[AMPS] Tetrode Amp articles needed

[w8jitom@postoffice.worldnet.att.net]

Hi Ian,

> W8JI wrote:
> >In the practical world, ALC is actually MORE necessary in a class AB1 
> >PA. The slightest grid current will send the PA into distortion 
> >quickly with most AB1 grid driven designs.
> >
> Very true - but how much of this is due to the tube characteristics, and
> how much to the power supply? Many power supplies designed for AB1 have
> very high output impedances (several K or more), so the flow of grid
> current shifts the operating point dramatically. In such cases you can
> expect distortion.

Absolutely correct. If the PA was built like an AB2 PA, even 
though operated AB1, problems would be greatly reduced. Unfortunately 
MANY published designs use a high value grid source resistance (and 
unregulated screen supplies to boot). 
> 
> Ironically, much of this additional impedance comes from the circuits
> intended to generate enough ALC voltage when grid current begins to
> flow. This turns the need for ALC into a self-fulfilling prophesy.

I again agree. Collins does this, but in that case the system is 
justified because the tube they use is easily destroyed by ANY amount 
of grid current.

> >The correct way to install ALC in an AB1 PA is to sample PEAK RF grid 
> >voltage, and activate the ALC as peak positive voltage approaches the 
> >quantity of negative bias voltage. 
> 
> Yes, but even then it's a mistake to rely too heavily on ALC, because
> there will always be overshoots.

Absolutely, but with a long time constant the overshoots will be 
infrequent. Using Rich's suggestion of setting the bias and 
operating conditions properly PLUS ALC would be the most worthwhile 
improvement.

> Like Rich, I tend to run my own tetrode PAs open-loop, with no overall
> ALC feedback, and instead use speech clipping and careful setting-up of
> the whole system to control the drive level. On the other hand I also
> sympathize with Tom as a designer of amplifiers that could be used by
> any operator with any driver, and therefore has to provide ALC.

That's correct.

But you, Ian,  know what you are doing and always keep 
things adjusted. In a contest station (or with guest ops), in the 
frenzy of a pile up, if the exciter changes, or if you change 
processing, ALC can be very helpful to people sharing the band with 
you. Bump the drive control or processor level, or move from 160 
to ten meters, and you have no safety without ALC.

While it should never be depended upon, neither should Rich's method 
of setting bias. Virtually ALL exciters have overshoot on leading 
edges, and change output power with load impedance or frequency 
changes. Unless the operator wants to spend his life adjusting bias 
and screen voltages, ALC would be a good additional safety feature.

I'd like to add this thought. I think a PEAK grid voltage warning 
indicator makes great sense in a grid driven linear PA. A simple 
LED driven by a sample system with hang time on the warning 
coupled with slow decay ALC would be the best idea. Then the 
operator would get a warning while the ALC attempted to limit the 
problem.   
  
Watching a grid meter is NOT good enough.
73, Tom W8JI 

--
FAQ on WWW:               http://www.contesting.com/ampfaq.html
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm