HAM Radio 1/HamRadio.cdr/tech/amnoise/amnoise.txt

Text File | 1987-09-18 | 15KB | 256 lines

                         Simple Method used at KCRW
          to minimize Synchronous AM Noise in FM Transmitters.

    Over the past several years there have been a number of articles
    and papers on minimizing the effects of AM noise in FM
    transmitters.  I am sure that by this point the majority of
    working Broadcast Engineers are aware of the effects of
    mistuning or narrowband tuning of FM transmitters.  Increased
    crosstalk and/or multipath are the resultant problems.

    The one common thread that the authors of these articles have
    stressed is to measure the Amplitude Modulated portion of the
    signal using a "precision envelope detector".   There have been
    few suggestions as to whose unit to use or even possible sources
    for these devices.   They are not commonly found in Broadcast
    facilities or through the common Broadcast Equipment vendors.
    At the 1986 SBE National Convention in St Louis, this subject
    was discussed again by Geoff Mendenhall of Broadcast
    Electronics.   He suggested that such a device should introduce
    a VSWR of less than 1.06:1.   He further stated that the
    Modulation Monitors in general use are not accurate enough for
    proper measurement of Synchronous AM Noise in FM transmitters.
    At the same conference, this subject came up again during a
    panel discussion.   A suggestion was made from the audience by
    John (J.D.) Strahler of KTMS, Santa Barbara CA.    He felt that
    the simplest means to tune for minimum incidental (Bandwidth
    related) AM noise in an FM transmitter was to monitor the output
    of the forward sample diode on the transmitter output
    directional coupler with an audio amplifier and speaker.
    The physics of the subject are simple.  Any variations in the
    amplitude of the carrier detected by the sampling diodes will be
    carried on the rectified RF (which is to all practical purposes
    DC) and can easily be used to determine the "relative level" of
    the Synchronous AM Noise in an FM transmitter.
    The suggestion struck me as extermely sensible and I decided to
    try this method.   I will detail the results of this trial.


                                                    AM noise        pg 2.

    KCRW operates with an ERP of 6900 Watts at 1100ft AHAAT.  The
    antenna is a HARRIS (ERI) FMH-2AE 2 bay fed using 3" air
    dielectric co-axial cable.   The FM transmitter is a CSI
    FM12000E operated with a TPO of 7410 Watts.   This transmitter
    is a two tube transmitter with a grounded-grid final tube.  The
    tubes are a 3CX10,000A7 final and a 5CX1500B driver.  The
    exciter is use is a HARRIS MX-15.  This exciter is backed up by
    a HARRIS MS-15 on hot standby.  KCRW operates in Stereo with
    subcarrier tenants on both 67KHz (voice & music) and 92KHz
    (9600bps data.)

    At KCRW we have been tuning for minimum AM noise for quite some
    time.   The first step in the tuning is to set the tuning for
    minimum observed reflected between the exciter and the
    transmitter input.    We know from long experience that the most
    critical tuning is the input tuning of the transmitter (input of
    the 5CX1500B).    The normal method that we use at KCRW for
    monitoring AM noise out of the transmitter is the RCA BW75 Main
    chan and BW85 Stereo/test set monitors.   Though these monitors
    say RCA on the nameplate they are also known as the FMM-1 and
    FMS-1 by BELAR.   It is fair to say that they are in common use.

    The RF input to the FM Modulation Monitor/Test Set is taken from
    an RF tap on the PHELPS-DODGE Directional coupler that was
    provided with the transmitter.  The RF tap is variable and is
    set to the minimum point.   The coupler is installed after the
    transmitter output low pass filter.   This filter is considered
    by the FCC to be a part of the transmitter.   Sampling at the
    output of the filter prevents the harmonics normally found in
    the PA from causing inaccurate readings of AM Noise.
    Connecting the Modulation Monitor to the PA sample point may
    cause inaccuracies in AM Noise readings as well as Modulation
    percenage.   The placement of the sample coupler prior to the
    low pass filter will cause similar inaccuracies.

