| |
G8MNY > TECH 01.03.20 23:28l 217 Lines 10529 Bytes #999 (0) @ WW
BID : 46098_GB7CIP
Read: GUEST
Subj: FM Deviation Calibration
Path: IW8PGT<HB9ON<IW2OHX<IR2UBX<IZ3LSV<I3XTY<I0OJJ<I0OJJ<GB7CIP
Sent: 200222/1135Z @:GB7CIP.#32.GBR.EURO #:46098 [Caterham Surrey GBR]
>From g8mny%gb7cip.#32.gbr.euro@i0ojj.ampr.org Sat Feb 22 12:40:42 2020
Received: from i0ojj.ampr.org by i0ojj.ampr.org (JNOS2.0m.1) with SMTP
id AA100769 ; Sat, 22 Feb 2020 12:40:42 +0100
Message-Id: <46098_GB7CIP@i0ojj.bbs>
>From: g8mny@gb7cip.#32.gbr.euro
X-JNOS-User-Port: Telnet (i0ojj @ 44.134.32.240) -> Sending message
From: G8MNY@GB7CIP.#32.GBR.EURO
To : TECH@WW
By G8MNY (Updated Feb 20)
(8 Bit ASCII graphics use code page 437 or 850, Terminal Font)
PHASE MODULATION
This is a form of FM where the carrier phase is changed not the frequency. It
used to be quite common on Xtal bound PMR rigs as the Xtal Q did not affect
modulator gain, as varicaping different crystals in FM designs does. The
modulator is slightly different as the varicap is not used directly on the
crystal oscillator, but in the following buffer stage.
It is measured differently to FM, in Rad/s, & 1 Rad/s = FM Modulation Index
(Dev/ModF) of 1 at ANY Mod Frequency. However it gives much the same results as
FM except the modulation is treble lifted at +6dB /Octave.
______ _______ _________ _________ _________ __________
Mic>´PreampôClipperô6dB/O cutôMod LevelÃÄÄÄÄÄÄÄÄÄ´Phase ModÃÄ´MultiplierÃÄ>
~~~~~~ ~~~~~~~ ~~~~~~~~~ ~~~~~~~~~ ________ ÀÄÄÄÄÂÄÄÄÄÙ ~~~~~~~~~~
³Xtal OscÃÄÄÄÄÄÙ
~~~~~~~~
So peak modulation has a different meaning in PM, unless there is a clipper set
up before a 6dB / Octave treble cut filter in the modulator (to turn it back to
FM), to limit the frequency related deviation!
FM MODULATION STANDARDS
Using Carson's rule the width of an FM signal is approx..
2x deviation + 2x highest modulation frequency.
This is not all the sidebands the FM process generates, but most of them, to
see the rest of them look into the Bessel function FM sideband harmonic series.
Carson's rule is simply explained by considering what happens at an instant
when a low frequency has given almost full deviation. The instant FM frequency
is at 1 side of the deviation window, & there is still some treble syllabance
mod to carry with its ñ3kHz sidebands like an AM signal. This gives the
diagrams below for 12.5kHz & 25kHz systems, where the lowest & highest
modulation sidebands are added to either side of the FM deviation like 2 side
by side AM signals. So the Rx also has to let in & correctly demodulate all
these wanted sidebands if there is to be no distortion.
12.5kHz CHANNEL FM SYSTEM
TX Bandwidth
Lowest _____________ Highest Rx .--------------. _-3dB
Lower /'³ deviation ³`\ Upper Bandwidth ³ ³
Sideband/' ³ +/-2.5kHz ³ `\Sideband (ideal) ³ ³
_________,/_____³___________³_____\._______ ____,'_-70dB `._____
Next ||<3kHz><----5kHz---><3kHz>|| Next |<----11kHz--->|
Channel ||----------11kHz----------|| Channel |<-----12kHz---->|
|----------12.5kHz----------|
N.B. there is next to no Rx protection "GUARD BAND" between channels on the
12.5KHz system! For this reasion commercially adjacent channels are NEVER used
in the same area!
For the 12.5kHz system a MAX of ñ2.5kHz Peak deviation is used, giving a
modulation index of 0.833 (Dev/ModF), which gives little capture effect over an
AM system.
The Tx also needs to have the AF response VERY WELL FILTERED, if the FM
sidebands are to be kept out of the adjacent channel.
0dBÄ´ .-ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ-. ÃÄ100% 2.5 kHz
-6dBÄ´ / >Ý ÃÄ 50% 1.25kHz
³ / Very tight audio filtering>Þ ³
³ / > Ý ³
³ / for no adjacent ch QRM > Þ ³
-70dBÄ´ / > Ý ÃÄ0.03% 0.75 Hz
ÃÄÄÄÂÄÄÄÂÄÄÄÂÄÄÄÂÄÄÄÂÄÄÄÂÄÄÄÂÄÄÄÂÄÄÄÂÄÄÄÂÄÄÄÂÄÄÙ
0 150 200 300 400 600 800 1k 1k5 2k 3k 4k Hz
The ideal Rx IF filter can't be made (see Rx bandwidth diagram), so in practice
narrower filters give better adjacent channel performance, but with quite high
audio distortion (10%), as some of the needed spectrum is lost.
With tight filters, the channel carrier frequency accuracy is important to keep
the Tx signal centre of the Rx IF. This is not so easy on higher bands without
very good Xtal stability (ovens etc), so the 12.5kHz system is NOT used above
VHF!
25kHz CHANNEL FM SYSTEM
TX Bandwidth |GUARD |
Lowest ____________ Highest BAND| Rx .------------. _-3dB
Lower /'³deviation ³`\ Upper Bandwidth ³ ³
Sideband /' ³ +/-5kHz ³ `\Sideband ³ ³
___________,/_____³__________³_____\._______ ____,'_-70dB `.____
Next | |3.5kHz<--10kHz--->3.5kHz| | Next |<----17kHz--->|
Channel| |---------17kHz----------| |Channel |<------25kHz----->|
|-------------25kHz--------------|
N.B. Here there is the luxury of an 8kHz GUARD BAND between channels on this
system, which is why it can work much better with strong adjacent channel
signals, than the 12.5kHz system, & with very little distortion!
For 25kHz system a MAX of ñ5kHz peak deviation is used, gives a modulation
index of 1.4 (1.6 if 3kHz) & has 2x (6dB) more noise rejection & capture effect
than the 12.5kHz system.
The Tx AF filtering, Rx IF filter & frequency accuracy, are far less stringent
than for the 12.5kHz system, & the comms sound quality can be quite a bit
better. Due to the wider guard band the adjacent channels have less Tx QRM in
them & Rx filters can more easily remove the adjacent channel signals.
EMPHASIS
With FM it is usual to apply some Tx treble pre-emphasis & Rx treble de-
emphasis, this is to mask the increased treble Rx noise (hiss) with the FM
system, & reduce overall system AF harmonic distortion. With comms bandwidth
the amount of emphasis cannot be very great, but up to +6dB @ 2kHz can be used.
FM Rx Tx Pre- Rx De- Overall Audio
Noise Floor emphasis Response
/|\ _.-' _.-' ```ÄÄ..__ ..............
Level__..,,-Ä' __..,,-Ä' `Ä._ _ _ _ _ _ _ _ _ Noise
2 3 5 1k 2k 3k 2 3 5 1k 2k 3k 2 3 5 1k 2k 3k 2 3 5 1k 2k 3k Floor
Freq ->
DEVIATION MEASUREMENT
Here are 3 simple methods for FM deviation calibration. Phase modulators use
6dB / Octave LF lift on the modultion to mimic FM & these need some care when
scoping @ the Tx to realise what you are seeing!
