In this video I am going to make a heavy-duty CNC machine but as cheaply as I can. It will be different from the previous Moot One and Two machines which are designed to be more easy to calibrate, and can hold multiple tools. This machine will be equipped with one tool – a spindle router, and will use the cheaper SBR20 rails instead of the HGR types I used previously. Overall, there should use a lot less parts and some of these parts can be bought in bundles – keeping the cost down. So here goes.
On screen now I am cutting a cradle from some MDF to hold 30×60 aluminium profile which I’m planning to machine mounting holes into. I’m doing this on the Moot One desktop CNC machine – which is the first CNC machine I documented as a manual in case anyone wanted to do the idiotic thing of building one too.
I slot the profile into place and clamp it down using some mending plates. Oiled the area about to be milled and then plunge my 6mm single flute cutting tool into the material by accident. If you don’t have the occasional blunder while CNC’ing, well you’re not mortal. In this instance I didn’t change the depth of the material when going between the cradle job and hole milling one. But look at that – how rigid must my CNC machine be to get away with that. I can also hide my mistake by placing the profile on the inside face when joining these parts up later.
I resaved my files and had another go – first without the profile present, and then with. I need to make four sets of holes on either side. One pass will make space for the head of a socket cap machine screw. I’ll list my feeds and speeds on screen now.
- Aluminium Cutting Feeds/Speeds:
- Pass Depth: 1mm
- Step Over: 20%
- Spindle Speed: 18000 RPM
- Feed Rate: 1000mm/min
- Plunge Rate: 700mm/min
- Machine Acceleration 60 mm/s^2
I also listed the machine acceleration – although that is something normally determined by the CNC controller, I think it’s important to mention as this has a knock-on effect on actual feed rates.
Once those were all done I swapped over to a thinner 5mm tool and drill the through holes. I’m using a slightly thinner tool than the 6mm hole requires so that there’s more room for chips to be evacuated. I mill halfway and flipping the material to cut from the opposite face, and I have to do it this way because the tool can only cut a maximum depth of 17mm.
Altogether this took me a couple hours of work – and of course would have been a lot quicker on a larger machine but as you can see the 350x450mm bed of the Moot One can be expanded with the right techniques.
I’m now cleaning any swarf from the profile and preparing to tap the ends of the four joining pieces. I use a bit of WD4 0 which I spray into a small vessel to dip my spiral tapping bit into and drill out the M6 taps.
And as simply as that I can fasten the base section together. The squareness of the part is reliant on the how the profile is cut – and I had this done professionally when I ordered mine from KJN. This is not a paid sponsorship – I shopped around and so far these guys are the best. Delivery is more expensive so you will want order everything in one go but it’s worth paying for – if only because their slide in t-nuts which have a special feature which I will show you in a moment.
I’m now unwrapping the SBR linear rails. These are 1 meter long or 1000mm long. I have clamped my base frame to my worktable and will loosely fit one rails on now.
Ok I’ve just clamped this to the table so it won’t fall over and I’m planning to fit these linear rails into position and with this new profile I am using, the t nuts are slightly different and I’m able to slot these in from the through-face. So I don’t need to slide them in from one side and remove the machine screws in any way. It is a slightly different way of assembling but what this means is I need to mark the holes which I’ll do as simply as this. And I can now start to pop these int and align them up.
I mark the location and as you can see fir the t-nut at the points they are needed, without the need to slide them from the cut end. This alone makes this profile system a more user friendly experience, but also any building process far easier to document.
And to secure these I use M5x14mm machine screws. I gently thread these in by hand but I’m not going to secure them…
As I’ll need to level these at some point. The more expensive HGR rails have a ground flat edge which makes this process a lot easier – but for the SBR types I’ll have to work out a more rudimentary way to do this.
I slid one carriage on, which came supplied with my rails – and these use small captured ball bearings and can sounds a little noisy. There is a more expensive option of polymer bearings although I’ve not seen much comparative information for this particular type.
For the next step I’m cutting the front and rear base plates from some scrap Valchromat. One side will also hold the stepper motors and I’ve made a recess for proximity sensors on all four sides. These are my feeds and speeds and I’m using a two-flute cutting tool.
