leg of the x-axis will have 2 of these plates, one mounted to run along each side of the rail. This also means I could stretch the base of the legs wider ,or have them as narrow as 100mm if I felt like experimenting.
The square bar is 25mm wide, and the metal plates are 4mm thick. So I need to create a structure to hold 2 mdf panels 37mm appart the extra 3mm is to stop the metal plates being pushed against the bar, it just wants to touch on the bearings.
In 4 hours I made 2 plates, 2 sides to one leg, and assembled the whole thing. So I currently have just one x-axis leg. However I am pleased with it. There is vertical play in the motion, the bar runs smoothly through the bearings and the whole thing feels like it will bear the weight of the other axis without binding up or rattling around.
Hopefully one more weekend and I can make the second x-axis leg & assemble all three axis together. Then I'll need to try mounting a threaded rod at one side and see if it will work, or if I need a more complex arrangement to push at both sides simultaneously.
After that it will be probably another weekend to make up a router mount on the x-axis. Before I start to think about buying motors etc.