Printed Circuit Boards

roddie wrote:I obtained some small boards from work which are totally flat with only "tracer-circuitry" running through them. The board-thickness is .067" (1.7mm).

I'd read somewhere that this type of stock can be used for C/L bellcranks. It's very strong/rigid material. I'm not sure how easy it is to cut through this stuff.. but it would be interesting to trace a bellcrank profile onto it.. and try using a scroll-saw to at least rough it out. A bearing is easily made from alloy-tube and some flat-washers. Have any of you heard about or tried this?

If it is a true fiberglass board it ought to work. If it is a material such as phenolic, then material is weaker and probably not worth trying. Give it a shot, tug, bend. If you have a CL model with bellcrank exposed, would be easy to test.

Back when I worked at McDonnell Douglas in the 1980's, we'd send minicomputer boards in for repair. They were multilayer trace (wiring levels embedded within the layers). If broken trace could not be jumpered, repairer would locate broken trace, mill down to the trace and repair it. Simpler boards might have 4 internal traces, but I heard as many as 10 layers. Due to expense of manufacturing multi-trace level boards, you'd only see this on more expensive equipment such as mini-computers, mainframes, aircraft avionics, etc.

Cutting fiberglass FR4 circuit circuit boards is hard on blades. If you can get the round carbide blades for the scroll saw they might last longer. We have circuit boards at work that are close to 30 layers with blind and buried vias that are close to 1/4 inch thick. These cost $8000 before any parts are put on.

Jason_WI wrote:Cutting fiberglass FR4 circuit circuit boards is hard on blades. If you can get the round carbide blades for the scroll saw they might last longer. We have circuit boards at work that are close to 30 layers with blind and buried vias that are close to 1/4 inch thick. These cost $8000 before any parts are put on.

Jason, thanks for the update. When I worked in the scientific computer aerospace industry 35 years ago, real men programmed in FORTRAN (and assembly) and quiche eaters wrote in PASCAL.

inf.ed.ac.uk: Real Programmers Don't Use PASCAL

We were still dealing with much stuff made in low and medium scale integrated circuits with a lot of discrete components. When I arrived at McDonnell Douglas in 1982, I remembered an old SEL hybrid analogue data acquisition computer for aircraft structural testing. It had 8 kb of ferrite bead core memory that the engineers and techs had upgraded with another bank of 8 or 16 kb static ram. It had just been removed from service (took up about 8 wall locker sized double cabinets). John, one of the techs would enter the boot loader via a line of binary toggle switches on the front panel, which then loaded the OS and application software from paper tape.

Our laboratory data computer was a then new Perkin Elmer 3240. I think it was 4 mips, with two 300 mb mulitplaten disk packs and 3 "high speed" 1600 bpi drives. It had a total of 16 mb dynamic ram. Each 4 mb board was about 2 feet square populated with 2kx1 or 4kx1 IC ram chips. High speed communication was via 1200 Baud modem. For tests we used smaller P-E 3220's (1 mip machine) and 3230's (2 mips I think) with 1 to 2 mb ram. 3230 was used for more demanding tests and had a higher capacity 30 mb hard disk the size of 2 shoe boxes.

Regarding cutting PC boards, I'd say yes if they are that thick. However, we're only talking material that is about 1/16" thick and small quantities, which a fine hacksaw blade (suitable to cutting EMT) would be able to cut.