At 02:32 PM 6/23/2005, Ray J wrote:
So.... what are you saying? that this( sdr) is not a 100 watt radio when
used on ssb....
I don't get it... u say you use a scope and set the sdr peak power out to
the same as the other radio.. and then u say the output of the amp will be
different..... that makes no sense to me... of the power input is set "the
same" it should be the same out of the amp.. ?? am i missing something?
Ray
W9RAY
Measuring RF power on a SSB signal with a highly varying envelope is a
tricky business, especially if the radio does (or doesn't) any speech
processing.
The "real" limit on power for most solid state amps isn't really power, but
is the maximum voltage (limited by the Vdd or Vcc power supply). Over
drive and it clips the peaks, generating all sorts of distortion and
harmonics. So, if your signal has a high peak/average ratio, and you set
it up so the peaks are just shy of the clipping level, the average power
(what you'd read on a calorimetric power meter, or on a standard thruline
Bird type meter) will be a lot less than 100W.
The run of the mill "peak reading" meter is typically some sort of
diode with an RC time constant filter scheme, with a time constant chosen
so that "typical" voice patterns keep it near the true peak.
It's pretty easy to do some simple speech processing (e.g. an audio
compressor) which reduces the pk/average ratio, so now, the average meter
will read higher, with the rig still set up so the peaks just avoid
saturation. The other thing is that mechanical meters tend to read average
voltage (and are calibrated in watts, assuming a particular impedance), as
opposed to power. If the ratio between RMS and average is different than
for the waveform the meter is calibrated for, it will read differently.
You can also get different readings depending on the spectral content of
the speech signal. Imagine the run of the mill voltage probe into a diode
and RC time constant sort of circuit, with a TC around 100
milliseconds. Put in a 1% duty cycle pulsed RF carrier, pulsing at 10
kHz. Your meter will certainly read the peak value. The average value is
1% of that peak. Now change the pulse rate to 20 Hz, but keep the duty
cycle the same at 1%. The average power remains the same. The peak meter
will read less now, because the voltage on the RC has time to decay between
pulses. The mechanical inertia of the meter will make the needle hold steady.
(In the precision RF measurement world, you use a resistor for a load and
measure how much it heats up, giving you a true power measurement,
regardless of waveform, but you still have to deal with the sensor time
constant vs the source envelope problem.)
A voltage probe into an oscilloscope is a "true" peak meter, but, of
course, reads voltage, not power. You have all those peak vs average vs
rms issues to worry about.
So.. relating back to the SDR1000. In a vanilla implementation, there's
not much speech processing going on, so properly set up (i.e. max power
without saturating the PA), the average power will read low compared to
another 100W transmitter, that does do speech processing.
If you want to do an apples/apples comparison, you need something like a 2
or 3 tone test, where the beat frequency between the tones is high enough
that average power measurement isn't adversely affected.
James Lux, P.E.
Spacecraft Radio Frequency Subsystems Group
Flight Communications Systems Section
Jet Propulsion Laboratory, Mail Stop 161-213
4800 Oak Grove Drive
Pasadena CA 91109
tel: (818)354-2075
fax: (818)393-6875
