G8MNY  > TECH     29.11.21 08:00l 214 Lines 9923 Bytes #193 (0) @ WW
BID : 54363_GB7CIP
Subj: CW/SSB Rig Frequency
Sent: 211129/0800Z @:EA5URD.EAMU.ESP.EU #:41615 XFBB7.07 Bid:54363_GB7CIP

R:211129/0758Z @:EA2RCF.EAVI.ESP.EU #:23209 [Vitoria] $:54363_GB7CIP
R:211129/0758Z 22305@ED2ZAE.EAVI.ESP.EU LinBPQ6.0.14
R:211129/0757Z 45394@F1OYP.FAQI.FRA.EU BPQ6.0.22
R:211129/0757Z 30027@CT1EBQ.CTLX.PRT.EU LinBPQ6.0.22
R:211129/0756Z 20036@CT7AXA.CTLX.PRT.EU LinBPQ6.0.22
R:211129/0756Z @:GB7CIP.#32.GBR.EURO #:54363 [Caterham Surrey GBR]


By G8MNY                                 (Updated Sep 13)
(8 Bit ASCII graphics use code page 437 or 850, Terminal Font)

Rigs frequency readouts are not all what they seem. Here is an explanation of
why they can be wrong. With the older pre PLL rigs, tiny frequency errors of a
few 100Hz was considered irrelevant, but with 10Hz digital readouts it looks so
bad. A rig should be warmed up before any re-calibration is done!

On modern PLL rigs there is an oscillator that is used to lock the VFO.
Changing this will alter all the Tx/Rx frequencies.

┌──────┐      ┌─────┐     ┌──────┐
│MASTER├──────┤MIXER├─────┤ LOOP │        The PLL needs to do the whole
│  OSC │      └──┬──┘     │FILTER│        of HF down to say 10Hz for
└──────┘         │        └───┬──┘        good tuning, however correcting
          ┌──────┴─────┐      │           a VFO only 10 times a second
 DIGITAL ­│PROGRAMMABLE│      │DC         to get 10Hz loop steps is not
FREQUENCY­│   COUNTER  │      │Control    a good strategy, as there is
  INPUT   └──────┬─────┘      │           slow tuning response & higher
                 ├──────>OSC  │           noise sidebands (wobble)
              ┌──┴──┐         │           than a faster loop.
              │ VFO ├───<─────┘
To solve this problem another method is used...

This is a pre calibrated offsets of a crystal oscillator (often the master one)
Sometimes used for just a 5kHz FM rig shift, but more sophisticated may be up
to 256 fine steps. You can tell if this is the methord used as there is always
was click heard on tuning over the PLL steps.

More typically 100x 10Hz steps to fill in frequency gaps of a 1kHz PLL step
oscillator. Correctly set up the result can be perfect, but incorrect it will
give uneven jumps for each PLL step.

      Just right              step gain to low           step gain too high
  interpolation steps        interpolation steps         interpolation steps
 ┴┼┴┴┴┴┴┴┴┴┴┼┴┴┴┴┴┴┴┴┴┼    ┼─┴┴┴┴┴┴┴┴┴┼─┴┴┴┴┴┴┴┴┴┼     ┴đ┼┴┴┴┴┴┴┴┴đ┼┴┴┴┴┴┴┴┴đ┼
    Course PLL steps           Course PLL steps            Course PLL steps

So the gain of the interpolation offset steps needs to be just right, so that
at the PLL step exactly equals the interpolation frequency jump. An error here
may CORRUPT the PLL Master Oscillator calibration!

For direct conversion rigs the PLL is used directly as the Rx & Tx frequency in
SSB, but there is a Rx offset to generate a CW tone of the heard station.

  SSB Direct     CW Direct        FM Only                 Multimode
  conversion   Conversion      Rx   Rigs   Tx                      Tx/Rx
      │          │Tone│        LO           │           LO           │
    ──┼──       ─┼────┼─      ─┼────────────┼─           ┼───────────┴─
     PLL         Rx  Tx         <----IF---->              <----IF---->
    Dial        Rx   Dial     PLL  Jump---->Dial        PLL         Dial

For superhets the Rx LO frequency is offset from the Tx frequency by the IF, so
the PLL is setup to do that.

In an FM only system, the PLL will be asked to jump to the Tx frequency on Tx.
(N.B. it must not Tx while jumping frequency!).

In a multimode the Tx signals are made at the IF frequency & may share the IF
filter for bandwidth limiting.

With these modes the Tx frequency is the same as the dial. If the rig uses a Tx
IF for these, it's frequency will need setting up to generate the dial
frequency. If this is not done after the PLL is set up, & the Tx freq used for
PLL calibration a frequency scaling error will occur & the PLL will be out at
other frequencies!

