[ARC5] Interesting occurrence...

[mac]

Yep, the 5963 (for ex.) is just such a premium version of the 12AU7.   
One of my first projects as a new engineer back in the 50's included  
hundreds of 5963s in "flip-flop" and other switching circuits.  Only  
failure mode I remember with these tubes were a (very) few instances  
of heater to cathode leakage.

Dennis D. W7QHO
Glendale, CA

*********

On Apr 2, 2011, at 7:54 AM, Dennis Monticelli wrote:

> Regarding leaving on filament without applied B+, there are some  
> lessons
> from early vacuum tube computer circuits.  In binary logic  
> operations a tube
> could spend a great deal of time sitting at either an ON or OFF  
> state.  It
> was found that tubes were failing bofore their expected MTBF and that
> problelm was traced to tubes that had been in the OFF state for an  
> extended
> period.  This premature failure was nick-named "sleeping sickness"  
> but is
> more correctly termed "cathode interface resistance."  The mechanism  
> is the
> formation of an interface layer between the nickel cylinder (in an
> indirectly heated cathode) and the active oxide coating.  The series
> resistance of this layer reduces the gm.  The work-around is to run  
> a little
> plate current during the "idle" state or better yet use a tube  
> purpose built
> for the job.  There are tubes in the premium industrial series of  
> minatures
> that are specifically rated for computer use.  These types are  
> virtually
> immume to the formation of the problematic interface layer.
>
> In view of this known problem it wouldn't seem to be a good idea to  
> run
> indirectly heated filaments over extended time periods without some  
> trickle
> plate current.