- Soft Wood Cutting Feeds/Speeds:
- Tool: 4mm Carbide, 2 Flute, Straight V Flute
- Pass Depth: 4mm
- Step Over: 40%
- Spindle Speed: 21000 RPM
- Feed Rate: 1600mm/min
- Plunge Rate: 750mm/min
- Machine Acceleration 60 mm/s^2
I really need to order a single flute up spiral bit – as the one I’m using is actually designed for plastics and it’s noticeably nosier than what I would prefer to use. I’ve split the job into two parts beginning with all the inner shapes. Now I can check that I can fit a captured nut into a recess – which it does nicely so I can carry on with the second part of the job.
Ok I’ve got the plates here and I’ve already done one side, so I’ll show you how I’ll do the opposite side. This is the front end where the stepper motor be mounted, and this is the back end. I’ve provided the option for proximity sensors, but I’ve got an idea of setting this particular CNC machine up with sensorless homing. So potentially we may not be using these.
Before you do that you need to prepare your ball screws and the first thing you want to do is to fit the nut-housing on with some M5x16mm long socket cap machine screws. You’ll then want to install the BF12 end support.
This end gets secured with a circlip however the bearing only push-fits into place – to allow for micro expansion and contraction along the ball screw length.
This is done by using one of these circlip pliers – locking it into place like that. [Distant bus horn sound] Thank you very much.
On the moot One CNC I didn’t use this equivalent part – instead locking at the opposite end of the stepper motor and essentially relying on the motor coupler to do the job of the floating end support – which his very little work at all. This made for a more compact machine but work out less cost effect as the FK12 part had to be bought separately.
The BK12 fixed end support gets attached to the opposite end. This is secured with a nut which is threaded and locked into place. The square recess protrusion facing outwards. I’ve also removed the grub screw on the locking nut and will replace it with a M4x6mm cap screw – which is less likely to thread.
To tighten this up, and really this locking nut is the most important part of what prevent any backlash from the ballscrew fixing anchors or fixing points. And to do that I’m going to use a very thick piece of neoprene rubber within the jaws of a mole-grips which I’ve worked fully opened and tightened bit by bit until I can secure it without feeling I am damaging the ball screw. Because that is one thing you really want to avoid. And with that in place I can use a spanner to tighten that up like that. And that should do it. I then tighten this and that should stay in place.
You can see how that pops into position, and the socket cap locking screw can move freely without impeding on the plate. On this side I’m using 45mm machine screws to hold the parts together. On the 3D model I didn’t use washers on this side but I have the capacity to add then so I’m using. And I’m only hand tightening these for now as there will be some alignment callibration between the ball screw to the linear rails which will have to be done later.
You can see here we still have room for the captive nut for the stepper motors.
Back in the floating end I can add the opposite plate and this time I’ll be using 40mm long M5 cap screws which I again hand tighten.
Ok I just realise I made a mistake. The square top section have to face outwards.
While doing that I found my grease nipples – I have to of them, as do most people and I’m adding onto the y axis for now while I’m waiting for more to come in the post. I have to loosen the machine screws as the hex section at the base is a little wide. When assembling you’ll need to imagine the plate position and how you can access it with the grease gun.
Ok the next stage is to add the carriage blocks onto the linear rails and there doesn’t seem to look like there is a space for a grease nipple on these so what I’m going to do is squeeze some into the ball run. [mumbling]
I’m now securing the front and rear plates to the base aluminium frame. Using M6x20mm socket cap machine screws, washers and the T-nuts. I also make sure to check the plates are square before tightening.
This larger profile is much thicker than the stuff I used on the Moot One, and while it does make this a lot heavier it does feel a lot more rigid.
You’ll want to check this has been assembled squarely – and you can do that by measuring between the diagonals. That is 1101mm, and that is 1101mm as well. Another method you can use is the 3,4,5 method – where you measure 300mm along one side, 400m along the adjoining side and the hypotenuse should equal 500mm if square – which my base is.
Before attaching the Y plates, I am going to make sure that both Y axis linear rail are as parallel as they can be. This is difficult to do because the carriage don’t have a ground reference edge and they also twist on the rails, so I can’t use those as me reference points. I tried to measure the round linear rails themselves, finding the apex with my dial gauge magnetically fastened to an engineers straight edge – which sort of worked on one end of the machine, but I found the tolerances difficult to adjust along the entire length. Eventually I decided to measure from the top edge of the aluminium support section. This is not ideas but a reality of using these type of linear rails. I will later take other readings and measurement as the machine is more assembled.
