[Amps] grid fuses

[R L Measures]

On Jul 19, 2006, at 7:29 PM, Paul Christensen wrote:

>> I joined this list to learn about amp design, not repair.
>
> It's a tough neighborhood here -- no doubt about it.  However, I have
> probably learnd as much about amp design from the hostile  
> discussions as
> much as the friendly ones.  Make no mistake: I am not advocating  
> hostility.
> I'm simply stating a peronal observation.
>
> I can see the
>> point about not designing a fuse or limit resister in the grid  
>> circuit
>> because it seems to be a band-aid for poor fault design.  How do you
>> justify
>> designing it in?
>
> I don't know whether or not grid fuses are appropriate in some amp  
> designs,
> but I will offer this answer to your question:  Sometimes, the  
> least complex
> solution is the best compromise.

Bingo, Paul

> While electronic grid protection may offer
> the quickest response to a grid fault condition, the designer must  
> design
> the circuit in such a way that its failure causes no additional  
> harm to the
> amplifier above and beyond that caused by the underlying fault  
> condition.
> For example, if a switching transistor is used for grid protection,  
> do we
> have assurance from the designer that a collector-emitter short  
> will not
> cause additional harm to the amp?

... next question: does a C-E shorted transistor limit current?

> Is there even the remotest chance that
> the circuit can fail and not provide 100% assurance of grid  
> protection?

I've seen the grid protection pass-transistor short in an Alpha 3,  
8874 amplifier, and one of the 8874s had gold sputtering leakage.    
Dick Ehrhorn once told me that he had over a hundred gold-sputtered  
Eimac tubes at his factory.  Previously, he told me that no Alpha  
amplifier has ever had a parasitic oscillation.  When I quoted and  
glued these two statements together, he said I was insane.
>
> Some good electronic grid protection circuits do exist.  But in  
> terms of the
> complexity/effectiveness ratio, I still ask this question: are most  
> grid
> failures the result of instant, excessive Ig, or sustained Ig?

IME, instant. With only 100J to 300J stored in the filter-C, a gold  
boiling episode probably could not last longer than a few mSec.
By its very nature -- i. e., skin-effect at high frequency, gold  
sputtering would have to take place only on the surface since  
bringing the entire grid up to the boiling point of gold would take  
Much more energy than is available.

> If a fused
> grid protection circuit can be made effective in the design, is it  
> necessary
> to add additional cost, complexity, and additional potential points of
> failure in the system?

Not unless one is trying to impress ignorant buyers.
>
> By analogy, we could employ electronic current sensing on all our home
> appliances.  When AC current rises above a safe level, the AC  
> circuit could
> be made to immediately open rather than depend on a fuse or circuit  
> breaker
> to blow or pop.   Notwithstanding UL and code issues, if a cruder  
> method
> (the fuse) is just as effective, less complex and less costly, is  
> it right
> to condemn the designers of all home appliances because they chose  
> to use
> fuses rather than electronic current-sensing switches?
>
> Paul, W9AC
>
>

R L MEASURES, AG6K. 805-386-3734
r at somis.org