Sunday, January 20, 2008

Thinking about the Machine's Design

Machine Specifications

One of the first things that was clear about this machine is that it would be nice for it to hold a CAT40 tool holder, if only for popularity reasons. With 12,000 rpm max, maybe 15,000 rpm should be enough for most jobs. By deciding on a CAT40, the max rpm is set and other characteristics of the machine are also set. This machine will not move very fast (800in/min but rather more like 300in/min) when cutting. Given the rpms, it will be a strong machine in the sense that to make it an efficient machining center it will need to take larger bite - hence requiring more torque on the spindle and thrusting force while cutting. This is very different from the other end of the spectrum – high-speed routers that are kind of flimsy. But given their high cutting speed, they can experience reduced cutting forces to maintain accuracy. The CAT40 will allow for large diameter tools which are ideal for some jobs. Plus in the future we should be able to have a different tool holder.

Now what about work volume? Most entrée machine start at 20x20x16in and this sounds like a good place to start.

What about machine configuration? Should we go for a traditional C shape (Bridgeport style) of gantry style? After a lot of thinking and reading about this I really like the gantry style. You don’t see it on entrée level CNC machining centers but given a new choice of material (which I will talk about later), and the fact that I want to design a style of machine that could scale to larger size, I think it is a perfect fit. Gantry also is perfect for limiting thermal distortion caused by hot chips and reducing the need for way covers if all the moving components are kept high above the work piece. I also like the fact that there is less variation of translating mass when you move the work head around instead of the work piece.

Construction Choice

Lets talk a little bit about the vision. One of the goals is for this machine to be accessible to many. And by this I mean almost anybody. So I am also focused on the fact that I want to make a design that can be built in any moderately equipped shop. Also the cost of material is low ($10k), so some traditional technologies will have to be abandoned.

First, iron castings: Although at productions scales they can be cheap, this is not a technology that scales down nicely - so castings are out. Maybe components like T slotted tables (which you can get as a standard component) may fit the bill but a machine base will not. So what will I use instead? This takes me back to some of the work done by one of my favorite college professors and precision machine design guru Alex Slocum. He came up with base designs with 4x improvement in vibration absorption performance over concrete filled in steel structures. Here is link to the article it basically consists of using thin layers of high viscous shear damping material in-between hollow sections. Sounds fancy, but it is not. These materials are everywhere from honey to tar. And if used in the appropriate place in a structure they can significantly dissipate energy. In the machine world this translates to reduce transmission of cutting vibration force and results in smother finished surfaces. Square and rectangular steel tubing will be ideal for this application given the low cost, availability and easy of manufacturability.

Second, nice quality linear slides and precision spindles: These are expensive. I want this machine to be as good as it can be for $10k and to keep the cost down. To do this, I will have to resort to two strategies. I can look for these components in China - hopefully their price has come down or I can down grade them accordingly, but that will be my last resort. There is an alternative, which I plan to develop in parallel: Low viscosity hydrodynamic bearing design, again by Slocum. Hydrodynamic bearings have been around for ages, even before ball bearings. The concept is to slide on a thin layer of fluid. In the past, the fluid of choice has been oil, but oil had its disadvantages like high resistance at high speeds and specialized skill involved in the manufacturing of race ways. Although hydrostatic oil bearings are still used in the high-end precision machines, ball bearings with their high efficiency, relative low cost and cleanliness have taken over the market. They are not as precise but, although cheaper, they present a challenge when trying to keep the $10k budget. The alternative I am proposing is pressure assisted low viscosity hydrodynamic bearings. These are bearings that use water or a fluid like water instead of oil - they don’t require the manufacturing tolerances of oil bearings and they can move with little resistance, given the fluid’s low viscosity. If low cost spindles and linear ways are not available, I am positive this is the way to go. The plan is then to design a hydro-block that anybody with a Bridgeport can make, and to reuse them all around the machine. These bearings also have very high load capacity - if done right, they will never wear out and have much better vibration damping characteristics over ball bearings.

No comments: