Hello, I was wondering, has twenty teeth been calculated as being the most efficient at achieving the finest resolution on the "belt type" cnc routers, specifically the OX? What factors go into calculating resolution for a given setup with respect to the number of teeth needed for optimizing pulley/belt/stepper ? On either side of the 20T pulley, what at the effects of say... a 16T vs. 24T pulley the same system? Thanks,
so, with GT3 belt which has a tooth pitch of 3mm and a 200 step per rev motor and 16x microstepping 1 step = (pulley teeth * pitch) / (steps * microsteps) = (20 * 3) / (200* 16) = 0.01875mm or 0.000738"that is much smaller than we generally need! We benefit from more power at lower microstepping settings 8x microsteps give us twice the above 0.00147" or 0.0375mm per step but quite a bit more power (which translates to more acceleration and higher top speed, up to a point). I must do some experiments with this, but really, if we are cutting wood, do we need more than 0.1mm resolution? Even that much? The wood will change size the next time the aircon kicks in anyway! Now, if we use a smaller pulley, we get more resolution, but we get less speed because we need more steps to go an inch. This might be useful if we are cutting PCB's for surface mount components. 10 teeth would give us roughly 0.01mm resolution(1) which I think is largely a pipe dream with belts of any kind. If you want to do PCB's with small features then please get a C-Beam machine (with leadscrews) and build it carefully. You decide what you want (-: (1) resolution : when microstepping this is theoretical, since the microsteps are never all the same size, ie calculated. resolution doe snot equal accuracy. The magnetic system inside a stepper motor is non-linear and external loads on it are also non-linear, so there is no way to predict exactly where each step will stop unless you have a feedback loop. You can buy that, just bring money. The fewer microsteps (down to 2x, full stepping suffers from horrible resonance) you can use the better for accuracy. so, to answer your question more directly, no, I don't think any kind of 'optimum' calculation has been done for pulley size, but you can use the information above to choose what you want in your machine. And later, you can change it if you like (-:
As David alludes to, optimization is related to what you intend to cut and if you are intending to cut a broad range of materials, your choice of optimization is really only between course (belts) and fine (screws) movements. And you'll find the difference between 16, 20, and 24 tooth pulleys on a belted system is truly insignificant in the overall scheme of things. While dropping to a 16 will offer a slight improvement in the resolution, there's still a significant gap between this and the resolution of a screw. The best way to achieve a substantial difference is through belted reductions. (The example shows a 1:2 ratio using GT3 belts but GT2 belts and pulleys would probably be just as workable and are far more economical and available and also easier to create alternate ratios.)
Very neat information guys !!! Okay so I have committed to building an Ox. What I will be doing, mainly, is cutting guitar bodies and necks. Maybe some other stuff here and there but nothing planned other than the guitar stuff. I can modify the Ox design to use a screw, rack and pinion etc... and long before I posted this thread I had been thinking about using a rack `n pinion on the ox. I was going to use a few parts from CNC Router Parts on their Rack and Pinion Parts Page. Being that I am building a 1m x 1m router the rail they sell is the perfect length and would not need to be cut. So that is one way to mod the Ox. I am not sure how I would use screws on it... or at least I have not thought it through... Any ideas? Here is one of their videos showing off the newer version of their R&P.
Actually screws are the easier of the choices. There are several examples in the forum including one here. R&P is what is going to be difficult. Integrating the rack on the face of the rails pushes the side plates out even farther requiring extensions to support and carry the wheels back to the V-slot grooves. Doable but a fair amount of work and coordination.
If we picture the Ox configuration... The V-Slot wheels remain in place on top of the 20x80 and the rack is mounted on the lower outside face of the 20x80. The lower v-slot wheels would still be there to make contact with the bottom of the 20x80, they would need longer bolts. There is also a heavy duty spring providing a load on the pinion to keep it firmly on the rack. In addition, on each 1000mm side of the router there will be 5 pieces of 6" x 6" x 3-1/2" aluminum angle bolted to the inside to provide rigidity/support to the system. I think you were on a thread where I mentioned the design once before... I got the idea from the charging ox cnc router build on this site. The maker did the same thing with aluminum angle to provide rigidity to the 20x80. I would probably need to go with hardened bolts for the wheels because they will need to be longer due to the rack placement. The gantry plates already sit out X amount of inches/mm from the 20x80. So with the addition of some hardened bolts that are longer and some spacers the .5" rack will cause the gantry plates to sit out a bare minimum of ??? from the 20x80. Another modification I will need to make is a swing arm for the roller drive spindle.