I can now start mounting the Y Axis plates which I cut out from some scrap moisture resistant MDF. I am planning to replace these at a later point but for now I just want to make sure the 3D model is accurate. I attach these on either side using M6x20mm long machine screws to the rail carriages and M5x20mm ones for the nut housings. I have the stepper motor Y plate to the left-hand side, after which I fit the X axis profiles.
With all the plates that have carriages on them you want to make sure those are installed square to the plates – assuming the plates are cut square. This is part of the rudimentary or course alignment. After the machine is assembled, we will do the fine adjustments.
I’ve include space on either side for an additional large bracket which address a little more rigidity. I’m now unpacking the x axis ball screw and linear rails. I grease them as per usual using a lithium calcium grease. I then add the t-nuts onto the x axis profile and loosely secure them in place.
My trusty red pen.
I’m not going to overtighten the machine screws here as again they will need to be aligned at a later point. I’m going to install the side profiles – which you can see in the 3D model here. These will act as a support for the Y axis drag chain but also a physical barrier to the machine moving parts – which should reduce the likelihood of stepping into the machine and getting caught. You could also use these as a fixing point for a rope switch estop if health a safety is a legal requirement.
I’m using scrap profile for this part, so the mounting hole happen to be a lot larger than what I designed for. To get around this I use a spiral thread repair kit to give myself something to anchor too at the size I ned. Essentially what I’ll be using are these coils. To prevent the profile from twisting I’ve 3D printed some parts which will go over the profile and endplate and add a bit mor support. This particular piece here will connect to the drag chain. If I was lifting this machine, I’d still do it from the front and rear as these are not designed to be handles.
Next I add the stepper motors for the Y axis on their 40mm standoffs. In retrospect it is worth putting the nyloc nuts at an earlier point. It’s a little fiddly to do this as I am now demonstrating.
I then added the X axis ball screw into position – using a piece of scrap material to clamp the profiles too so that nothing bends and breaks. Again I only hand tightening the fixings for the fixed and floating end supports. Before I align these I need to cut the carriage and Z axis plates – which I happen to cut from birch. As per usual on screen are my feed and speeds and I’m using a single flute up-spiral bit to cut this out.
- Hardwood Cutting Feeds/Speeds:
- Tool: 4mm Up-Spiral Single Flute
- Pass Depth: 4mm
- Step Over: 40%
- Spindle Speed: 18000 RPM
- Feed Rate: 1400mm/min
- Plunge Rate: 700mm/min
- Machine Acceleration 60 mm/s^2
The parts cut out quite nicely – but before I assemble this I want to cut out some 40mm washers from 5mm thick aluminium. I’m going to attach these to the base of the z axis SBR20 linear rails and will be the part the carriages hit against while sensorless homing. Even though I have a coolant nozzle on the machine I haven’t set this up with an air feed or coolant – so I just place cutting oil onto the surface of the aluminium and hope for the best. I mill the centre hole first – and then cut the outer path of both circles using large tabs to prevent the parts moving and becoming damaged. I then screw down the disc with a cup-washer and re-cut the outside path this time without the tabs.
I now need to add this discs that I’ve cut out to the base of the Z axis linear rails. And I need to drill a hole in the centre here. I don’t have a lathe what I have done is milled this little jig which fits into place and gives me a centred hole I can drill by hand. It doesn’t need to be perfects, and just needs to be centred so the rail carriage housing knocks into so I can use that for the senor-less homing.
You can see that spot looks pretty good. I’m now going to clamp this into my engineers vice and see if I can do a little bit better.
It doesn’t look bad. It looks more or less centred.
Ok the machine looks a little bit different here. I’ve cut some of the plates out of birch just to make it look a bit more uniformed… And solve a few problems that I didn’t notice in the CAD model. I’m now going to show you how to level the X axis, and the way I’m going to do it is from the base section – using an engineers straight edge as the reference plane. I first make sure the aluminium profiles are parallel, and then turn my focus to the linear rails.
The first thing I want to do is make sure the machine screws are only a fraction loose so there is no wobble in the piece but I can still adjust. Essentially tight it up and a ¼ turn so it can still move up and down.
I begin by finding the halfway point between the lowest and highest positions of the top rail. I take a measurement with my vernier calliper – half it, adjust the rails and secure it with one machine screw. So this is our reference point and this is where the pivot will occur. I followed the same steps as with the Y Axis and squared the X axis linear rails to the base section – measuring from the linear rails’ aluminium support. And using the dial gauge on an engineer’s straight edge to measure from – making sure it is supported by the joining base profile.
