In the last video I re-assembled the CNC machine with the freshly cut acetal plates, and in this video I am going to level the waste-board, square the axis, and tram the spindle. But first I’m having a go at cutting some aluminium.
I didn’t put any tabs on this so it’s obviously cut on the side there but… Considering I didn’t use any cutting fluid the 1/8 of an inch bit went through the material quite well.
I was cutting at 15000 RPM, with a feed rate of 200 mm/min, plunge rate of 80mm/min and step down of 0.2mm.
The 15mm square came out 15.06mm along the x axis but 14.82 along the y. The y axis on this machine is between the front and back, and it’s likely the spindle is nodding which is contributing to the error.
You can see the opposite side is when the last time I cut this material was with the x carve.
The nutblocks, locking collars and couplers are all secured, and there is no obvious movement in that direction. I have already surfaced the waste-board and after measuring between opposite corners x and y axis are square. However there is a noticeable pattern in the MDF because the spindle is dipping along the y axis – this is called a nod.
I can see from the surfacing that I’ve done of the wasteboard, there is a slight raised on the back of the machine. So this is actually tilting forward. I can tell just by looking at the pattern left behind. From side to side through it looks pretty good.
I am going to check the alignment of the spindle by placing a dial gauge on an arm, into the collet of the spindle the checking how consistent opposite points are.
If the spindle is tilting the dial gauge will not read the same at opposite points, and I will be able to identify this.
But first I need to make an adaptor collar to fit the arm onto the shaft of a ¼ inch cutting bit. I picked a CNC bit which happen to have a short cutting head and long surface to clamp against.
So with this open I’ll just check the inside diameter, which is about 12mm. And this is obviously a 1/4 inch.
I cut this out from a scrap piece of acetal, and then cut through the wall so the collar could be compressed.
This time I’m going to cut with a step down of 2mm at a time.
That’s the side with the dust shoe kept on. Barely anything escaped. On that side – well you can see how messy this machine can be.
Ok that feels nice and tight.
Ok that’s how I turn off the signal to the spindle so if I accidentally press something and it were to turn on, or the signal were to be sent it won’t actually receive it.
Ok I’m going to take the dust shoe off.
Because this is a desktop CNC machine the profile is quite small and I couldn’t get the dial gauge directly behind the spindle and along the y axis. I should have also put some glass onto the waste-board to help with tramming and avoid some of the holes I’d already made on the MDF. It looks pretty good. But I managed to get some readings. It’s only moved two lines.
So across here it’s only off by 0.02mm… However form the front to the back it feels like 1mm or maybe a little less.
So the spindle is leaning forward like this, and according to this it’s close to a mm. The error with the nod is mainly caused by the weight of the spindle pulling the v-wheels on the gantry. Now this is where I have to admit that I didn’t tram the plywood machine. In fact I completely overlooked it, and I did notice this after I cut new plates, all the openings were ever so slightly narrower at the bottom of the cut than they were at the top. This would have been along the y axis predominantly and may have also contributed to the nod – but it’s mainly the weight.
If I place an engineers square on the gantry across the two plates you can see the gap on the plate closest to the front of the machine is about half a mm.
Ok so, these are some of the offcuts from the actual cutouts of the plates. This bit here is 0.6mm. If I cut that out and push that onto the back for now – I’ll tighten it up and have a look.
I placed some shims directly behind the spindle but it later occurred to me that despite this squaring the bit to the waste-board, but error lies further back – and the bit would travel along the y axis as the z axis moved up and down.
I decided I should check the rest of the machine a little more thoroughly.
While checking the y plates I noticed the aluminium cbeam on the righthand side had an imperfection.
When I slide this across it hits onto the edge here. It’s almost as if this section here is a bit lower.
A section seems a bit lower than it should, and I am not sure if this is effecting how the wheels rest?
You can see this side is a lot flatter.
I didn’t surface the acetal to make the two outside faces parallel, so I’m in no doubt that they are anywhere near perfect.
That means there’s something wrong with the profile (that I didn’t buy from China)
Ok the Y plates are a little bit out. It could be that when I was fitting all the corner plates, that I just tighten them up too close together and they’re pulling the y plates in a bit. Maybe if I loosen these up, I’ll be able to slide it around a bit.
The x axis c beam is also not entirely 90 degrees to the waste-board, but the holes for the x axis are 90 degrees to the fix wheels further down the plate, which run along the y axis.
Ok I’m loosening all these machine screws on this aluminium extrusion at the front to see if I can adjust this. I’ve also loosen all these here. I don’t have any clamps that I can reverse and push these out so I’ve used the hammer.
I moved the gantry to one side and tighten the machine screws, and then back to the opposite side where I secured the remaining ones.
Ok 496mm along.
I can’t tell if that’s made much of a different.
I am going to square the z axis now and to do this I had to take the entire gantry off. I had some thin 0.35mm washers which I placed between the gantry and the back of the z axis c beam. I placed two washers on either machine screw at the bottom, and one washer on the two middle machine screws. To square the z axis along the x axis I simply pushed the c beam to one side. That’s how I did it previously and the results were pretty good.
Ok I think I’ve got this as square as I possibly can. If I use a light. I checked for squareness by placing an engineers square against a bit in the collet and the wasteboard.
I then placed the dial gauge and checked the opposite sides. The x axis looked less than 0.1mm – so maybe not as good as the first time I checked which was 0.02mm but still pretty good. And along the diagonals it was similar so about 0.08mm.
I then surfaced the waste-board this time with a large step over of 90%, and used a straight edge and torch to visually compare the cuts along the x and y axis.
There is a pattern. It is a little bit higher on the right hand side which means the spindle is tilting to the left at the top. In terms of the front and back, it’s actually still nodding down a tiny bit but it’s a lot less than the x axis.
There is a little bit of light bleeding. You are talking very very small – maybe 0.01 or 0.02mm. Barley any light coming through. About 0.02mm. I can live with that. I’m just going to run the are clear again with the step over a little bit smaller so I can get rid of these corner bits that were missed and then that should be it. The machines ready to go.
This is still an improvement, bit I don’t particularly want to spend more time calibrating. Without designing something to adjust the axes incrementally I don’t think I’ll get it much better. I decided to try machine some aluminium again. The settings were the same as the first time but this time I cut a 20x20mm square.
So I put some tabs just to stop it popping out. It looks like it’s cut really well. The results were really impressive. 20.04mm along the y axis and 19.99 along the x axis.
So this problem with the not has prompted me to change the plate design and I’ll show you what I’ve done. The holes on the y plates which accept the mounting machine screws that go into the x axis c beam. I’ve actually stretched them. I’ve made them 4 degrees wider on either side. What this means is I can loosen those and pivot the z axis. I’ll just edit profile sketch so you can see this a little bit better. The shape is based on a centre point which is the centre of the threaded rod, so this means I don’t have to change or adjust anything to do with the coupler, threaded rod, the nut block or the stepper motors. And once I’ve found the position I can lock the four down plus the two angle brackets under neath on either end of the x axis. And it actually works very effectively. If you imagine the clamping pressure of the machine screws is quite a lot – maybe 30 or 40 kilos, when all of these on either side are tighten I think it’s more likely the bit will break if the machine ran into problems than for the pivot to come out of alignment. So this is my design innovation to deal with that problem of the nod. I’ve not seen any of this on any of the other CNC machines. I think it’s going to make a massive difference to the calibration time and also make the calibration so much easier. You can imagine – let me leave the sketch. You can imagine possibly having a machine screw on the inside here, which can be set so once these are loosen you can incrementally adjust the angle and lock it back down again.
I cut out a test section of the y plate out of some scrap MDF, and you can see how the x axis c-beam can pivot here – and in this shot how the centre of the acme threaded rod opening remains centred.
So I think I’m going to make this version of the plates available on openbuilds. I did start from an opensource design so it’s only fair to end there. But I’m going to also start working on a manual to accompany these videos and project, I will be dedicating a page to thank all the patreons who have supporting this channel and this build.
As per usual please sacrifice a thumnb, and let the algorithm gods know when you think. If you’re new to this channel and want to support it you can becoming a paid subscriber on patreon, where any changes to this CNC machine or future designs will be made available there before general release. I’m going to also try make more patreon only feeds – so that’ll be something else to look forward too.
Any thanks again and you’ll see me at the next one.