In the last video I tested whether I could butt together MGN12 liner rails to create larger expanses along the axes of a CNC machine, and I also tested some methods of improving the rails performances. I eventually decided to go with longer rails, and I also realised simply cleaning the rail blocks in alcohol significantly improved their movement. In this video I will reveal the latest plates designs, and talk about where I’m at, in this ongoing CNC-building saga.

When I began designing my CNC machine, I decided to learn how to use a CAD software to help me work out positioning for things which I would otherwise find difficult. You can see the design here. It is incomplete, but I’ve done enough to work out all the part for a completed machine. I’m keeping some things from the previous version such as the pivot mechanisms on some of the plates, which make tramming the spindle a piece of cake, and I’ve removed some things such as the proximity sensor mounts – which I plan to replace with limit switches which will be positioned on moving plates instead of the receiving ones.

So this should be 140mm.

Eventually got some plates water-jet cut in 4mm 6082T aluminium but to get there I had to do a lot of tests. I made some plates out of 12mm MDF, and I even 3D Printed some. Both of which are valid methods of prototyping. One I could do during the day while I was at the workshop. And the other I could do into the evening, knowing the noise wouldn’t inconvenience anyone.

I’m going to jump straight to the aluminium plates. Unfortunately I don’t have any footage of this being done as I dropped the material off one morning before work, and come back a week later to collect.

It was done on a water jet cutter in a place local to me. They sorta blasted through this like you wouldn’t believe it. So there is a little bit of a taper on all the cuts, and the bearing doesn’t fit all the way in yet so I’m waiting for a reamer to come in the post to open that up. I also make these mounting holes a bit smaller so I’m going to have to drill them out and tap them. I feel quite confident that this will work. Here is this part – you can see the holes minus the recesses, so if I wanted to recess I could get some special drill bits for doing that but I don’t think I need too – I think I’ve given myself enough space for everything. The only problem is they made one mistake. This part here. I don’t know what went wrong there because everything else cut out ok so I’m not sure what he was doing to that plate. Really annoying. I think I can get around this so either I’m going to counter sink the hole or make a bushing to go in there, just to keep the machine screw centred. In anycase they’re not going to get a plug from me.

The mistake was really annoying, but I’m going to ignore that for now and check the one thing I didn’t test which was the intermediate plate. It was a bit of a gamble to decided to cut all the plates out without having tested every single aspect but I felt quite confident that the 3D model I designed would give me all the information I needed to continue.

So to save the need for countersinking or recessing the heads of the machine screws, with the latest design I’ve made this spacer plate and the idea is that this fits between the plate that runs along the c-beam and the pivot plate and the machine screws from either side, should fit in this space. And it should allow for enough room for this to pivot.

The shapes cut out are an accumulation of all those heads, and any possible movement created by the pivot –

The inclusion of this plate, and the use of flanged bearings means all the the parts can be very easily and quickly cut whether on a laser, water jet or eventually a CNC.

What I need to do is drill out these holes here to 4.2mm and then tape them.  I mark the holes I need to drill out with a permanent marker, and use some cobalt stub drill bits to get them to the required size. If I want more accuracy I could use a reamer – and I will later use one for the bearing openings but for everything else – it’s simple drill bits.

I tap the 6 holes to M5 by hand. I later bought a 40 degree helix spiral taps – from a place which even listed the newton meters of force the bit could withstand.  This bit can be used in a drill and make tapping a lot quicker and easier.

Ahh I’ve stuck it the wrong way around.

I assembled this and luckily I doesn’t seem that I’ve overlooked anything. It’s moving correctly and feel quite solid. I decided to mark up all the other M5 holes that needed drilling and tapping.

I now need to address the issue of parallelism of the rails. Ok at the moment this slides really well because I’ve only screwed the bottom rail down to the plate. The problem I have is when I screw the top down, it because difficult to slide along. It’s not impossible and I think the leadscrew could move that but it’s not ideal. Really when you look at professional machines they often have one rail on either side of the machine and on the gantry one is at the front and the other is at the top. And the reason that is, is even if there’s a little bit of deviation with the two rails not being parallel because the movement is in line with the face of the plate plate with move backwards and forward. When all the the machine screws are tightened the play is along the thin edge from top to bottom of the plate, and essentially the plate is being pulled and pushed together. I tried to compensate for this by using this jig to line the two rails or guide blocks up and I got it to about 0.02mm which I think is pretty good. But potentially the plate isn’t entirely flat – there’s inconsistencies with the parts and like I said when the top rail is tightened its a little bit harder to move along.

One way I could get around this is to obviously just redesign the face plate but I kinda wanted the gantry to have space along the top so I could have the drag chain suspended on there so I didn’t have to add additional brackets to hold this in any other place. We’ll see how well that works but at least for the Y-axis I can still use this method if I change how I secure the top rail.

What I’ve found is, that instead of using 8mm to screw the plates to the rail blocks, if I use 10mm machine screws and a washer which when screwed bottoms out and leaves an ever so slight gap which allows the machine screw to travel within the slot while the plate is moved along the axis. You can see the washer under the machine screw head still being able to move.

This solution doesn’t seem to interfere with the overall movement and should help prevent the plate from nodding once the weight of the spindle is in place – but in reality this design goes against some major convention in building CNC machines. Because the rails are on the same face I cannot make any adjustment to compensate for any contours in either how the rails were ground, or any created by the surface the rails are clamped too. If this is a really big problem I could move one rail, to either the top and bottom of the gantries aluminium profile, and make a connector piece with some angle section which the gantry plate screws onto and in turn fixed to the re-positioned rail blocks.

Ok coming back to the mistake on one of the plates, one way I can get around it is by countersinking the hole a little bit. And then using a taper on the head to self align the bolt so I think I’ll just carry on – I think think I have enough time to go get a plate re-cut.

I set my depth stop and counter sink the openings. To get the faces lining up just right, I use a bit of paper under the aluminium plate which lifts the plate up and allows the countersink to sink a further in.  I also drill  out the remaining openings to their correct sizes before tapping – these include all the M5 threads and some M2.5 threaded which for the limit switches.

And finally I use a 16mm spiral reamer to open up the bearing hole to the correct size – which happens to be 16mm. The flange bearing pops in and it held quite nicely.

I’m going to leave this video here – the plates are all more or less ready for the assembly – I just need to decided whether to Loctite the machine-screws along the opposite rails or move the rail to another face of the aluminium profile and use a connector bracket.

Thanks again for watching and all my patreons for all their support. I am completely sick of CNC machines but I’ll persist. We’ll get there in the end.