|
The telephony signal report
(published in Electron # 11, 2007)
Introduction
Can somebody give me a quick report? This
question is often heard on the HF amateur radio-frequencies. I also am into
the 'habit' of issuing a reception report as a first exchange of information,
since an important part of a radio QSO is about gaining insight into the
performance of our radio station and propagation conditions. Reporting is
being performed according to the internationally accepted R S T code, with T
for Tone being omitted in radio telephony.
Readability
The first number in the RST code is the R for readability and is
ranging form 1 to 5, with 1 for Unreadable, 2 for
Barely Readable, occasional words distinguishable, 3 for Readable with
considerable difficulty, 4 for Readable with practically no difficulty and 5
for Perfectly Readable.
At many amateur stations this is watered down
to 5 for a well readable signal, 4 for a more difficult connection and 3 for
practically unreadable. At the last report the QSO will be quickly ended
because of the required additional effort.
It is a pity that no accepted code is
available for telephony modulation quality, especially since this is
determining to a large extend the success of a connection. Occasionally a few
words are added to the RST-report to specifically address this issue. In the
era of the radio valve, remarks about an audible hum could be heard. In our
days it usually is the frequency response curve that is playing havoc. Modern
transceivers currently are equipped with so many possibilities to 'play' with
the microphone signal, that it usually is far more easy to find a non effective setting.
Furthermore it is technically not very
difficult to compress the audio signal to enhance the average power output in
SSB-mode, to show high (and so 'nice') average figures on the output meter.
Since a feature is always a selling argument, most transceivers are equipped
with such an audio-compressor, usually
referred to as the 'processor'.
In the article about a digital speech processor to
enhance noisy speech signals at the reception-site, the reverse phenomenon
was exploited. To distinguish speech from noise signals, the signal envelope
is being exploited in this device, since one of the specific characteristics
of human speech is in the repetition frequency of utterances in the 5 - 10 Hz area. This distinguishing factor
also appears to be exploited in the human hearing system.
This has also been recognised
in the middle of last century when modulation schemes have been investigated
and patented that transmitted the information about the audio-signal envelope
and frequency content through different channels to enhance audio quality.
Bottom line is the importance of the envelope
signal in the comprehension of human speech. It therefore is somewhat peculiar
that audio compressors are popular at radio-amateur stations, since these
devices are designed to diminish or even remove this important envelope
information.
An other peculiar habit of some radio
stations is to go for 'studio-quality' in the microphone signal. Next to
expensive microphones, speech processing is added to fabricate or enhance a
deep and booming voice sound. This habit especially may be found in countries
with a strong masculine culture.
This again is counter
productive when trying to communicate in a noisy environment, when the
number of successfully transmitted speech syllables is far more important
than the speech quality. It often is overlooked that due to the particulars
of human hearing, much more energy is needed to perceive frequencies below
say 500 Hz., that carry little or no message information. To spend much of
the legally limited HF energy in the non-information frequency range
therefore is not very effective.
An efficient amateur station is utilizing a
simple but adequate microphone (e.g. electret type), followed by an equalizer
to enhance the speech frequency range that is 'carrying the message signals',
500 - 1500 Hz range, rather than the 'dominance signals', at below 250 Hz.
Setting of this equalizer is depending on the voice characteristics of the
operator and should be tested in practice.
My personal station is equipped with such a
simple but very effective mechanism, that often is yielding a full
readability report with the S-meter at the other side hardly showing above
the noise. More on this subject in the article on the 'Hambus
'.
Reporting on signal strength
The second
figure in the RST code is about the perceived signal strength. The figure '9'
in this series is well known and almost started a live of its own by
sometimes meaning no more than 'a good signal'. The meaning of the S-series
has been discussed for a great many years in local and international fora but it seems difficult to reach agreement between
the amateur and professional world.
Sometimes tables
are being published connecting the signal strength at the antenna input of
the receiver to the reading at the S-meter. Up to S-9 the formula often used
is:
meter reading in
S-units = 9 + (20 / 6) log(μV/50) and above
S-9, the formula
decibels above
S-9 = 20 log(μV/50). From these formulas it is
clear why the signal strength of 50 μV at the
antenna input for a meter reading of S-9 is well known by most
radio-amateurs.
In many books on
general knowledge for radio-amateurs these signal strength tables may still
be found starting with S-9 for an antenna input of 50 μV.
going down by one point each time this input signal is halved, related tot this logarithmic step of 6 dB.
More often these tables have nowadays been replaced by a more descriptive
table, with S-9 for 'Extremely strong
signals', S-8 for 'Strong signals', S-7 for 'Moderately strong signals', S-6
for 'Good signals', S-5 for 'Fairly good signals', S-4 for 'Fair signals',
S-3 for 'Weak signals', S2 for 'Very weak signals' and S-1 for 'Faint
signals, barely perceptible'. For signals above S-9, the old decibel
description is used again.