    We measured the output of the tap (at minimum setting) at 0.48
    Watts into 50ohms resistive.   The feed cable to the monitors is
    RG58-U in two pieces of 12ft each with a "BNC" "tee" between
    them to allow us to connect a Spectrum Analyzer for measurements
    of transmitted spectra.   This cable also drives (through
    isolation/loss pads) test or monitor receivers belonging to our
    subcarrier tenants.


                                                    AM Noise        pg.3

    The design of the Phelps-Dodge directional coupler is
    particularly well suited to easy connection of a speaker amp to
    either directional coupler output.  The directional coupler
    diodes are contained in an assembly that is "N" male at the end
    where it connects to the sample line on the "inline coupler" and
    "BNC" female at the end used to drive the meter (output of the
    diode).

    I simply installed a "BNC" "tee" between the output of the diode
    assembly and the cable to the transmitter forward power meter.
    I then connected another length of RG58-U to the tee and at the
    opposite end I used an adapter from "BNC" to 1/4in male phone to
    connect it to the input of a FOSTEX 6301 speaker/amp assembly.
    The simple method of tuning for minimum Incidental AM noise
    described above is ideal for the "working engineer".   It does
    not require any unusual or elaborate test equipment.   It can be
    done without any interruption of normal station operations.
    The tests we made at KCRW to verify this method of tuning for
    minimum Incidental AM Noise were empirical in nature.   The
    tests were done with the normal KCRW programming on the air.

    They are described in greater detail in the following text.
    KCRW FM 89.9MHz is licensed to Santa Monica CA with a
    transmitter located in the "Santa Monica Mountains" above West
    Hollywood CA.   One of the translators operated by KCRW is
    licensed to Ventura California.   This translator is located
    about 50 miles west of the KCRW transmitter site.   W. "Dow"
    Jones is an engineer for one of the Ventura CA stations and is
    under contract to KCRW to maintain the Ventura translator.   He
    has an SCA/Main receiver at his residence (15 miles from the
    translator) where he receives an adequate signal.

    We co-ordinated these tests so that he listened to the level of
    crosstalk from the KCRW programming into the (67KHz) Subcarrier
    while the transmitter tuning was varied.   The subcarrier
    receiver used for these tests was a Johnson DTR-6 tuned for the
    translator output frequency of 89.1 and receiving a subcarrier
    of 67KHz.   The programming on the subcarrier is simultaneous
    real-time translation into Russian of various Network TV
    programs.   It is a subscription service operated by expatriate
    Soviet Jews.   The nature of the subcarrier programming
    was ideal for the listening required for these tests.


                                                     AM Noise        pg. 4

    The point of minimum program audio level out of the test speaker
    was also at the point of minimum crosstalk of KCRW program into
    the Subcarrier on the receiver in Ventura CA.   It was
    simultaneously the point of minimum crosstalk from the 92KHz
    data bursts into the 67KHz receiver.   The point at which the
    level of the program audio out of the speaker was minimum
    differed from the point of minimum AM indicated on the RCA/Belar
    FM Monitor(s).   At the point of minimum level from the speaker,
    both the incidental and continuous AM noise, as indicated on the
    BW85 (-48dB), were about 8db higher than the minimum reading
    when the transmitter was tuned based on the indications from the
    RCA/BELAR Modulation Monitors (-56dB).

    KCRW also has a QEI 691-02 modulation monitor.   We do not use
    the QEI for AM noise measurements.   The QEI AM Noise readings
    are about 10dB worse than those on the RCA/BELAR.   I believe
    that the readings on the QEI are that much less accurate.
    At the point of minimum program audio on the speaker, the
    continuous components of the AM noise remain.   The residual hum
    from ripple components in the transmitter power supplies are
    clearly audible.   Given the proper instruments, the individual
    hum components may be identified and, working with the
    transmitter manufacturer, decreased.   As John Strahler of KTMS
    stated to me in a letter, "..it also makes it very easy to
    locate a rectifier stack gone sour."

    I chose to perform this test through the translator as a "worst
    case" test and was gratified to discover that, in Ventura, the
    level of crosstalk from KCRW programming (main and stereo
    subchannel) into our 67KHz subcarrier was not a function of the
    translator.   The TTC XL10FM series translator passed on the
    KCRW signal unchanged, including the main transmitter generated
    crosstalk components.

