[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Discuss-gnuradio] DC component
From: |
Marcus D. Leech |
Subject: |
Re: [Discuss-gnuradio] DC component |
Date: |
Fri, 23 Dec 2011 09:53:31 -0500 |
User-agent: |
Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.24) Gecko/20111108 Fedora/3.1.16-1.fc14 Thunderbird/3.1.16 |
I have noticed that there are some fixed frequency spurious signals in my
N210. These spurious signals are probably associated with the harmonics of
the clock. If your DC component is at some nice even frequency like 2GHz, I
would suspect a spurious signal to be the cause.
Evan
Spurious signals are a virtually-inevitable aspect of modern
receivers/transmitters. Many of us are used to radios that are
"purpose built", and probably don't realize that most such radios
have their own problems with spurious signals ("spurs"),
but that they get "tweaked" in the design phase to move those
(inevitable) "spurs" outside the operational envelope of
the particular application at hand. Your radio have a CPU? Move the
fundamental of its clock frequency so that the
fundamental and harmonics fall outside of your band of interest. But
in a radio whose "band of interest" lies anywhere
from DC up to a few GHz, that's a very tall order.
The good news is that most of the time, these "spurs" are quite weak,
and are generally overwhelmed by any actual signal coming
in from the antenna at the the same frequencies. For modern wideband
modulation schemes, an in-band spur that's 30-40dB below
your nominal incoming signal make essentially no difference to the
receive SNR. For narrowband weak signals that are coming in
just above the noise floor, it might be a different story.
I've attached a plot of 50MHz of spectrum (thanks to the latest
50Msps/sc8 support in UHD) around 1.420GHz, with the receiver input
terminated
in a 50 Ohm lab-grade termination.
You can clearly see spurious signals spaced every 5MHz, and a stronger
one right at 1.40MHz. The 5MHz may be from the ethernet clock,
not sure, but the stronger spur at 1.4GHz is very likely due to an
even harmonic of the 100MHz master clock. Even though this "spur"
at 1.4GHz is 40dB "out of the noise", in most applications the
signals themselves will *dwarf* that spur. The other spurs, across 50MHz
of bandwidth are no more than 20dB "out of the noise". They don't
worry me that much, even for applications like radio astronomy,
where the signals are really weak. Placing a low-noise gain chain
ahead of the receiver, with enough gain to "swamp" the receiver spurs
is all I need to make these go away.
It's true that a $40K laboratory-calibrated receiver like an (Agilent,
R&S, etc) spectrum analyser will likely have fewer "spurs". But if you
open one of those up, you'll notice a lot of individually shielded
sub-assemblies--that's not just for show. They'll also do tricks like
spreading the clocks for the control/monitoring CPU, shifting clocks
around, to make the "spurs" move away from the current band of
interest. And for the most part, things like laboratory spectrum
analysers are "deaf as a post"--they aren't designed, generally, to be
"off-air" receivers, so they tend to be less sensitive to their own
"spurs".
--
Marcus Leech
Principal Investigator
Shirleys Bay Radio Astronomy Consortium
http://www.sbrac.org
50msps.jpg
Description: JPEG image