<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 12pt;
font-family:Calibri
}
--></style></head>
<body class='hmmessage'><div dir='ltr'>Kok - <div><br></div><div>You really didn't address the main point - would you agree that 2 MSPS 24 bit IQ data is in need of GigE?</div><div><br></div><div>Chuck<br><br><div>> CC: charlesh3@msn.com<br>> From: chen@w7ay.net<br>> Subject: Re: [hpsdr] Status of Hermes Gigabit<br>> Date: Sat, 2 May 2015 09:13:21 -0700<br>> To: hpsdr@openhpsdr.org<br>> <br>> <br>> On May 1, 2015, at 3:09 PM, Chuck Hutton wrote:<br>> <br>> > With ham bands less than 400 kHz and 16 bit samples, we can live without GigE.<br>> > But jump up to 17 bits or more (requiring 3 bytes) and / or move up to wider bands such as the medium wave band (let's say 2 MSPS) or wider ham bands and you will need GigE. <br>> <br>> The I/Q data from Hermes to the computer is already in the form of a pair of 24 bit numbers (>145 dBDR). <br>> <br>> The decimated Hermes I/Q data at 96 ks/s rate only has about 110 dB of SFDR and has lots of headroom (5 to 6 bits out of the 24 bits) left before we need to increase the I/Q data width. <br>> <br>> A faster sampling rate (e.g., 192 ks/s) actually reduces the numerical dynamic range requirement. The raw ("16-bit") ADC at 122.88 Msps starts with just 78 dB of dynamic range (about true 12.5 bits). <br>> <br>> Hermes' single signal "500 Hz ARRL" dynamic range of 125 dB is achieved when you finally decimate and demodulate inside the desktop computer at 500 s/s.<br>> <br>> 73<br>> Chen, W7AY<br>> <br></div></div> </div></body>
</html>