This subjective
scale is depicting the impression the signal is making at the receiving party
and is unrelated to any field strength number. The decibel scale above S-9 is
a measure of the deeply ingrained use of the S-meter and its perceived value
as a measuring instrument. This descriptive scale once more underlines the
above mentioned difficulty to reach agreement between the various parties 'at
the table'.
Measuring the signal
Lets take a look into the way the S-meter reading is
accomplished at the average amateur receiver.
The S-meter
usually is showing a reading that is related to the voltage at the common
rail in the receiver, that is setting the automatic gain control, AGC. This
voltage is related in turn to the incoming signal and the total gain in the
receiver chain, as determined by the combined effect of control
characteristics of each individual RF and IF stage.
At no input
signal, this AGC control voltage usually is at zero Volt. This is the first
problem to solve since the logarithmic S-meter is showing one S-point down
for every time the input signal is halved. The reading of S-0 therefore is
showing that the incoming signal is half the value of the signal that was
received at a level of S-1, which is not the same as 'no voltage at the AGC
line'. A correct S-meter therefore should always show some reading, however
low.
An incoming
signal at a certain level is bringing the AGC voltage up to keep the output
of the receiver within a reasonable range. This may be better understood when
realizing these incoming signals easily vary over several orders of
magnitude, making 'unprotected' listing a tiring, if not deafening,
experience.
The actual level
of the AGC voltage is determined by all amplifiers in the receiver chain,
while the relation between this voltage and the individual gain per stage may
be quite different. Usually the AGC control voltage is setting the amplifier
gain by controlling the base or gate of a transistor to set the gain
determining current of that stage. The relation between this voltage and the
current will be according to some power law, with different laws for junction
transistors, field-effect transistor and again different for tubes. At one
stage in radio history special radio-tubes have been manufactured to obtain a
long sliding characteristic for this gain-setting law, the so called tail
tubes, especially for usage in AGC stages. Since all the active devices are
acting according to some power law, the control voltage is behaving as the
logarithm of the incoming signal, which is the kind of behavior we also like
to see at the S-meter.
Unfortunately,
at a certain level of the AGC control level, one gain stage is acting more
active than the next and also at high levels, one gain stage is earlier at
end-of-range than the next. The relation between the voltage at the AGC
control line and the incoming signal therefore is different per receiver
design and will also differ (slightly) for each individual receiver within
the same series. Further more, the slope of the
voltage control curve is not constant either and will differ depending on the
actual level of the incoming signal.
Meter deflection
Lets look at how the designer is solving these
complicated matters.
In the S-meter
circuit a small voltage is added to the AGC voltage, so the meter will always
show a minimal deflection, even without input signals. This solves the '0' in
the logarithmic scale. Usually this start-point voltage may be tuned with a
trimming potentiometer. Calibration is performed with no signal at the
antenna input. Meter deflection is set to S-0.
A second
potentiometer sets the gain of the metering circuit to control the
translation of volts on the AGC-rail to the reading at the S-meter. Gain
setting is performed at an input level when all gain stages are somewhere in
the middle of their respective control range, usually with antenna input
signals around 50 microvolt. The S-meter is set by the gain control trim
potentiometer to S-9, usually in the center of the S-meter scale. In this way
the meter is giving some 'head-room' below and above this reading.
Around this set
point the S-meter is behaving reasonably well and is showing the S-points
according to the earlier mentioned formulas, because all gain stages are
acting around their mid-range. Somewhat further away from this calibration
position, usually more so below than above this position, the readings will
be less accurate
At low antenna
input signals the various gain stages are operating closer to the start of
their individual control range. This is giving rise to larger inaccuracies on
the meter reading, usually translating into lower than appropriate readings;
the meter is getting 'pessimistic' e.g. showing S-5 when the actual level
should read S-6. Depending on the specific receiver, readings below S-6
usually tend to be less reliable with higher inaccuracy at lower
numbers.
The same is
true, but usually to a somewhat lesser extend, for
meter readings at high input levels. Some gain stages are sooner at the end
of their individual control range than others, again leading to less reliable
S-meter readings; S9 + 20 dB reading
could well be an actual signal of S9 + 30 dB.
It should be not
too difficult to correct these differences at the S-meter dial. This however
could lead to non regular markings, resulting in
customer questions. Further more there always will
remain the individual receiver difference. Professional receivers usually do
not suffer from these inaccuracies since signal strength metering usually is
derived in a different way.
Accurate signal strength metering
An exception to
the inaccurate signal strength meter may be found in Software Defined Radio
(SDR) type of receivers, especially those that convert the incoming RF
signals directly into Low Frequencies, e.g. those based on the Tayloe mixer principle. Usually these LF signals are
directly converted into the digital domain by means of the A/D converter at
the audio-board. Even the internal PC chip sets for audio processing,
nowadays usually integrated onto the PC motherboard, are showing high
accuracy and 96 dB (16 bits) of resolution or more. This means that the large
range of input levels are now dealt with in a digital way.