    I have taken the time to verify similar crosstalk performance on
    the main KCRW signal using the same model SCA receiver at my
    residence.   I am lucky enough to have excellent reception at my
    home.   The reception is so good that I am able to receive a
    KCRW signal (albiet a noisy one) from exciter leakage through
    the transmitter when the transmitter fails.   (An outside
    antenna does help!)  This is over a distance of 8 miles.


                                                    AM Noise        pg. 5

    The further proof was in observations that I made of station
    reception on the radio in my personal automobile.   There is a
    section of Interstate highway that I drive daily where I had
    observed a "problem area" of about 3 miles length.   For those
    of you familiar with Los Angeles it is I405 between I10 and
    Sunset Blvd.   While it is about 5 miles from the transmitter,
    there are high buildings in the near vicinity.   In this area I
    heard a whistle (high frequency tone) directly related to the
    presence of either subcarrier (worse with both subcarriers
    present) on the radio.   It had been evident in this area ever
    since the subcarriers were installed and tested.

    The problem is due in part to multipath distortion.   If you are
    unfamiliar with multipath, it is simply the summation of two (or
    more) received signals, one delayed in time from another
    (others.)   One of the signals is direct and the other
    reflected.   The direct may not be the stronger signal.   The
    analogy can be drawn to "TV ghosting" or standing wave problems.
    It can be evidenced by a "tearing" or "picket fencing" of the
    signal as you drive through an area.   In some cases the result
    can be nearly complete loss of signal (picket fencing.)   In
    others it is a loss of carrier but not sidebands (tearing.)  In
    some cases, particularly with clock radios in apartment
    buildings, the opening or closing of a nearby refrigerator door
    may result in a complete loss of the desired signal.

    As is typical of problems related to multipath distortion, the
    problem came and went as I drove along this stretch of freeway.
    After the alternate method of tuning the transmitter was used,
    was very pleasantly surprised to find that the problem area was
    nearly gone.   There are now only two short stretches of about
    100ft each where the problem with the whistle is still evident.
    now use that section of road as a daily check on the operation
    of the station.   Whenever the problem is present, we now check
    for some transmitter deficiency or, at the minimum, improper
    tuning.


                                                    AM Noise        pg. 6

    With this in mind, I will speculate on the mechanism that caused
    the problems with reception on the radio in my automobile.   The
    mistuning of the transmitter caused the transmitter bandpass (or
    the bandpass of the mistuned stage) not to be centered on the
    KCRW carrier.   When modulation was applied, the result was non-
    symmetrical sidebands.   The absence of sidebands on one side of
    an fm carrier is received as distortions of the original signal.

    The mistuning of the transmitter had a more severe effect on the
    subcarriers as the sidebands generated by the subcarriers fall
    further from carrier center.   The result of any mistuning of
    the transmitter appears at the receiver as an insidious type of
    intermodulation distortion.   The presence of additional
    distortions in the form of multipath bring the first distortion
    into clearer focus.  (a multiplicative effect)  Correct tuning
    of the transmitter bandpass alleviates much of the perceived
    problem with reception.

    These results fit the conclusions of Geoff Mendenhall of B.E.
    that the point of minimum AM noise may not be at the point
    indicated by the Modulation Monitor/test set.   They further
    suggest a simple means to verify the proper tuning of an FM
    transmitter.   It makes use of items commonly found at a
    transmitter site or readily available at most FM Broadcast
    stations.

    This method of tuning for minimum Incidental AM noise is ideal
    for the "working engineer".  It can be done without any
    interruption of normal station operations.  It requires no
    expensive or unusual test equipment and makes use of nearly
    every Broadcast Engineer's most useful tools - hearing and
    judgement.

    We are continuing to operate the transmitter tuned as per the
    above method for minimum AM noise based on the tests above.
    have sent a copy of this letter to a few others for review and
    comments.  Should you have any contradictory findings let me
    now.  ì
PN