<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<p>Ok, I'm trying to understand it. I've played with manual Alex
filter settings so I can see a bit better what's going on. I'm
assuming that with multiple rx the Alex filters on auto are opened
up so I would need to manually apply the correct filters for the
radio with focus or use auto if listening to multiple rx. As far
as the WBS is concerned it's not very useful with the noise floor
so high that signals can't be seen. Is there no way to improve
that?</p>
<p>Bob - G3UKB<br>
</p>
<br>
<div class="moz-cite-prefix">On 4/23/2018 2:50 PM, Tom McDermott
wrote:<br>
</div>
<blockquote type="cite"
cite="mid:CACO3nRTD=L9-BtsTLezJDnOT10YjDLqQk3SgymnivtkUD8q10g@mail.gmail.com">
<div dir="ltr">
<div>
<div>
<div>
<div>
<div>
<div>HI Bob -<br>
<br>
</div>
<div>The sample rate of the ADC is 122.88 Ms/s. The
samples occur at 1/122.88e6, or <br>
</div>
<div>one 16-bit sample every 8.13 nanoseconds. This
is about 3 gigabits per second.<br>
</div>
<div><br>
</div>
So the hardware takes 16384 time-consecutive samples,
then sends them to the host.<br>
</div>
Then it waits about 75 milliseconds (This is the
large-timegap.)<br>
</div>
Then it takes another 16384 time-consecutive samples....<br>
</div>
<div><br>
</div>
There are 256 real samples per USB frame. So 16384/256 =
64 USB buffers.<br>
</div>
Metis packs 2 USB frames into one Ethernet packet, so 32
Ethernet packets per<br>
</div>
<div>16384 real samples.<br>
</div>
<div><br>
</div>
-- Tom, N5EG<br>
<br>
<br>
<div>
<div>
<div><br>
</div>
</div>
</div>
<div class="gmail_extra"><br>
<div class="gmail_quote">On Mon, Apr 23, 2018 at 5:59 AM, Bob
Cowdery <span dir="ltr"><<a
href="mailto:bob@bobcowdery.plus.com" target="_blank"
moz-do-not-send="true">bob@bobcowdery.plus.com</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
<div text="#000000" bgcolor="#FFFFFF">
<p>Hi Tom,</p>
<p>Thanks for the info and the code. I must admit I'm
still pretty confused on several fronts.</p>
<ol>
<li>In my code I'm collection 16384 samples starting
when the sequence number is dividable by 32 with no
remainder. I don't know if that's right as I didn't
understand what you meant by a large time gap. <br>
</li>
<li>To confuse matters the KISS code I have which I
think is the latest only collects 4096 samples. I
think my firmware is the latest and it looks like I
updated it in 2014.</li>
<li>I've been away from this so long I forgot Alex
sits in front and is set to auto. Then I thought,
what does Alex do on auto when several receivers are
running. If I run one rx my wbs display looks a lot
more like what I see on KISS but with 3 receivers
the noise floor is much higher and signals are
hidden. <br>
</li>
</ol>
Bob - G3UKB
<div>
<div class="h5"><br>
<br>
<div class="m_-8936519424513210642moz-cite-prefix">On
4/23/2018 12:32 PM, Tom McDermott wrote:<br>
</div>
<blockquote type="cite">
<div dir="ltr">
<div>
<div>
<div>
<div>
<div>Hi Bob -<br>
<br>
</div>
The number of samples was increased from
4096 to 16384 a few years ago in
Protocol 1 FPGA code.<br>
</div>
<div>I don't know the effective FPGA
revision number, but it was probably
2013 or so.<br>
<br>
</div>
The data rate would be very high if sent
continuously, so 16k samples are
time-contiguous but then<br>
</div>
there is a large time gap until another 16k
contiguous time set, etc. I treat each
block as a 16k vector<br>
of time samples.<br>
</div>
<div><br>
</div>
These are 16-bit raw samples, 2's complement
fixed point, real (not complex), and have not
been decimated.<br>
So the noise floor is not lowered by decimator
integration. If you do an 16k FFT on the
vector you will<br>
get a two-sided result.<br>
</div>
<div>
<div>
<div>
<div><br>
</div>
<div>If there is a bandpass filter in
front of the ADC then the spectra will
be shaped. In practice most<br>
</div>
<div>receive antennas / tuners will also
shape the spectra. A very broadband
antenna and no filtering<br>
</div>
<div>would be needed to achieve relatively
flat spectra.<br>
<br>
</div>
<div>I implemented a wideband gnuradio
module for this, you can look at the C++
code at<br>
</div>
<a
href="https://github.com/Tom-McDermott/gr-hpsdr/blob/master/lib/hermesWB_impl.cc"
target="_blank" moz-do-not-send="true">https://github.com/Tom-<wbr>McDermott/gr-hpsdr/blob/<wbr>master/lib/hermesWB_impl.cc</a><br>
</div>
<div>I don't know if that will help as much
of the processing is gnuradio specific.<br>
<br>
</div>
<div>-- Tom, N5EG<br>
</div>
<div><br>
</div>
</div>
</div>
<div class="gmail_extra"><br>
</div>
</div>
</blockquote>
<br>
</div>
</div>
</div>
</blockquote>
</div>
<br>
</div>
</div>
</blockquote>
<br>
</body>
</html>