For this type of
signal processing in the digital domain many SW programs are freely available
on the internet, one of which is Flexradio. This
program is designed with the radio-ham in mind using a lay-out of a familiar
radio radio-receiver, be it a sophisticated one with a large panoramic
display of signals at and around the tuned frequency. The 'receiver' is
exhibiting many more interesting features, but we will concentrate further on
the panel meters for this story.
Because of the
above mentioned digital signal processing, the meters displaying the tuned
signals in dBm. and S-points are so accurate and constant over the range,
that this may easily be regarded as an fine measuring instrument, once the
system is accurately calibrated. The dBm. meter is showing a resolution to
0,1 dBm. that is every bit true and accurate over the entire input range from
around 10 dB above the noise level up to - 0,1 dB from saturation. Actual
range of this 'instrument' is somewhat depending on the specific
'audio-card', but tests are showing that even using the internal audio
chip-set, dynamic input range is at least over 95 dB.
Calibrating the S-meter
Most
radio-amateur receivers still have to face one more problem.
Earlier we
discussed the S-9 display at the meter for an antenna input voltage of 50
microvolt. For calibration purposes, this voltage is delivered from a HF
generator. Formerly, the output meter was displaying the voltage at the
output terminals of the generator. When connecting to a load, this voltage
dropped to a lower value, depending on this load and the output impedance of
the generator, the latter usually 50 Ohm, but 60 Ohm or 75 Ohm were not
uncommon either. This output metering circuit was designed to show the
correct output voltage, no matter the frequency or output level. This was not
no simple requirement, that become more difficult when next generation
generators were designed at even higher frequencies. Therefore modern
generators have a metering circuit at the input of the output attenuator,
when voltages are at high enough level. The output attenuator now is to
operate over the full frequency range, which in practice is difficult, but
possible. With metering at the attenuator input, the output level accuracy is
to a large extend determined by the load, usually specified at exactly the
characteristic load impedance for this generator, often at 50 Ohm. All
equipment and connecting cables are now required to show this characteristic
impedance for the generator output meter to show the correct value at the
terminals.
Receiver input impedance
The receiver
input impedance is one of the important factors that make calibration of the
S-meter a tricky operation.
Transceivers usually
are specified to be terminated into a specific load to comply to factory
specifications and also for safety reasons. For amateur radio equipment this
load usually is 50 Ohm. To many people this implies the output impedance of
the transceiver therefore is also 50 Ohm. In a different article on the output impedance of an amateur radio transceiver
this has been described and measured and appeared to be depending on
frequency and on specific power setting. Actual output impedance was varying
between 25 en 180 Ohm, depending on settings for a particular device under
test.
Next
implied idea about amateur radio equipment is the receiver input impedance to
be 50 Ohm. Again this is a precious wish but no practice for most
transceivers and when measured this again appeared to be frequency
dependent. When connecting a generator
to such receiver with a generator output reading of 50 microvolt, the actual
voltage at the receiver antenna input therefore is not well defined. Since
manufacturers of radio-amateur equipment were not able to guarantee the
receiver input impedance at 50 Ohm, the following compromise once has been
reached:
At calibrating
the S-meter, the calibrating generator should have an output impedance of 50
Ohm. The output voltage of the generator will be set to an open voltage of
100 microvolt and will be connected to the antenna input with as short as
practical leads. The S-meter display of this receiver should read S-9.
Although
this may seem to end all further discussion, keep in mind that the receiver
input impedance is varying with frequency. A well calibrated system at one
frequency will show a different value with the same antenna input voltage at
a different receiving frequency. On top of this, keep in mind many receivers
to also be equipped with input amplifiers and attenuators with their own
peculiarities.
Reading is field strength?
To
complicate matters further, the antenna (system) is part of the story. We may
be aware of the fact antennas to exhibit a different gain depending on
antenna design. When operating an antenna over frequency, the radiation
impedance will vary and so will the impedance transfer characteristics of the
lines connecting this antenna to the transceiver. In the end, there is no
clear and constant relation between the field strength at the antenna and the
voltage at the antenna terminals of the receiver. More on this subject in a
different article; 'Where does the power go?'.
For
all of these reasons, the reading of an S-meter usually only is of some
comparative value when discussing signal levels of different stations
operating at the same frequency, or when performing tests at different power
settings at 'the other side'. Since meter readings usually only are
reasonably accurate between S-6 and S9 + 20 dB, these comparative tests
should preferably be limited to this meter range.
Conclusions
In
view of all of the above, it now may be better to comprehend that the S-meter
readings in most radio-amateur publications nowadays are presented in a
qualitative way, rather than quantitative. To really appreciate the meaning
of the S-meter report, it should be a good idea when next to the reading on
signal strength, also a reading of the background noise is reported. In this
way the other station will obtain a more balanced view on the effect of his
radio transmissions at the receiving site. Reporting an S-9 signal against
S-6 background noise, is meaning more when compared to an S-9 report against
a S-9 background situation.
Bob
van Donselaar, on9cvd@veron.nl
|