As well as the LO being offset by the IF frequency the modulation &
demodulation uses an IF BFO in the product detector/mixer. So there is another
frequency to set up. With USB & LSB modes it is placed above or below the IF
filter. The PLL system is programed with these mode offsets....

   USB Dial                                                LSB Dial
     Freq                   IF SSB Filter                    Freq
   0 _│             ______________________________             |
 -6dB_│           /~                              ~\           |
      │          /                                  \          |
-40dB_│      _.˙'                                    '˙._      |
      0  100 200 300              USB             2k6 2k7 2k8 2k9 Hz
     2k9 2k8 2k7 2k6              LSB             300 200 100  0

On non PLL rigs the BFO is normally set up to place the IF filter at it's best
place for good communications frequencies e.g. 300-2600Hz. But with a PLL
design a BFO frequency error will put the dial out!

      │BFO      ___________________________
      │ │      /                           \
      │ │     |                             |
        0     300                         2k6 Hz
     >   <

An exacting way of checking this is to tune into an accurate AM broadcast
signal & select AM, USB, & LSB modes, & see if there is any pitch difference,
there should be none. (If the rig display is one that jumps in frequency
between modes, it will need resetting to the station's frequency when U change
    ________________________ │ ________________________AM signal
   /         /~~~~~~~~~~~\  \│/  /~~~~~~~~~~~\         \
  |         |     LSB     |  │  |     USB     |         |
   -4k5   -2k6          -300 0 300           2k6    4k5 Hz

  <----------------------AM Filter---------------------->

Use the RIT to eliminate USB/LSB pitch difference. Fixed RIT difference may
indicate the PLL is off, a different amount of RIT for USB than LSB indicates
the IF BFOs are off.

Rigs using double conversion & 2 IFs it is possible to overlap filters of
different frequencies to change the BFO to IF frequency without an apparent
frequency shift or to get variable IF widths.

    Dial                        IF Shifts
    Freq              LOW                                HIGH
      │         ___________________________________
      │        /               ˙~~~~~~~~~~~~~~~~~~~\~~~~~~~~~~~~~~~~~˙
      │       |               :                     |                 :
      0  100 200 300 400 500 600                   2k2 2k4 2k5 2k6 2k7 2k8 Hz

To do this the BFOs/LO are mixed together so that the overall IF difference is

    Freq                          W   I   D   E
      │            ____________________________________________
      │           /            ˙~~~~~~~~~~~~~~~~~~˙            \
      │          |            :       NARROW       :            |
      0  100 200 300 400 500 600                  2k2 2k4 2k5 2k6 2k7 Hz

Different makers use different controls for these functions. Simple ones let
you narrow the width by removing top or bottom bandwidth in a single control,
more complex ones use 2 controls to give a full combination of the 2 functions.

When set up properly IF shift & bandwidth should not affect calibration, but
they will need calibration in them selves, so that the filters do overlap fully
for the widest IF, say when the control is in the middle. This may have
consequences for the Tx IF/SSB tone.

This is a crystal that is tunable across the IF to filter out unwanted signals
on spot frequencies.

    Freq                       TUNABLE NOTCH
      │             ________________   _______________
      │            /                \ /               \
      │           |               <- U ->              |
      0  100 200 300 400 500                 2k4 2k5 2k6 2k7 2k8 Hz

It will not affect frequency calibration, except to itself so the centre of
control is centre of IF.

Note this is a split frequency mode! In Rx the offset PLL or IF frequency is
set up to produce a CW tone ideally equal to the CW Tx sidetone say 700Hz.

   Rx ---- BFO offset ---- │         Like the SSB mode to do
    ┼──────────────────────┼─        this the PLL or IF BFOs
    ├----- Rx CW tone ---->│         are shifted between Tx
                          Dial       & Rx & the 2 frequencies
     <─ CW AF Sidetone Osc ┤         will need setting up if the
          Fixed 700Hz                rig is being calibrated.

Then tuning in to a CW station at the same pitch has your CW Tx sidetone you
will be able to Tx EXACTLY on the same frequency & not a few 100Hz out!

If there is a narrow CW IF filter then the BFO will need to be calibrated for
that to do the same.

Of course the above is all true for Digital Signal Processing as used on
Software Designed Radios as well, except that all the calibrations are in the
software, so if the master PLL is OK all should be OK.

But generally to calibrate a Rig's frequency there are several BFO oscillators
to set up.. USB, LSB, CW Tx, CW Rx, AM Tx & FM Tx, as well as the PLL master
oscillator. VHF/UHF sets may have other up mixer oscillators as well.

See also my TECH buls on "Calibrating Frequency", & "Off air lock ref osc".

Why don't U send an interesting bul?

73 de John G8MNY @ GB7CIP

Read previous mail | Read next mail