More top secret information declassified now.
The sensors were dual "hot wire anemometers" that had to heated elements positioned next to each other in an airflow from two nozzles and air exiting or entering a bottle due to ascending or descending would flow through these nozzles. The wire in the air stream shadow of the other wire would not cool as much as the wire facing the air stream and the balance of the bridge would shift. Thus the wire changing the most would indicate whether air was entering (descending) or leaving (rising) the insulated bottle.
The sensor elements were the filaments of tiny lamp bulbs. I developed a system where I could break the bulbs to expose the filaments to ready the lamps for construction of the sensors. If the manufacturer of the lamps didn't make a gas tight crimp between the wire and the filament then the tungsten filament would oxidize over period of time and the contact would fail. This was the main cause of "infant mortality" in new varios. In order to weed out as much of this problem as I could, in a reasonable amount of time, I would bake the newly exposed bulbs in a low temperature oven overnight and then they would sit for a couple of weeks. When I was ready to start using them, I would run each one through a test system. Not only did it identify bad parts but I also paired them, within certain parameters, to use in the sensor assemblies. There were times when I would have to trow out about half of the bulbs. Obviously, over time, I determined the most reliable manufactures to purchase from.
One German lamp manufacture claimed to weld the filament to the wire posts. Those were very expensive lamps but I figured that if I didn't have to trash any I would be ahead in the long run and fewer varios would experience failure. It turned out that I had a very high failure rate when testing those lamps. When I told the sales rep that the connections were not getting welded as claimed, the factory replied that was impossible because it was done automatically by machine and therefore could not be in error (hmm.... sounds typical German design philosphy doesn't it?). The factory didn't even check to see if there might be something to what i was claiming. Just blind faith in their automatic process to work correctly. Of course I quit buying those lamps.
Frank Colver
- Filament Test System.JPG (536.48 KiB) Viewed 7376 times
- Filament Test Current.JPG (345.15 KiB) Viewed 7376 times