RadioBanter - View Single Post

October 27th 04, 05:27 AM

In article ,
(Mark Zenier) wrote:

In article
,
Telamon wrote:
In article ,
(Mark Zenier) wrote:

Huh? There're no fast edges. It's a COFDM signal that consists of
hundreds of closely spaced subcarriers modulated as some slow baud rate
(around 30-50 Hz). It shouldn't slop over much, but it'll fill up all
the spectrum it's using. It's very similar to a FDM telegraph (Droning
DC-3) signal, only about 5-20 times as wide.

Mark Zenier Washington State resident

There are not hundreds of carriers. I believe the number is sixteen.

It is not the rate that they are switched at but how fast they actually
turn on and off. Think dV/dT.

The unfortunately vague article I found (Elektor Electronics, Dec. 2002)
said that the (sub)carrier spacing is 66.666 Hz, and that they use 64
QAM modulation. So there are somewhere around 160 subcarriers in a 10
kHz wide signal) and at 6 bits per baud the baud rate must be in the
20-40 Hz range. (As I understand it, the signal is around 40 kBPS).
(There's more tolerant form that uses 16-QAM).

Any time you digitally modulate RF it will spread out from the carrier
frequency. Now you can use a system like QAM -16 or 64 to minimize this
but it still spreads out and occupies more spectrum than the bandwidth
allocated. Their are many switching and mixing products that cause
digitally modulated RF to spread out above and below the carrier
frequency.

From plots I've seen you are at best *(averaged) 60 dBc 30 KHz away from
carrier center. This means it will easily wipe out the adjacent channels
and weak signals 2 channels away from it.

* Peak is much worse.

For weak signal work that's 40KHz of spectrum with digital buzz for a
9-10 KHz DRM signal and that is when things are working right.

--
Telamon
Ventura, California