I've been interested in the 630 meter band between 472 and 479 kHz for some= time, I would like to see just how good a receiver I can make. 5W EIRP maximum, and contacts are already being made over huge distances. -----Original Message----- From: piclist-bounces@mit.edu On Behalf Of Harold= Hallikainen Sent: Wednesday, January 9, 2019 2:38 PM To: Microcontroller discussion list - Public. Subject: Re: [EE] DSP algorithms for very narrow band applications Sounds like an interesting project! I have very little DSP experience other= than telling a DSP engineer what I want, and the code magically appears. B= ut http://www.earlevel.com/main/2013/10/13/biquad-calculator-v2/ has proved useful. I'm currently using a biquad as a CW filter for receptio= n by ear. Picking up signals below the noise level is interesting. This, of course, is what FT8 and similar weak signal techniques do using mu= lti-frequency shift keying. Are you also looking at machine decoding of the CW? I note that there is a = CW decoder in KiwiSDR, but it is very error prone with noise and fading. I am, as a background project, looking at using SOMETHING for amateur radio= applications. My transceiver has an RS485 control bus along with a balance= d audio bus and PTT. I'm thinking of maybe using an old netbook to control = it, stream audio, etc., OR using a PIC32 (maybe MZ) to do the same. I'm MUC= H more familiar with PIC32 software, including streaming audio (PCM and ADP= CM) than I am familiar with linux talking to hardware like a USB audio inte= rface, USB RS485 interface, etc.). So, still playing around with ideas. I look forward to hearing other comments you get. Harold http://w6iwi.org > I had posted a couple weeks ago and I guess it fell in the bit bucket. > > I'm looking for info on DFT or similar approaches to detect very weak=20 > CW signals buried in noise. > I'm only interested in a 200Hz window, and I can control where it=20 > falls frequency wise, between 150Hz (center) and several kHz. > I need as much frequency detail as possible, sub 1Hz definitely, but=20 > the "dot" is pretty short. At 20WPM that works out to well below. > https://www.w8ji.com/cw_bandwidth_described.htm > > This will get implemented on an AVR or similar, with input from the=20 > on-board ADC. > > I expect I will need a rather long array of samples, but I am starting=20 > to think of a circular buffer there and sending chunks of to the DSP proc= ess. > > I have looked in Smith's DSP book, but I haven't found what I'm=20 > looking for. > > -- > http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive=20 > View/change your membership options at=20 > http://mailman.mit.edu/mailman/listinfo/piclist > -- FCC Rules Updated Daily at http://www.hallikainen.com Not sent from an iPho= ne. -- http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/chang= e your membership options at http://mailman.mit.edu/mailman/listinfo/piclis= t --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .