AVOR – a steel-framed medium-size router

[Neale]

Basic Structure

As described in the intro, I used 3mm wall thickness box section steel throughout for frame and gantry, with a mix of 50x50 and 100x50. The frame is pretty straightforward. I found that steel has one brilliant advantage over aluminium extrusion – it is remarkably forgiving. The combination of angle grinder and MIG welder (I’m not clever enough for or could justify TIG) means that building the frame is not difficult, and in particular those two tools will also deal with the odd mistake. Chop and reweld to your heart’s content. What I did realise, after the event, is that I should have spent a lot more time jigging and clamping and aligning before welding. Obvious to the more experienced out there, not so obvious at the time to a beginner with a welding torch. Close scrutiny of any of the welds will indicate my lack of ability, but fortunately it seems that a strong enough weld does not have to look pretty. In many cases, I welded as a series of “spots” as I had found that there was more tendency for things to move if I made continuous welds, and what I have done is probably good enough.

3mm is too thin to take a decent tapped hole for fittings, especially for things like M5 cap screws to hold down the Hiwin rails. I have used the same idea as other people, therefore, and fitted a steel strip inside the relevant box sections to provide added thickness for tapped holes. To keep tapped holes square, I used the old toolmakers’ trick of turning up a short mild steel cylinder, maybe 20mm long, ends faced square, with a hole up the middle a close but not binding fit on the tap. Hold guide to the work and turn tap with other hand. Very easy to make, very easy to use, does a good job. And use spiral-point taps for these kinds of through-holes – enormously better than conventional hand-taps for this kind of work. Mine were cheap from China but stood up well enough to this steel with a bit of tapping compound, even when I was turning them with a cordless drill. I “marked out” the hole positions by clamping the rails in place and then using a cordless drill with a drill bit that was a close fit in the rail’s mounting hole to make a small mark in the steel. You can then swap to the correct tapping size drill which will pick up the mark and drill a hole where you want and not where the drill point drifts to. Using M5 cap screws in the rails gives just a little bit of wiggle room for final alignment as long as you were fairly accurate to start with.

Note that the additional short vertical braces (see photograph in previous post) which take loads from the main rail supports to just above the diagonal braces are not shown on the original drawing. These were added as a result of a sophisticated deformation analysis. I clamped a DTI to the bed structure and measured the deflection when I put my weight on the middle of the (unsupported) X rail. From memory, I think the bend was around 0.25mm which seemed a bit too much. So I added the extra supports and the deflection was very much reduced. Again, I forget the numbers but I think it was around 0.05mm. Gut feel said that this was better, and good enough.

The gantry is also a welded steel structure. More 50x50x3 box. Why not aluminium? Well, I had the steel to hand and the tools to cut and weld it, and weight saving is not as important as you might think at first glance. The steel gantry is almost certainly heavier than an aluminium version but not by an enormous amount. Things like the two 20mm Ecocast slabs forming the Z moving and fixed platforms are pretty heavy, and the spindle and mount add a chunk of mass. There are also Y and Z motors, ballscrews, and rails. As mentioned in the intro, the rotational inertia of two 20mm ballscrews 1700mm long is quite significant as well, so a bit of weight-saving in the gantry structure doesn’t make a big percentage difference to the total inertia as seen by the X drive motors. I reckon that the total finished gantry weight with everything in place is about 40kg. Of that, the steel weighs about 18kg. If you just replaced the steel with identical section aluminium it would weigh about 6kg, but in practice the aluminium would need to be of much heavier section, so maybe around 12kg? In other words, aluminium would only save about 6kg out of 40kg.

A partial view of the gantry but it shows all the important bits. There are also a couple of short upright sections of 50x50 bracing top rail to lower rails not shown on the picture.



The gantry structure has Hiwin rails mounted on top and bottom. I considered using the front face of the rails but top and bottom gave more separation. I was also able to achieve the ideal of the ballscrew geometrically centred between the rails, by running it through the gantry uprights which might be more difficult with aluminium. All the relevant holes have sufficient clearance to allow tweaking on assembly to get the alignment right – I allowed something like +-1.5mm. You can see one ballscrew bearing at one end of the gantry and the clearance/mounting holes at the other.





Of all the things that are important with a home-built machine like this, the most important has to be, “If you can’t build it accurately, make it adjustable!” I knew that I was not going to be able to keep the “feet” of the gantry accurately aligned due to welding distortion (I should have done better – see comment about jigging and clamping) but the machine design allowed for that. The X Hiwin carriages bolt to a flat 12mm aluminium plate. The gantry feet - lengths of 50x50x3 - bolt into tapped holes in those plates via a set of holes just inside the ends of the feet. However, the feet were not truly horizontal, as expected. I therefore made up a thick epoxy paste, smeared a good layer on each plate, covered it with a sheet of Cling film, and gently placed the gantry on top. Excess epoxy oozed out and was removed, and the whole thing left to cure. Once cured, I lifted the gantry clear, removed the epoxy from over the tapped holes (a small square of sellotape over the hole ensures that the threads stay clear), and the gantry then dropped neatly back on to its prepared bed with no wobble. As I said, allow for inaccurate build! The holes in the gantry feet were also made a little oversize for their bolts, to allow a small amount of movement for squaring the gantry (at a later stage).

The whole machine sits on castors bolted to simple welded brackets so it moves easily around the workshop if necessary. However, in use, it sits on home-made feet. I welded M12 nuts to flat plates across the bottom of the legs, and turned down the heads of 12mm bolts so that they fitted into counterbores in wooden pads to sit on the concrete workshop floor. I also welded on M12 nuts so that the feet would be easily adjustable with a spanner (as I had just turned off their heads which were hidden in the wooden pads anyway…), and a locknut makes sure that they don’t move afterwards.

A lower shelf of thick MDF rests on the bottom rails, and I welded on a set of brackets to carry a second shelf halfway up and across half the area of machine. There is just enough room to bolt the control box to the frame so that it sits, in effect, on the lower shelf but is easily accessible, and at the moment the PC and monitor also sit on the lower shelf. I’m still making a swivelling bracket to hold the monitor at eye level, and I use a wireless keyboard with built-in touchpad. However, I have just started using one of the wireless Mach3 MPG devices from Aliexpress which is making things easier and I wouldn’t want to be without it now.

The bed is supported by a simple grid of 50x50x3 box welded in place. I did my best to ensure that the grid was flat, but I had a plan to allow for inaccuracies anyway. I have bolted a set of strips of hardwood to the bed support rails. I used U-bolts sunk into grooves below the top surface of the wood. I then machined the top surface of these wooden strips using a 50mm face cutter, and using the machine itself. This guarantees as far as possible that the bed supports are parallel in all dimensions with the machine X and Y axes. The bed itself is a simple sheet of 19mm ply which seems to be flat enough for my purposes (mostly woodworking) and if I need more accuracy for small components I machine a spoil board in situ. I have also added ¾” aluminium extrusion track to take T nuts and bolts for work hold-down; by machining the grooves for the track with the machine itself I can also align work pretty accurately using the tracks. The track comes from Rutlands – they have it on special offer fairly often. I also bought a big set of T-bolts and clamping knobs, again at special offer prices. They are intended for ½” track, but I made up some sliders to make use of them on the bigger track. Really useful.