[Amps] SWR and amplifiers

Thu May 25 13:32:18 EDT 2017

Bill, and all,

this discussion is just too much fun to abstain from adding my own input 
to it!

So, here are my answers to your questions, and then an explanation:

> Let's say my SWR meter indicates 1000 watts forward and 100 watts
> reflected. 
> 
> 1. Does this mean the actual radiated power from the antenna is 900
> watts?

Yes.

> 2. How much power is the amp actually putting out - 1000 watts or just
> 900? 

900. But either into a load resistance that's not a precise 50 ohm, or 
into a combination of some resistance and either a capacitance or an 
inductance.

> 2. If the answer to #2 is only 900, why does the SWR meter indicate
> 1000? Is that a false reading caused by the actual 900 added to the
> 100 reflected? If so, why does reflected and actual power add to
> produce a false reading?

Because all usual RF power meters don't really measure true power!!!

Now the explanations:

Firstly, there is no such thing as one power flowing from transmitter to 
antenna, and at the same time some other power flowing back over the 
same line. This is just a model of thinking, extrapolated from applying 
short pulses to a transmission line. These short pulses can indeed flow 
forward, be reflected, and flow back, AT SEPARATE TIMES. But when a 
continuous RF signal is applied to the line, we can no longer really 
talk about power flowing forth and back simultaneously. It's just a 
single power flow, in one direction. What "comes back" is absorbed by, 
and integrated with, the power that flows forward.

Secondly: All forward/reflected power meters I have ever seen in my life 
  DO NOT measure power!!! The "forward power" indicated by a typical 
meter  is the vectorial (two-dimensional) sum of properly scaled samples 
of the voltage and the current on the line, and this is displayed on a 
meter scale (or digitally converted) so that if, and ONLY if, the load 
resistance is 50 ohm without any reactive component, the indication will 
be the real power dissipated in that 50 ohm load.

The "reflected power" in such a meter is simply the vectorial difference 
between the same two samples. Bird meters need the slug rotated by 180 
degrees, to change between "forward" and "reflected" indications. When 
rotating the slug, capacitive voltage pickup from the line remains the 
same, while inductive current pickup reverses in phase. This changes the 
measurement from the sum to the difference between the two samples.

Some numerical examples:

Take a meter connected to a clean, perfect 50 ohm load. Apply 200V of 
RF. 200V divided by 50 ohm causes a current flow of 4A, and thus a power 
of 800W. The current will be in the same phase as the voltage. The 
meter, having been designed for 50 ohm lines, divides the voltage sample 
50 times more than the current sample. Let's say the the voltage gets 
divided by 50, and the current is converted 1:1 to voltage; then we get 
a voltage sample of 4V, and a current sample that also transforms to 4V, 
and both are in phase, so the "forward" output is 8V, and the meter is 
calibrated to read 800W when getting 8V. Note that this is not a direct 
1:100 meter scale calibration, as power varies with the square of 
voltage. Twice the voltage is four times the power, and the meter scale 
is calibrated accordingly. 4V shows as 200W, 2V as 50W, and so on.

The "reflected" output is 4V-4V, thus zero. That means an SWR=1:1.

Now let's remove the load. We still have the same 200V, but the current 
is now zero. The "forward" output is 4+0 volt, the "reflected" one is 
4-0 volt, so both are 4V. Half as much as before, so one quarter of the 
power. So this meter will read 200W forward power and 200W reflected 
power, meaning an infinite SWR. The infinite SWR is indeed correct for 
the open load, but the 200W reading is a gross lie! In truth there is 
zero power.

A more sophisticated meter that displays SWR, and forward minus 
reflected power, would give correct readings: Zero power, infinite SWR. 
But such meters are exceedingly rare in the ham environment!

Now lets take the same meter, same 200V, but a load that's a pure 
inductor (or capacitor) having a 50 ohm reactance. The current will then 
be 4A again, but it will be 90 degrees out of phase with the voltage. 
The reactive power in the load will be 800W, but the true power will be 
zero. Let's see what the meter will do:
The "forward" side will add 4V at zero degrees to 4V at 90 degrees. The 
result of this is 5.66V at 45 degrees. The phase angle of the result 
gets lost in the detector diode, and the meter gets 5.66V, making it 
read 400W. Meanwhile the "reflected" side will take 4V at zero degrees 
and subtract 4V at 90 degrees. The result is 5.66V at -45 degrees, so 
the "reflected" meter will read 400W too. Again, the power indications 
are total nonsense, complete lies, while the infinite SWR displayed is 
correct, and a more sophisticated meter displaying forward minus 
reflected power would correctly display zero power.

Now let's take the same meter and 200V signal, but with a 100 ohm load. 
There will be 2A flowing, putting 400W into the load. But the "forward" 
side will add 4V to 2V, getting 6V and indicating 450W, while the 
"reflected" side will see 4V-2V=2V, so the meter will indicate 50W 
reflected power. If the SWR scale is correctly drawn, this will show the 
correct 1:2 SWR, and the net power (forward-reflected) is indeed the 
actual power going   into the load, but the 450W indication is a lie. 
There are no 450W anywhere in this system!

What you should take away from all this:

1) Reflected power is a mental construct, and is not real.

2) The forward power readings of usual meters are complete nonsense 
except when the SWR is exactly 1:1.

3) The SWR readings of any decent meters are usually correct, within 
their precision capabilities.

4) The forward power reading minus the reflected power reading is the 
actual power produced by the amplifier and delivered to the load.

5) Power meters that do this forward minus reflected calculation 
internally, and display just "power" and "SWR", are usually correct, 
within their precision limits. But not many ham-grade meters do that. 
It's more common in professional laboratory gear.

A power meter that does the calculation internally can easily be 
recognized by applying some significant RF voltage to its input while 
leaving the output open. It should indicate exactly zero power, no 
matter how much RF voltage you apply. And when shorting the output, you 
should be able to apply RF current to the input, with the meter reading 
zero power.

Manfred

========================
Visit my hobby homepage!
http://ludens.cl
========================