Hi all,

I have a circuit where I need to detect the amplitude of an incoming
10.7MHz signal. The signal consists of 10 microsecond long pulses of a
10.7MHz carrier. What I tried to do was to use a ceramic filter with a BW
of 360kHz (roughly corresponding to a rise time of 3 microseconds) to pass
the minimum BW needed to allow the pulses to pass, and then feed the result
into an Analog Devices AD606 Logarithmic Amplifier IC. However, I
discovered that when I try this, the background noise causes the output of
the log amp to be very noisy at low signal levels. I had originally thought
(from looking at the AD606 datasheet) that Gaussian noise was detected by
the IC much the same way as a sinusoidal signal, yielding a DC voltage out
which is proportional to the log of the power in the noise.

This output noise causes a problem because the output noise level is so
high that it extends to about 10dB higher than the average level (in other
words, let's say that the noise power is -70dBm and that this should
correspond to an output DC voltage of 0.5 V. What I actually see is a 0.75
pk-pk noise signal on top of a 0.5V offset as the output, and the slope of
the amp, in V per dB is such that the extra 0.75/2=0.375 V corresponds to
10dB, so if I sample the output of the log amp, I can see results which are
as much as 10dB above or below
-70dBm. This effectively means that my SNR needs to be > 10dB for even
remotely reliable detection of the pulses.).

Upon thinking a little further, I think my problem may be a fundamental
one: since my log amp can respond to fast edges (and must), it is
responding to every edge in the noise amplitude which is slow enough for it
to respond to, and there are plenty of these since the noise has just gone
through a ceramic filter, which removed all the really fast edges.

Now, finally, to my questions :-)

#1) Can someone either confirm my diagnosis of what is going on or tell me
what I might be doing wrong?

#2) Is there a better way to measure pulse amplitudes in the presence of
noise (without having to make a matched filter or do DSP with high sample
rates)?

Thanks,

Sean

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body