Easy Prototyping
Comments from a Compulsive Gadgeteer

By Roger Baker

Although inventing may often be 1% inspiration and 99% perspiration as Edison claimed, our goal should be to reverse this ratio however possible. One of the things I firmly believe an independent experimentalist needs to be best at is in building cheap and workable prototypes quickly. By prototyping ideas as cheaply and rapidly as possible one can get the information needed to encourage more effort or to help build an improved model.

If you can build and test and quickly rebuild ten "cheap and dirty" versions of some gadget or principle while someone else is coming up with one beautifully machined model made in a well-equipped machine shop, who is usually going to make the most rapid progress? When and if you get to the stage of commercial production, the customers will want to see fancy enclosures and shiny well-machined parts, but the goal of the experimentalist should be to get to that stage the fastest and easiest way possible.

If you understand the underlying physical principles of some idea, then you can build something just good enough to demonstrate the key principles with the least possible effort and make the most progress with the least effort. Things rarely work perfectly the first time and few are clever enough to anticipate way in advance. Imagine what would happen if a computer programmer had to wait weeks to see if their code revisions would run? Inventing is a process that demands the most immediate feedback possible.

But enough philosophy and on to the specifics. If you don't have a big machine shop but want to build fast prototypes of a wide variety of instruments and gadgets, I suggest you start with a decent jewelers saw from a local supply store for handcrafted jewelry materials. The good ones are imported from Europe and last a lifetime. Use medium or coarse blades which don't break as easily as the fine blades. If you have good eyes and a steady hand, you now have the perfect tool to cut any cross section through anything from an inch thick pine board to sheet brass to large hunks of non-hardened steel with tolerable precision. This one tool comes pretty close to having a machine shop in its ability to shape a wide variety of materials with acceptable physical effort. Of course, this is not to criticize the abilities of a wide variety of hand tools and power tools and a lathe and milling machine. However, most of the time you won't even need the expensive tools to make basic mechanical gadgets if you are clever at reusing precision mechanical movements scrounged from junk consumer electronic equipment supplemented with odds and ends from a good hardware store.

Next is a mechanical prototyping technique which I have never heard advocated before but which I swear by from years of personal experience building every sort of instrument! Try making mechanical prototypes from a combination of window glass and silicone rubber. These two materials are a nearly perfect mechanical combination in terms of the fact that they allow extremely inexpensive construction of permanent and rigid mechanical arrangements of a wide variety of three dimensional shapes. Glass is cheap, extremely hard and rigid and chemically resistant but easily cut to give a flat and rigid surface. On the other hand silicone rubber bonds to clean glass perfectly when you squeeze it out of a tube. Furthermore, silicone is non-toxic and forms a flexible but water resistant permanent bond with glass. Silicone rubber sets up in an hour or so when exposed to air. Moreover, many complex silicone and glass arrangements will hold together enough until the rubber has set without clamping, due to the viscosity of the uncured rubber.

If you want to make a rigid right angle fixture to hold two components of some device, you simply cut two rectangular pieces of glass with a carbide wheel glass cutter and brace these planes at right angles using two right angle triangles cut from glass. You put a bead of silicone on the edges of the glass pieces and prop everything in position until it sets up. Of course, the rubber bond in itself is stretchy, but this rubber bond is relatively small in its dimensions compared to the glass. Moreover, the glass can be cut accurately or ground so that the rubber bond is arbitrarily thin, allowing the mechanical rigidity of the bonded combination to be improved as required.

I think I could often functionally best someone with a machine shop in building basic mechanical fixtures to hold working components together by merely cutting and assembling strategically positioned pieces of glass bonded together with silicone rubber. Nothing is as hard and cheap and rigid and chemically inert as glass nor anywhere near as easy to cut to shape as glass, while practically any complex shape can be assembled out of pieces. If you are dissatisfied with the result, you simply run a razor blade through the bond and reassemble it.

I haven't got around to it yet, but I have a vision of making a twelve inch telescope mirror blank by taking two circles cut from plate glass and bonding and bracing them together with a hexagonal honeycomb array of many precision ground little glass rectangles bonded between the glass discs. Composite mirrors have been tried before, and are mentioned in Ingall's books on amateur telescope making (they were said to debond eventually), but I think this idea might have been prematurely discarded before the uniquely permanent properties of glass and silicone rubber bonds was discovered.

I make a lot of my scientific instrument prototypes from wood or on wooden bases made from pine boards. You can mount electronics over to one end, often a six volt alkaline battery supply is convenient for such trials, and the mechanical element can be mounted on the other end of the board. I use silicone rubber as a glue since one can squeeze it out of the tube and leave it to cure and then break the bond and make mechanical revisions whenever necessary. Although not as good as glass, clean aluminum makes a fairly good bond with silicone as compared to most metals.

Wherever you need mechanical rigidity for small complex shapes, few things are as convenient and cost-effective as small pieces cut from sheet brass and brass tubing with a jewelers saw and then carefully soldered together with a handheld propane torch. Many hardware or hobby stores stock a selection of brass. Brazed joints are strong, but are rarely necessary for strength if the solder joints are made flat and thin and the shapes are braced. Mild steel solders well with zinc chloride flux, but one should work in a well-ventilated area and wash up afterwards because lead chloride is generated. When you need dimensional precision without a machine shop, it is helpful to assemble soldered brass subcomponents with metal bolts through oversized holes or slots that allow you to position everything before the bolts are tightened.

A lot of my prototype instrumentation makes extensive use of inexpensive electronic sensors and analog circuitry. There is nothing nearly so cost-effective and easy to use for prototyping circuits as the hobbyist plug-in boards like those they sell at Radio Shack, or better quality versions of the same thing. Getting a proposed circuit to work right is largely a process of trial and error. Being able to switch components in seconds at will until it works right is made vastly easier if a circuit is prototyped on such a board, at least initially.

Can such a board be made part of a permanent instrument? I think the answer in some cases is yes. It is mostly in the case of low resistance stability in the realm of milli-ohms or else high frequency circuitry that a simple friction plug-in connection is unsatisfactory. They usually seem to work reliably with moderate frequency CMOS circuits. If one wanted to make a plug-in board circuit more permanent, it would be advisable to mount such a circuit in a hermetic enclosure along with a desiccant like silica gel so the metal contacts will not oxidize and to use short trimmed leads.

Actually, a lot of the inaccuracy of sensitive and accurate instrumentation circuits comes from thermal non-homogeneity. Thus if a circuit is mounted in a metal box rather than in the open air, all the components will be nearly the same temperature and the overall output will tend to drift in a predictable way as a function of ambient temperature change. In fact, the use of a miniature constant temperature "oven" maintained slightly above the ambient, using a little resistance heating and temperature feedback circuit, is probably one of the cheapest and easiest ways to get very good accuracy and stability from a wide range of small physical and electronic sensors, at the expense of a few watts of extra power. These small electronic sensors, which are almost invariably temperature sensitive, are the key components that determine the final accuracy of instrumental measurements.

.