One of my first jobs in QC was to look over some boards for an adjustable solid-state grid-pulser I started hooking up the boards and trying to get it going. Most of it worked, not the pulse. We blew some transistors. We were plucking them off of a big board with a bunch of them stacked. It was a 1,700 V NPN by Telefunken, the BU 209, surprisingly advanced for the day. Chuck boxed it up and got it finished on the fly. The problem turned out to be the push and pull were driving "on" simultaneously.
That circuit went to drive the grids of 3ea. 35 KW tetrodes in paralell, for 50 KV 60 Amp Peak, 100 KW Average loading. The customer also gave us three spare tubes and sockets to do our own load. The tubes are under oil, which is pumped through a heat exchanger for cooling. The power supply was a tetrode test set. John McAmmon at Litton vouched for us to the Navy. But that thing would't ship. John was embarrased. Then the program was over. The Navy wanted us to add another cabinet, loaded up to support the electrodes of two high-peak current TWT's. But they couldn't order anything from us again until they'd received the first unit.
So we shipped it shot--to beat the first-article-clause. They sent someone to accept it. The guy took it without hint of impropriety, it was amazing. Jimmy, running the customer acceptance test, comes and borrows a pulse-generator for a procedure in the test. He hands it back. It's blown. He didn't know he blew it out though. He pushed the button. A single shot successfully emerged. Crowbar firing executed flawlessly. End of story. Uh, er, time to buy more pulse generators. That unit had all the gags. It still had straight capacitive discharge breakdown internally bypassing resistors under the oil. Getting those resistors into the circuit made it a different machine; it sounded nicer too.
The Crane 50 K shook out to be a good unit. It repetitively fires a triggered vacuum gap crowbar. It dogged Chuck, because he was in a tizzy. The things that went wrong were mechanical. It has a two-terminal reversable polarity 50 KV 2 ADC 60 APk supply. There is an oil tank full of ohmweave resistors, strappable for various values, with forced oil cooling. It has an automatic foil tester which is pretty neat: with foil running through a regular chart recorder, in a front panel. A "foil-test" button cycles an arcdown by advancing a 50-KV-potential corona-ball up to the back of the grounded foil. Action is damped by a glass dashpot. An "advance-foil" button unfurls the test specimen. However nice, at one point It turned out the ball was arcing to something else back there. That's how it passed the ATP. A ground braid was tucked under a rail where you couldn't see it. As the ball swung forward, blammo. It sounded loud but sure was soft--didn't mark the foil at all. The flash was right behind the foil. It was totally convincing. Jim Zeranski figured it out after the ATP.
When I first got to ETM, Chuck was hot-sticking the 50 K. I was busy with a cookie-cutter. Raunchy splashing-metal--"sputditty-pop" sounds came from the cabinet with each arc. He'd call out the voltage before each arc. Those sounds begged for investigation. I thought I'd try to get a glimpse from the rear. I got back there okay. Creeping up, the four chain-linked Powerstats hunted and jerked, sounding like a chain gang, then silent. He called out the voltage and that thing issued the raunchiest blast. I was duped dumbfounded. Oil rushed at my feet. Mist rolled out and dropped. It didn't make my ears ring, not for lack of stored energy. It was more like the thud you feel in your chest, from a bass drum. The crowbar tank had ruptured. It looked like a 12 gauge magnum slug had hit it on the inside. And that tank had forced-oil-cooling too. It was a mess.
Chuck deduced graphite-ceramic shaft-seals were incompatible with transformer oil. So he switched to teflon/fiberglass seals. Crowbar tank dimensions were increased to fill alotted space. And that was the last really bad report the machine issued. I went through it later and it wasn't so bad. It blasted periodically through the shakedown process. And I had that crowbar tank out too many times for my taste. But I came to like the machine. Later I modeled some aspects of my own repetitive firing crowbar after this one. The Hiltek transformers were so large they did fine. Ample isolation was a given. |