Once I was happy – and there was roughly a 0.15mm discrepancy across the entire length, I secured the top rail to the profile. I then fitted the carriage plate to the top pair of carriage blocks, and square those – well I tried too.
It’s not perfect but that’s due to the carriages being a little bit off. We will be able to get around that when we add the z plate because that has its own calibration process. If I forced it up into that position and adjusted the ball screws (and linear rails) based on that. What you would fine is when this moved this would travel down or up like that.
I adjusted the base linear rail to line up with the lower mounting holes on the carriage plate and again used the dial gauge to align that across the entire length. The final thing was to drive the carriage to either side of the X axis, and secure the fixed and floating end supports. Now just in case you were concerned that my plate might not be square – here I am checking it… It’s pretty good. And I also took some measurements from the bottom of the carriage plate on either side of the X axis.
So there’s a 0.05mm difference between either side which is excellent, and mean the carriage plate is moving parallel to the base.
But how do I explain the discrepancy we saw earlier? If I measure across the top of the x axis carriage blocks with my engineers square, you can one carriage is noticeably lower and that is to do with how the part was extruded and machined. We have compensated for that in this case, and also will also have some pivot adjustment when fitting the z axis assembly, so I will carry on.
I am now assembling the Z axis assembly. First hammering in the seated nut and then installing the ballscrew. You may notice the bit of 3D printed plastic which acts as a packer between the nut housing and the carriage plate.
I use the dial gauge to align the screw to the side of the plate. Locking the fixed end support first and opposite floating end afterwards.
I also re-machined the end-stops with an edge cut off as I had widened the plates which created clearance issues.
So there is plenty of wiggle room in the linear rails which we will use to our advantage, when it comes to levelling this so that it is 90 degrees to what will be the wasteboard.
And just be careful with this bit, so you support the z plate and it doesn’t accidently drop onto your fingers or something and get damaged.
That is pretty good.
I’m using the dial gauge again to check the z axis spindle plate movement for squareness along the y and x axis. As I already concluded, it’s pretty good. After which I installed the z motor plate, bracket and other parts. I’m using some smaller stand-in motors while I wait for a set of more powerful 2.8amp ones to arrive in the post. I also cut the spindle mounting adaptor plate off screen as I’m sure you are probably sick of CNC by now.
So, this is the plate that will mount the spindle onto, and the spindle bracket. And the only difference here is I used a drilling toolpath where I’ll be mounting the spindle. And I’ll just list the feeds and speeds on screen now.
- Hardwood Drilling Feeds/Speeds:
- Tool: 3 Flute HSS 12.4mm Countersink
- Depth: 2mm
- Step Over: NA
- Spindle Speed: 18000 RPM
- Feed Rate: NA
- Plunge Rate: 500mm/min
- Machine Acceleration 60 mm/s^2
The final few things to do included securing the drag chain to the X and Y Axis of the machine – including an aluminium support bracket at the rear of the X. The drag chains are fitted onto 3D printed parts, and I also completed the wiring to the machine, from a new CNC controller. Here I wired the stepper motors and limit switches using in-line leaver connectors to the shielded cable – which are nice and easy to use.
The sub waste-board and sacrificial cutting surface are things I’ll add later. So now I’ll show you the new controller. I didn’t document the process as it would have made this video twice as long but I do have a series of videos about the previous one which may be a good reference and worth a watch – because I’m so enthusiastic and positive.
I was really impressed with the sensor-less homing process for the z axis. Which was easy to set up and calibrate. Had I known this, I’d have cut out more aluminium stops and fixed them to the end of the linear rails for the X and Y Axis. The last thing I did was fit the sub-wasteboard – which I had cut in four sections. While a solid piece is nicer to work with, having made it this way meant I could cut it all out on a much smaller machine and had I owned a domino or biscuit jointed, aligned the faces before fixing down.
So that is the moot three CNC machine – which awkwardly follows the Moot One, while the Moot Two is still a work in progress. There’s some stuff to be done including selecting and installing a spindle router and finishing up the system files, but I’ve come to a natural end for this video. If you’d like to support this channel please check out my patreons page, or check out my shop where the files for the machines I’ve designed can be bought. I’ll link to those in the description, which leave me the final thing to say… If you liked or found this video useful, sacrifice a thumb to the algorithm gods and let me know what you through of the video in the description. And you’ll catch me in the next one.