29
Appendix A Cont’d
Peak Power Readings:
How should peak readings compare with steady state readings? Because of the speed and response of the LP-100A, this is a loaded question, as you
will see by reading on.
First of all, here’s a little background on digital wattmeters. Almost all digital wattmeters use a diode peak detector to detect power. There are a few
which use log detectors such as the Analog Devices AD8307. I know of none which use a true RMS detector, although Analog makes several of these
chips. The Analog chips have advantages, but the biggest problem with them is accuracy, generally stated at about 0.3dB to 0.5 dB (6-10%). This error
could be calibrated out, of course, but to do so at all power levels would be expensive, and not amenable to user adjustment.
A peak diode detector rectifies the RF envelope and charges a small filter cap to the peak voltage level of the applied signal. The response time of the
detector is determined by the time constant of the cap and load resistance. In the case of the LP-100A
, this time is very small. To obtain a true “average”
reading requires taking many samples over a period of time and calculating the arithmetic mean. This slows down the response of the meter to power
changes in the average mode, of course. In the case of the LP-100A, the sampling is adjustable, and can reach levels as high as 40,000
samples/second in the Peak-to-Average mode.
Getting back to the original question, there are several
things that affect peak readings…
Where on the envelope the sample hits
The response time of the rigs ALC
The power supply regulation of the rig/amplifier
These items can all result in higher readings than steady state. All ALC circuits have overshoot… meaning that they can’t respond instantly to set the
requested power output level, especially at the leading edge of a CW character or voice syllable. This behavior also depends on the resting time
between characters. The effect is exaggerated when the SWR
is high, since the rig will try to back down the power due to high SWR, but can’t respond
quickly enough. Because the meter reads and holds the “highest” peak, it will usually be a value greater than the steady state reading for this reason.
The same effect happens with a linear amplifier whose power supply is usually not regulated. There is more voltage available at the onset of power
delivery than there is once the supply has stabilized. This is because the no load voltage will soar to a higher voltage than the voltage under load.
These phenomena only last for milliseconds, but the LP-100A will catch and display them until the next peak hold counter reset, which results in you
seeing the transient peak power being delivered. It is up to the user to determine if this is important to you in the tuning of your amplifier.
Here are some examples of overshoot
The first picture is my TS-480S in CW mode. This is in the middle of a string of dits. The effect is exaggerated if the character is the first one after a long
pause. The second picture is from Jack, K8ZOA’s Elecraft K2 and shows a tone burst in SSB, with the input signal on top and the RF output envelope on
the bottom. It is taken from a digital scope, and illustrates the effect after a pause
. Note: Since these are voltage displays, you would have to square
them to get the effect on power overshoot
. In Jack’s photo, you can see that the voltage overshoot is about 12.5%, which represents a power overshoot
of about 26% (1.125*1.125=1.265). This is the value the LP-100A would grab and display if the sampling window caught the leading edge. Sampling is
done at close to 100 samples/second, and the detector time constant is fast enough, that there is a good chance that peaks like this will periodically be
displayed. This is normal and represents an actual event.
Summary of Contents for LP-100A
Page 18: ...18 Schematic Page 1...
