Geoff, I looked at the 250TH specs, but I really can't make a solid judgement on acceptable filament voltage at reduced specs for long life. Generally I would be conservative; where you might reduce filament voltage a great deal, I recommed that you only reduce it slightly.
Emission falls off very rapidly with reduction in filament voltage, and a small reduction probably has a large effect on tube life. I really can't specify how much you can safely reduce filament voltage, but with thoriated-tungsten filamentary cathode tubes, I would recommend that you don't reduce operating filament voltage more than 10% from spec. And make sure your socket connections are clean, etc. Ref: http://www.bext.com/filament.htm Ref: http://groups.google.com/groups?q=%22reduced+filament+voltage%22&hl=en&lr=&ie=UTF-8&safe=off&selm=3kp9s8%24o68%40indy-backup.indy.net&rnum=9 Some recommend not reducing filament voltage on large oxide-cathode tubes such as the 8877, etc. Ref: http://www.w8ji.com/vacuum_tubes_and_vaccum_tube_failures.htm Others claim remarkable improvements from filament voltage reduction. Ref: http://www.broadcast.harris.com/ support/kb/supportdocs.asp?fid=1050 (Note: I believe the 4CX20000 is an oxide-cathode tube.) Other factors probably have as much effect on tube life. Excess current, of course, or operating before full emission temperature is reached, and thermal shock at turn-on. ref: http://lists.contesting.com/archives/html/Amps/2001-06/msg00213.html Burn-in: cook long-unused tubes at rated filament power for 100-200 hours. The idea is to allow the "getter" elements to absorb gas that has leaked into the tube or out of the materials in the tube. In tube operation, gas becomes ions which attack the cathode. Ref: http://www.rfparts.com/tubeapp.html Filament and cathode structures do not like thermal shock. The sudden application of filament voltage to a cold filament is a significant shock, because the resistance of a cold filament is much lower than its resistance when it is hot. A gradual application of filament power is best. This can be done with a resistor in series with the filament transformer primary, a small filament transformer that overloads at turn-on, etc. Filament voltage regulation is a good idea. Ref: http://www.vt52.com/diy/tips/tt_filaments.pdf But, be careful not to make the turn-on shock situation worse. Basically, I suggest one or two big constant-voltage transformers and a few small filament transformers, possibly with a series resistance of a few ohms in their primary circuits. The little transformers will overload on turn-on, limiting thermal shock to the tubes, yet the system will maintain operating filament voltage nicely, over a wide range of mains supply voltage. Then you can tweak the filament voltage to extend emission life, and the voltages will stay put, and turn on will be gentle. The gentler, the better. Bacon, WA3WDR