1 Bessel carrier null method.
Mr Bessel modulation index graphs show the 1st order carrier null occurs when
the Modulation Index (Dev/ModF) = 2.4, then again at 3.142 intervals after
that. This means a 1kHz sine wave modulation tone will produce a 1st carrier
null at precisely 2.40kHz deviation & a 2nd at 5.54kHz deviation. USEFULL!
dBs ³ |
³ ³ || || |, ,|
³ ³³³ |³³ ³³| .|³³³³³³³|. ,|³³³³³³| |³³³³³³|,
ÄÄÄÁÄÄÄf ÄÁÁÁÁÁÄ ÄÁÁÁÁÄÁÁÁÁÄ ÄÁÁÁÁÁÁÁÁÁÁÁÁÁÄ ÄÁÁÁÁÁÁÁÁÁÁÄÁÁÁÁÁÁÁÁÁÁÄ
No Mod Some Mod 1st Null More Mod 2nd Null
MI= 0 <2.4 2.4 >2.4 2.4+Pi
To monitor the modulation spectrum, you will need a SSB Rx with RF gain
control, ideally with a very narrow CW filter, or a Spectrum Analyser with a
narrow filter. E.g. a sound card from SSB AF output & an AF Spectrum Analyser
programme, or just good ears listing to just the carrier whistles nulling
while the other sideband tones get stronger.
Also an accurate & pure AF 1kHz sine wave generator is needed to feed the mic
circuit via a suitable attenuator (series 100kê ?)
Method.
Using a 1kHz sine wave tone, adjust modulation level (mic gain/deviation) to
produce no carrier on a SSB/CW Rx.
Now note the modulator drive level (e.g. scope it) @ the modulator, & ensure
that the AF FM clipper now hard clips anything at this level by adjusting the
deviation pot with the mic gain set at max (e.g. shout into the mic etc.)
2 Discriminator DC & Scope method. (can be used on air with a Rx/scanner)
Access to monitoring FM Rx's discriminator is needed to display the DC level
on a oscilloscope. Make sure the scope is connected to the discriminator
point before any de-emphasis components, & that the deviation sidebands being
measured will all fit through the IF filter, otherwise the display will lie.
SCOPE TRACE Fc+5kHz _____ _
.' `.
Fc ÄÄÄÄÄÄÄ MOD | | |
Fc-5kHz _____ ñ5kHz `._,'
Send a carrier, & change the Rx/Tx frequency +/-5kHz & adjust the scope gains
& position to give a +/- 5 division display.
Now anything you can Rx, will instantly show you the on channel "frequency
error" & "deviation" on the scope trace. N.B. any AC loading of disciminator
my give errors for this methord!
Method
For Tx deviation setting, just ensure the clipper hard clips anything @ this
level by adjusting the deviation pot with the mic gain at max. (E.g. shout
into the mic.) Then set deviation = ñ2.4 or ñ5.0 divisions peak to peak.
3 Accruate by design DDS Osc source.
With a DDS osc programmed to construct FM, the modulation source is correct
by design. I have a "FeelTeck Dual DDS Osc FY6600-60M" one, that generates
many orders of FM deviation accuracy, from a fraction of Hz dev to MHz
deviation. Above 60MHz signal harmonics are be needed. Use as the calibrated
source with method 2 above.
SETTING UP A DEVIATION METER.
At this point with a calibrated reference, it is relatively easy to make a peak
reading meter display, & calibrated in peak deviation for your Rx. Once you
have a calibrated source, it is easy to put a peak reading meter circuit (not
an average VU circuit!) onto any Rx & calibrate it. For accurate work a wide Rx
is needed.
HARMONICS
Using these the deviated signal will increase by their hamonic factor. E.g. 2M
rig at ñ2.5 kHz dev will have a 70cms harmonic deviation of ñ7.5 kHz, a usefull
check.
N.B. With some deviation meters, it is easy to accidently measure a harmonic &
not the fundimental's deviation!
Also see my buls on "FM Stereo Radio Principles", "RTT Mod Meter (100)" &
"FeelTeck Dual DDS Osc FY6600-60M".
Why don't U send an interesting bul?
73 de John G8MNY @ GB7CIP
Read previous mail | Read next mail
| |
