Process R-D and one piece flow

This week besides being buried in numerical analysis was a time of process R-D. One of the things I worked on in my plasma metrology days was one piece manufacturing flow. At the time, we were building tons and tons of assemblies, and in the past ran in bath mode. Eg, cut thousands of pieces of metal, weld thousands of joints, assemble thousands of units, test thousands of units, life cycle and thermal screen units, and finally calibrate them. The problem with batch mode, is anytime you have a process error, you risk ruining thousands of units, and totally hose up the process. We had SPC charts, and bureaucratic procedures left and right to reduce and hopefully eliminate any problems. Six sigma was just making inroeads at that time. Most of the time, such procedures worked pretty well, but they are not cost effective. Everytime you add a step, you add labor cost, and bump out leadtimes.

With one piece flow, the idea is to set up processes such that 1 unit or a small quantity of units can be assembled with the same or greater efficiency as 1000 piece batches, or in some cases 50,000 piece batches. This is where life becomes incredibly complex, and tradeoffs are made. In electronics manufacturing, instead of running a 27000 piece build over a few day period, maybe one might drop the batch size down significantly and build to completion, rather than to subassembly and by ditching much of the BS achieve simliar efficiency.

So the problem I am trying to tackle is bumping up efficiency in microbatches. While I would like to be able to crank out 1 unit at the same efficiency I might hit while running 100, I know thats probably impossible with the equipment at hand. Its odd, but our new product line has over 74 products, some of which are going to be decent volume, vs others might only be sales of a few here or there. As a result, the game is to get the efficiency up, so we can sell the incremental products, like the 16 channel thermocouple adaptor, or the high res photodiode amplifier at a reasonable price. Who knows, maybe we will even do a lock in amplifier module, or NMR spectrometer module if we can get our efficiencies in line.

Hey, anyonw want a NMR (nuclear magentic resonance module for $100!!!) and no its not for sale, nor even on the planning board yet, but if 1000 people want one, it just might happen.