[Amps] Linearity metering

[Ian White, G3SEK]

Jason Buchanan wrote:
>
>I noticed in the 1972/1973 ARRL handbooks that I have there is a 
>4-1000A amplifier project with a linearity metering circuit.  Look for 
>the SS-2000 project.  This seems like an incredibly good idea, however 
>I have never seen anything like this in any other amplifier except for 
>something similar apparently in the ACOM 1000A (and presumably the 
>2000).  It seems to me that the ability to tune for optimum linearity 
>would be a "Good Thing"(tm) and the circuit design appeard to be simple 
>to implement.
>
>Was the original designer of this circuit blowing smoke or is this 
>something worthwhile to add to a homebrew amp?

I don't have that particular ARRL Handbook, but the circuit is possibly 
something like Collins use(d) as combined tuning/loading/linearity 
meter. It is described in 'Single Sideband Principles & Circuits' by 
Pappenfus and his Collins colleagues.

It's simply a pair of diode detectors. One looks at a sample of the 
cathode  RF voltage (typically at the in a GG amp) and gives a negative 
DC output. The other looks at a sample of the plate RF voltage, and 
gives a positive DC output.

These two outputs are combined through a DC balance pot, and fed to a 
center-zero meter. When the amp is correctly tuned and loaded, this 
internal pot is adjusted so the meter reads exactly zero. For improved 
linearity, you would probably load on the heavy side from the point of 
maximum RF output, in the usual way. Whatever you decide is "correct", 
you set the balance pot to read zero under those conditions.

To use the meter for pi-tank tuning, simply tune for maximum rightward 
(positive) deflection.

As a loading indicator, the meter is actually measuring the *gain* of 
the amplifier - it reads zero when the amp is loaded to give the correct 
level of gain. Too much gain (too light loading) makes the meter move 
right; too little gain (too heavy loading) makes the meter move left.
Any single-meter (or single bargraph) indicator probably has to be of 
this 'Collins' type. To use it, simply tune for max, load for center 
zero, and repeat those two steps until you're exactly there.

With normal modulation, the needle should stay pretty much on zero. If 
the amp is overdriven into flat-topping, the gain will drop and the 
needle will move leftward, so it acts as a simple linearity meter.

It's obviously pretty easy to automate the basic tuneup procedure in a 
computer-controlled amp. More advanced auto-tune amps use a separate 
phase detector to identify the correct output tuning point, which is 
0deg phase shift between input and output (for a GG amp).

Hmm... my amp has an unused position on the meter switch... thanks for 
making me think about this, Jason!


-- 
73 from Ian G3SEK         'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.co.uk/g3sek