Speaking from lack of any experience of D/A converters, wouldn't D/A
conversion at 136kHz to generate the I&Q channels, thus allowing filtering
and DC elimination to be carried out digitally? Wouldn't this overcome th
problems Jim rightly pointed out below?
Didn't Charles, G4GUO, complete a 136kHz transceiver thing along these lines
a year or so ago?
John, G3WKL
-----Original Message-----
From: [email protected] [mailto:[email protected]]On
Behalf Of James Moritz
Sent: Tuesday, February 19, 2002 15:07
To: [email protected]
Subject: LF: Re: Vectorscope display
Dear LF Group,
I had the usual crop of bouncing e-mails, but I hope I have now sent the
vectorscope circuit to all who have requested it. Thanks to G3YXM for
putting it on his web pages - so hopefully it will also be available there
shortly. Correction - I don't seem to be able to send it to
[email protected], Johan, so perhaps you can get it from Dave's site
later, or is there something else I can do?
Re: Andy's comment:
...how about
the next version, a direct conversion receiver generating quadrature signals
at 137kHz then having masses of low frequency gain before the X/Y plot ?
A low noise, high gain version of the circuit would certainly be possible,
but there are a number of problems to be overcome here:
The signal spectrum at the mixer output extends to DC, so any DC offsets
will be amplified as well as the wanted signals. This would saturate the
output unless offsets could be kept down to a few uV. Even then, the
remaining offset would effectively be an unwanted 0Hz "carrier" in the
demodulated output spectrum. A related problem is that the noise figure of
amplifers, etc. increases drastically at very low frequencies due to
flicker noise. Both these problems could be avoided by having AC coupling,
but this would put a hole in the received signal spectrum around the LO
frequency.
Also, if we low pass filter the I and Q channels in order to define the
bandwidth, any mismatches between the filters in the two channels would
result in errors in the phase quadrature between the 2 channels, which
would be a problem if good image rejection between the sidebands were
required in subsequent demodulation of the signal.
I think it would probably be possible to achieve a useable result for a
dedicated 136kHz RX - a fairly simple bandpass filter before the mixers
could restrict the bandwidth to a few kHz, and some pre-mixer gain could be
applied, reducing the post mixer gain required and reducing the effects of
offsets.
Cheers, Jim Moritz
73 de M0BMU
|
