by erichayes » Tue Feb 12, 2008 12:07 am
Hi Hembrook,
No offense taken at your post--a legitimate point-- and none was intended with my rejoinder.
I preface my admittedly anecdotal "proof" with a couple of statements that most of the folks on this forum who've been around for a while have figured out.
First, I have low regard for the wine-and-cheese crowd who've invaded the Hi Fi world in the last couple of decades. Unidirectional speaker cables and gold plated contacts don't even leave home plate, let alone get to first base with me. I build, along with Bud Wyatt of (among others) Sheffield Labs infamy, power amplifiers for recording studios, so snake oil doesn't impress me.
Second, this "experience" occurred in 1960 or '61, when nobody but the military gave a damn.
My electronics mentor was a UCB college buddy of my folks in the early '40s. He wound up in the Army signal corps during the war (WWII) and later worked for the Navy at Hunter's Point in San Francisco as an electronics maintenance specialist.
One weekend, when he came to visit, he saw that I'd inherited a Heathkit variable power supply with a bad PT. The pass tubes were 1619s, which were was a directly heated 2.5 volt version of a 6L6 used prolifically in signal corps equipment. This prompted a monologue on the failure rate of the 1619 and his part in the analysis of same.
Baldy (his surname was Balderson) told me that there was an inexplicably high failure rate of 1619s in mobile equipment, as opposed to stationary gear. His job was to find out why.
He (and his coworkers) found that, almost without exception, the failures were caused by the negative-connected end of the filament breaking. Their conclusion was that, due to the relatively unregulated voltages available to mobile equipment, and the fact that the tube filament was designed to operate on AC, the filament suffered from a byproduct of the Edison Effect.
Being the Baby Boomer version of Soren, I asked him if this was why directly heated tubes intended to be operated only on DC had polarity stated on the basing diagrams. He said "probably", but there were other factors to be taken into consideration, such as which leg the suppressor grid was connected to in pentodes, or where the diode plate(s) was located in a multifunction tube. I guess my question intrigued him, because he told me on a later visit that he'd talked with an RCA engineer and was told that they tapered the filaments on those tubes, with the "fat" end of the filament being connected to the negative pin.
I don't read other tube fora anymore; I got tired and disgusted with the "facts" that are propagated, spewed and flatulated on them. If someone will back up the phenomenon here, or on another forum, I, too, would appreciate some corroboration. All I have is a story that's almost fifty years old, told by someone who didn't really give a rat's ass about tube life. My guess as to why it wasn't publicized is that the JAN tube manufacturers were (duh) in the business of making tubes, the military wanted same--regardless of failure rate--and juicy contracts would have been jeopardized if someone had blown the whistle.
Sound familiar?
My point is, if you're going to invest a considerable amount of money in a pair of 2A3s or 300Bs, think about tube longevity vs hum problems that can be addressed by means other than running the filaments on DC.
Eric in the Jefferson State