[sci.electronics] AM Radio

elliott@optilink.UUCP (Paul Elliott x225) (01/30/90)

In article <26374@cup.portal.com>, ISW@cup.portal.com (Isaac S Wingfield) writes:
> 
> If the transmitter is a standard plate-modulated rig, the 5 Hz will
> certainly saturate the modulation transformer, with possibly dire
> results; the 20kHz will spill, provided you can get it through that
> same transformer...
> 
> Modulation transformers are big, heavy, and *expensive*. Rest assured
> that the manufacturer didn't use any more iron than was necessary to
> get down to 50Hz, nor pay any more attention to distributed capacitance
> than necessary to get to 10kHz. [...]

Slightly off the subject of the original question (re: improving the
transmitted frequency response), What _is_ the current practice in AM
transmitter design?  I am familiar with the classic plate-modulation
methods, but many years ago studied other methods such as using two
unmodulated carrier sources, phase-modulating them, and combining them to
yield an AM signal.  This had the advantages of not needing all the iron
of the plate-modulation schemes, possibly better efficiency, and better
fidelity.

Can anyone educate me as to the current state of the art?


-- 
Paul M. Elliott      Optilink Corporation     (707) 795-9444
           {pyramid,pixar,tekbspa}!optilink!elliott
                  "The dog ate my disclaimer."

hanavin@udel.edu (Chuck Hanavin) (01/31/90)

>Slightly off the subject of the original question (re: improving the
>transmitted frequency response), What _is_ the current practice in AM
>transmitter design?  I am familiar with the classic plate-modulation
>methods, but many years ago studied other methods such as using two
>unmodulated carrier sources, phase-modulating them, and combining them to
>yield an AM signal.  This had the advantages of not needing all the iron
>of the plate-modulation schemes, possibly better efficiency, and better
>fidelity.

>Can anyone educate me as to the current state of the art?

-----------------------
The current state of the art being used is called Pulse Duration
Modulation. Here, the audio signal is digitized with a switching
rate of 40 to 60 khz. A power supply of twice the unmodulated carrier
value is needed. The power supply is turned on and off by the digital
audio. The amplitude of the original audio signal relating to the
duration the power supply is on, and the frequency of the audio signal
relating to the frequency the pulse is on. A filter is used between the
modulator and modulated amp to filter out the 40 to 60khz switching
frequncy. The modulator is directly coupled to the modualated amp, hence,
no mod transformer and excellent frequecy response. Since the moduater
is operated as switch (on or off), it has very high efficiency (80 to
95 percent). 

Chuck (WB3FJJ)

 

markz@ssc.UUCP (Mark Zenier) (01/31/90)

In article <3044@optilink.UUCP>, elliott@optilink.UUCP (Paul Elliott x225) writes:
> Slightly off the subject of the original question (re: improving the
> transmitted frequency response), What _is_ the current practice in AM
> transmitter design?  I am familiar with the classic plate-modulation
> methods, but many years ago studied other methods such as using two
> unmodulated carrier sources, phase-modulating them, and combining them to
> yield an AM signal.  This had the advantages of not needing all the iron
> of the plate-modulation schemes, possibly better efficiency, and better
> fidelity.

One method, describe to me by a station engineer, is to jack up the
voltage by a factor of two and use a second power tube as a cathode
follower switch.  This top tube was switched with a 70 Khz pulse width
modulated version of the audio, and then filtered to provide the plate
voltage for the RF amp.  The filament transformer for the switching tube
must be something special.

markz@ssc.uucp

rmf@bpdsun1.uucp (Rob Finley) (02/04/90)

0000000014 0000000001 n
alt.hotfut 0000000000 0000000001 y
rec.arts.startrek.info 0000000002 0000000001 m
comp.sys.laptops 0000000191 0000000000 y
alt.hackers 0000000069 0000000000 m
clari.feature.dave_barry 0000000006 0000000000 m
clari.feature.mike_royko 0000000012 0000000000 m
clari.feature.miss_manners 0000000007 0000000000 m
clari.feature.kinsey 0000000000 0000000000 m
clari.feature.lederer 0000000002 0000000000 m
alt.cobol 0000000056 0000000000 y
alt.fan.dave_barry 0000000091 0000000000 y
news.software.nn 0000000057 0000000000 y
sci.virtual-worlds 0000000025 0000000001 m
Inventor of the Year (or whatever)
by the National Association of Broadcasters.  PDM is used in our SX series 
of 1, 2.5 and 5Kw AM transmitters.  To go to 10Kw, we have come up with an 
even neater scheme.

Phase modulation for audio works like this:  The audio is fed into one
side of a comparator while a 60Khz triangle wave is fed into the other side.
There are two comparators for the two phases of PDM.  (We also have implemented
a single phase but it isn't as efficient (uses fewer parts though)).  The
other comparator has a triangle wave 90 deg out of phase with the first and
also sees the inverse of the audio signal.  When the triangle wave level 
is higher (or lower, in the second phase) than the audio signal, the comparator
output turns on and drives a MOSFET-based class-D amplifier which allows
the carrier signal to reach the output filter network.  The higher the signal
level is, the longer the amplifier is going to feed carrier into the antenna.
Radio Electronics ran two construction projects of audio amplifiers based on
our PDM technology.  I disagree with their "digital" labels but Check 'em out.

For higher power levels (up to 50Kw now, higher when we get this project done),
we have gone to another technique developed by Mr. Swanson using digital
controlled arrays of class-D amps to produce levels of modulation (DX series).
We take an audio signal, feed it into a 12bit A/D converter with a sampling
rate higher than 400khz, decode the output into individual levels, and
turn off and on an array of switches.  It's kindof like a bar graph:  the
higher the signal level, more modules are turned on.  We also do a bit
of encoding of Big step and Little steps to reduce the number of modules,
but that discussion gets involved.
The neat thing about both of these techniques is the abundance of ways
to bypass sections if something dies (just in case...).  All of the
digital modules are redundant and plug-swappable.  If one quits, you can't 
see it on an 'scope, and you would have to lose quite a few before the 
listner can even tell.  Module failures are tracked with a generous number
of indicators.

For more information, visit the latest issues of broadcasting electronics
textbooks.  I apologize if any of this information is incorrect.  I am
not an engineer here but I am working on that problem too.  

-----
"We as a company know what we are talking about and have sales leadership
to prove it.  I am an individual and since I am not an engineer, I am trying
to help you guys out by typing in stuff as I read the instruction manual.
Please don't start a flame war because I misunderstood what I read.  I knew
what I was talking about once but have since forgotten quite a bit..."

quintro!bpdsun1!rmf@lll-winken.llnl.gov    uunet!tiamat!quintro!bpdsun1!rmf