Hi Reg,
I will be waiting (and I sure many more of us as well) to see the results of your vector network analysis
Applied to the BW of the DSO. I have been aware for many years that you could take the frequency domain response of a system (a scope and its plugin for example) and use it to push the bandwidth by applying the system response to a signal you were measuring that was at or slightly beyond the bandwidth of the scope. That was the best I could do to describe it. I'm sure you could say it much better.
The important thing is we are going to be eagerly waiting to see your results. Ideally you would provide enough of a tutorial so that we could try it for ourselves.
Dennis Tillman W7PF
toggle quoted message
Show quoted text
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of
Reginald Beardsley via Groups.Io
Sent: Wednesday, May 01, 2019 10:16 AM
Subject: Re: [TekScopes] My TDR Evaluation of the SG503 012-0482-00
Cable has been REVISED
Albert,
Pretty much. To increase the dynamic range I'll break up the trace from
the DSO, window, align and sum the reflections before doing the FFT.
My professional background is reflection seismology in the oil industry.
Typically impulsive sources are used, but Vibroseis is widely used,
though generally on land. However, I did process a marine Vibroseis
line in 1982-83 when I started work for Amoco. The data quality was
very impressive. They had put a truck or two on a barge and pushed it
around with a tow boat. (Despite the name, tow boats push barges)
Vibroseis is just FDR at low frequencies using elastic waves. It's very
popular where setting off dynamite charges would not be acceptable. So
it's the standard source in urban areas.
Because drilling a single deepwater well is a $150 million undertaking
and evaluating a discovery requires drilling 3-4 wells, oil companies
spend very large sums on seismic data acquisition and processing. A
rough estimate for a modern survey is $10-15 million for acquisition and
$5-10 million for processing. What's more that's just the first of many
surveys over the producing life of an oil field.
No entity spends as much money or applies as sophisticated DSP as the
oil industry. Though as processing power increases and data BW
increases, DoD is moving into the same space. They'll never catch up
because seismic has the luxury of working in recorded time acquiring
data over many weeks and processing it over many months using warehouses
full of computers.
I am often bemused by EEs telling me I don't understand DSP. It all
started with Norbert Wiener's Geophysical Analysis Group in the 1950's.
The oil industry had a major problem with reverberation in the near
surface, especially offshore Middle East. And they could live with 16
bit 250 Sa/s DAQ on 24 channels. TI was formed as a subsidiary of
Geophysical Services Inc. for the express purpose of building seismic
data recording systems. TI is still around, but they sold off their
parent which is now long gone. The sole claim to fame of my PhD
supervisor at UT Austin was putting a 2nd playback head on an analog
drum recorder and adjusting it to suppress reverberation by inverting it
and summing in the late 50's before digital gear became available. He
was one of the last of Wiener's students and later became head of
research at GSI before moving to academia.
The point of the preceding narrative being that when I say I can take
one of Leo's square wave generators and provide vector network analysis
to the BW of the DSO and >100 dB dynamic range I know what I am talking
about. And correcting for the screwball AFEs on DSOs is not a problem,
just another algorithm.
Have Fun!
Reg
--
Dennis Tillman W7PF
TekScopes Moderator
