There are certain pieces in my clocks that just give me fits trying to make them as small as I want them to be and as accurate as they need to be. In order to be able to test my ideas and designs I needed (wanted) to be able to remove the variables introduced by my shaky hands. In reality I can usually (but not always and not consistently) split a 1 pixel line on a printed pattern but at times that doesn’t seem to be enough.
The following photos outline the development of my CNC machine. Having completed the CNC I would say that this is not as nice and simple as one might think nor is it cheap. By the time I was finished I had spent over $2000 probably closer to $3000 (selective memory). The finished product will do what it is supposed to do and that is produce accurate items from a CAD drawing.
Having said that there are other things to keep in mind. First you need a CAD program and then the ability to use the program from a designers point of view. I find the ideas to be the easy part, now putting them into practice you need to be able to come up with a good tooth form. You can approximate these quite easily but then why go to the trouble of great accuracy in the cut if the mathematics aren’t there in the design at the outset.
Next you need a CAM program that will convert the the dxf drawing into gCode that the machine’s controlling program can understand. This is relatively easy as the program just does its thing. Editing to get it the way you want it is another thing altogether. In the end I decided to just let the program do its thing even when it didn’t seem too logical to me from a cut sequence point of view.
You’re not through yet because you have to have another program, in my case Mach 3 which feeds information to the board and ultimately the steppers. There are really good video tutorials for Mach 3 and I have watched these dozens of times to try and get the set up the way it should be. It is not as intuitive as one might think.
A CNC is not the answer nor the replacement for the scroll saw, unless you are going into “mass” production of a clock. However I still feel that it is worth all the work and effort to be able to cut an experimental escape wheel and pallets to an accuracy that ensures it isn’t the cut that’s in question when it doesn’t work the way expected. As to cost effectiveness -- NOT!
As Clayton said in an email “Luke!, come back from the dark side”
Like any project you start with the pieces. This photo shows the 3 - 305 oz stepper motors that will drive the x,y and z axis movements.
The driver board requires a substantial transformer that puts out a low voltage for the board and motors.
The large blue capacitor is used to “smooth” out the breaks caused by the rectification of the current.
The board is produced by Hobby CNC and is a unipolar microstepping driver.
Their web site is at
The products are well thought out and there are literally thousands of people on “Forums” to help with whatever problem you have
The unit comes as a kit and you have to put all the little pieces on the board and solder them in place. The board is clearly silk screened to identify all the locations of the pieces mentioned on the 4 pages of assembly instruction. None of that precludes one from making a mistake! I did.
This top view of the board shows the three blue terminal blocks at the top - these make the connection to the stepper motors. So in effect there are three identical circuits on the board. Actually there are four as this board can be used in a 4 axis application.
The unit connects to the computer using a parallel port.
When assembling the case the first installation is the diode block which will convert the reduced voltage from the transformer into direct current required by all the components
The 3 axis circuit board is finished and ready to be assembled with the rest of the components.
At this point I didn't know that I have soldered one of the voltage controllers in backwards so nothing works.
Next to be mounted was the transformer and the circuit board with its aluminum sheet heat sink. All the mounting holes are clearly identified and for the front and back a paper template is provided.
A 24 volt fan is installed in the back wall of the box and the parallel cable has been put in place and exits out the back. The multi colored wires now lead to the stepper motors. At the lower left is the main switch, and two fuses for protection.
A test set up shows the controller and the steppers such that they can be run without fear of “running away” and damaging the table.
I thought I could use an old windows machine and a PCMCIA - parallel adapter card but that proved impossible and it turned out to be money wasted.
I purchased a PC laptop that was old enough to still have a parallel port and it works.
I purchased a “Mini” CNC table from
It arrived in a box with some metal pieces and some plastic bags of fittings. This was nearly 1/2 the cost incurred as it was $1000.
Although there were no instructions on how to put it together the little beauty can be put together just by looking at the pictures of a finished table.
The fit and machining of the table is of very high quality and everything fit as it was supposed to.
If I were to give advice I would say get a profession table if you need accuracy.
The table is a beauty. It has a small table that will cut an 8 inch gear which is likely more than I will need. The stepper motors go on the top the near side and the back.
All the parts are at least now on stage -- if not in an operating condition. I am starting to get excited. I have been able to get the computer program to import a DXF file and convert it to g-code and then simulate the actual cut. No sawdust but encouraging!!
For a spindle Dremel has put out a 75th Anniversary Kit which is a really nice tool. The body is round and not contoured and so it fits into the tool holder of the Widget Master table. It is also very simple to operate just turn it on and select the speed you want and its a go.
This is the very first successful cut. I had experimented with a pencil as a spindle just to “check it out” and then put on the dremel. The lack of symmetry on the right side occurred when the connector between the x-axis motor and its screw came loose and allowed the motor to turn without the screw.
Well here it is a small gear that has the possibility of being a better cut and more precise element than I could do by scroll saw. I have pretty much solved the motor stalls through a lot of help from the on line forum.
The real test will be when I cut a full clock gear set and put it in beat.
The portable laptop computer was definitely problematic as it was inconsistent in what it would do. 90% of the time it did what it was supposed to and would cut a gear nicely - then obviously I would do something (don’t know what) and it would not co-operate. A used IBM for $200 turned out to be the solution
The CNC doesn’t have a special place of honor just a cleared area on a work bench - surrounded by small hand tools and jars of screws. Since only the gantry moves you don’t have to give special consideration to where the table needs to go.
The scale doesn’t show the best in this shot but the biggest pinion here is 0.6 inches or about the size of a dime. I cut 3 of each of these little pinions and they overlay perfectly.