Monthly Archives: June 2011

CNC run down

The cnc machine is now complete so I can explain what Ive done to turn a minimill into a working cnc mill. It will be fairly brief but youre welcome to email me and ask for more details.

First I bought one of these …

201106291224.jpgThis is a SIEG X1 super micro mill mk2 from Axminster tools. Its built in china, its cast iron and its relatively cheap. It has its downsides like poor lead screws and nuts, and a slow spindle speed (2K rpm) but its good enough for me for now. So this will work as a nice stand alone desk mill cutting most materials, although I haven’t tried any steel yet.

Next you need some stepper motors. I bought 3 of these hybrid nema 23 stepper motors from motion control products.


The stepper motors need drivers to power them with 24v 2.5A so from the same place I bought these.


The drivers need step signals to amplify to the motors, so you need some kind of controller. I wanted to run everything off a small net book so it needed to be USB. There arent many USB controllers on the market. I went with the ncPod as it was recommended by a friend. Its still in beta and doesnt look like its full anticipated functionality will ever be realised, but the company were responsive and the board is a good price and everything it can currently do is more than enough for me.


None of this will work without power. I originally hacked a pc power supply, but it didnt supply enough volts (24 is best for these drivers) or enough current. This means the motors will only go slow when cutting or they’ll slip. The quicker you can run your machine the happier you will be. I eventually went with this supply from farnel. XP Power PSU 160W . To power the ncPod you will need a 5v switching regulator.

To attach the motors to the machine you need some kind of motor mount. You can buy kits from cnc fusion for the X1 but theyre expensive and you can make them easily yourself.


To make a mount you just need to rigidly attach the motor in line with each lead screw, and connect the shafts and lead screws with motor couplings. Couplings can be expensive so I went to ebay to save money. They were not the best quality but they work fine.

With all the mechanical stuff attached you need to wire the steppers, drivers and controller. Spec sheets should show you how to do this so I wont give the details. I used 4 pin xlr connectors on the box that houses all the electronics. It makes the cabling robust and looks more professional.

To talk to the ncpod, and make anything move you need some kind of cnc software. There are free examples available, typically on linux, but they wont work with a usb controller, so Im stuck with windows. Mach3 is a well developed free hobby come professional cnc tool that reads gcode and deals with all communication to the electronics. Its well worth the £100ish pounds… Mach3

Once you have Mach3 up and running you have to configure it for things such as the number of clicks per rotation which is determined by your motor resolution, screw pitch and microswitch resolution. You will also need to determine the maximum speed and acceleration for the motors. This is trial and error. I have my acceleration quite low to reduce the risk of the motors slipping. If you press tab when in Mach3 it provide you with a jogging window to move the motors or you can input gcode commands into the command line.

The final piece of software you need is to produce the gcode. For this I use CamBam. Its still in development but its already very good. You can define all your cutting paths and even do 3d profiles. I tend to not use the CAD functionality as I prefer AutoCad/SolidEdge but it will do most things.

So all thats left to do is CAD -> CamBam for gcode -> Mach3 to execute -> Watch your machine cut stuff!

Mechanical Eye – cnc test

I am working on a new project to make a face tracking steam punk mechanical eye. This eye hides behind a mechanical iris. I have prototyped this iris with the laser cutter in mdf. There is also an outer ring which when rotated pulls and pushes the jaws open and closed. The idea is that this will all be cut from metals such as brass, to give it the authentic steam punk feel.

IMAG0281.jpg IMAG0280.jpg

mechanical iris.jpg

This is a perfect opportunity to use the new cnc machine. I used a 3mm cutter at the maximum 2K rpm, and managed to happily cut at 90mm/min (0.25mm depth cuts). So far Ive cut one jaw and one link…. I now need more brass! it all takes just over an hour to cut which Im quite happy about.



CNC Circles!

When I finally thought my machine was ready and working, I soon realised it wasn’t. I had to machine out some nema23 motor mounts for a new machine at work. So it seemed a good idea for a first real cut to test in MDF first. The results were a little skewed. Well all circles (drilled holes) were elongated in one axis, and it took ages to cut, mostly because it was drifting at any real speed. So after 6 or 7 hours, I found my work waiting for me in the garage. It wasn’t perfect but the wholes were roughly right. I then tried this is aluminium, and whooooah, drifting going on everywhere! This seemed strange as Im using quite strong motors at a slow feed rate.

So i had to solve this problem before any real CNC work can begin. This involved cutting a series of circles on top of each other and checking their concentricity. In mdf this almost worked, but with some slip. In acrylic the slip grew, and in aluminium it became ridiculous. In fact as you can see in the picture, the movement stops in one axis at the same point in each circle.


At first I thought it was because i was using a pc power supply (a bad choice at trying to save money) to power the motors, so I bought a good hefty one of almost 200W at 24v rather than 35W at 12v. The result is the motors move much quicker when jogging, but still cut the same rubbish circles! So back to the drawing board.

I decided to put a pencil in the spindle to draw the circles onto paper and start to measure the drift etc. to really find out whats going on. Maybe its the bed binding, the control board not working or the gcode being faulty. The results show almost concentric circles, only there is slight drift in the x axis.I tried this at different speeds around the workspace to to be sure it was consistent. The only consistent thing here is that its only the x axis that has problems…. so swap the drives around to see if the drift follows the y axis instead…. and it does! I have a duff driver. Its really odd that the driver half worked, I would have thought that if it was duff it would not work at all.

The good thing is that all the axis work as they should now… and I have cut my first official part, a voidnoise tea coaster!…