These Data  were submitted on: Saturday, November, 8, 2003 at 21:50:09
 
NAME:  Roger

CALL:  W3SZ

E-mail:  nospam

URL:   W3SZ 

CPU::  Pentium 3 

CPU Speed::  1 GHz

RAM in MB::  770 MB

SOUND CARD::  M-Audio Delta44

SOUND DRIVER::  OSS/Linux 3.9.7h 2.4.18-27.7.x

ADDITIONAL CARD1::  OnBoard AC97 Audio for VIA 

VIDEO CARD::  ATI Radeon 7000

VIDEO DRIVER::  Rage 128

LINUX DISTRIBUTION::  Red Hat

LINUX VERSION::  8.0

KERNEL VERSION::  2.4.18-27.8.0

GCC VERSION::  3.2 20020903

SVGALIB VERSION::  1.9.17

LINRAD VERSION::  01.07

TEXT:  This is my second system, Rein.

You can see my Linux experiences starting at:

W3SZ DSP Start Page
W3SZ DSP Primer, Linrad Section
Somewhat older W3SZ Linrad Page
W3SZ's Experiences with Linrad during the 2001 ARRL International EME                                                    
W3SZ Experience using Linrad in the 2002 ARRL Frequency Measuring Test
W3SZ FAQ Page for Linrad

I have used Linrad for over 2 years, on:
144 MHz EME
Microwave Weak Signal work
HF
VLF
with JT44 on 144 MHz EME (just to show it could be done).

with a variety of hardware including:

various homebrew front ends:
W3SZ FAQ describing one homebrew frontend for Linrad
Schematic of an early  W3SZ homebrew frontend for Linrad

Homebrew front end followed by Antennspecialsten RX2500:
Website for Commercial (Antennspecialsten) Linrad Receiver RX2500

The Time Machine:
The Commercial (Expanded Spectrum Systems) Time Machine Page
W3SZ Page Discussing The Time Machine

I have had Linrad running on several different computers and several different versions of Red
Hat Linux, starting with RHL versions 6.x

More details are in DUBUS Issue 4 2002 article by W3SZ.

As per 05/06/2004 Roger added the following:


  Hello, All!

  Following up on Joe K1JT's request for notes on people's experiences with
  Linrad, this is a brief note of my listening experiences with Linrad at
  W3SZ.

  For those of you who have been doing Linrad for a long time who have read
  my posts before, there is nothing new here, so feel free to delete this
  for-you-unnecessary bandwidth.  

  I haven't posted much in quite some time given the excellent quality of the 
  posts you guys are making, I really couldn't add anything worthwhile from a 
  technical standpoint; but a description of my experiences may be helpful to 
  those who haven't read my by-now-moldy posts from the distant past].
  
  I started using Linrad [before it was called Linrad] in 2001, and that
  year I used it during the ARRL EME contest as my primary receiver.  It has
  been my primary EME receiver ever since.
  
  My EME station location was so noise-filled that little EME was audible
  with conventional receivers even with their noise blankers and narrow filters.

  I could hear W5UN and KB8RQ and that was about it, even with 4 x 2xmp20
  array and 0.2 dB NF preamps on the tower.

  With DSP-Blaster and its noise reduction, I could pull quite a few more
  signals out of the noise, but it was still tough.

  The other software solutions that were [and still are] available were not
  really very helpful.  See
  
        W3SZ_DSP_new and

        LINRAD EME 2001 

  for a general discussion of things I tried, and a description of my first
  EME contest with Linrad, if you are interested in such things.

  Linrad was like a miracle for me in my high-pulse-noise environment.

  Leif's software noise blanker did an unbelieveable job of removing the noise at
  my receive station, and I have found ever since that when a two way qso
  doesn't occur now, it is nearly always now THE OTHER GUY who can't hear me, even
  though I am running 1500 Watts out to 4 m2 2mxp20's.  

  Most of my time during EME contests, after I have worked the first layer, is 
  spent answering very Q5 [on Linrad] stations who only seem able to send 'QRZ' 
  in response to my calls.  Note that you must CALIBRATE Linrad to make full use 
  of both the 'smart' and 'dumb' noise blankers.  The use of both is necessary 
  for the best results, but even the use of the 'dumb' noise blanker alone [that
  can be done without calibration] produces very good results.  I needed both
  where.  Leif's pages show just how to calibrate Linrad.

  So I have been very very happy with Linrad.

  It has been implemented here from the first with dual polarity.  At first
  I had only a moderately wideband system, allowing me to see 30-40 KHz of
  spectrum.  For a couple of years or so [I think] I have had wideband
  capability with about 90 kHz of bandwidth.

  The front end is homebrew, with two parallel channels using TUF-1H mixers.

  I have used PTS synthesizers as the LO's, as they have good phase noise
  for a non-crystal source, and they are frequency agile and I can control
  their frequencies from within Linrad via a patch that was added to the Linrad
  program see

        Lin2FT  

  This patch also controls from within Linrad the frequency of my FT1000 MP, 
  which is used as the transmit section of the Linrad Transceiver here.  
  Just a keystroke within Linrad sets the FT1000 to the Linrad receive frequency.

  Initially I mixed 144 MHz down to 40 MHz and then to baseband.
  
  When the RX2500 became available I mixed down from 144 MHz to 2.5 MHz and
  used it.
  
  Now I have just gotten around to installing the RX10700 and the RX70 and
  I mix down to 70 MHz.  When the RX144 becomes available, I will get rid of
  my trusty, simple, old front end and use it and all of my old hardware will
  then be out of the system.

  I have never used any commercial transverter / IF combination that has
  come close to Linrad and my simple homebrew hardware in my particular
  environment.

  For me the advantages of Linrad have been:

   1.  The noise blanker

   2.  Automatic received polarization angle control that follows the angle
       of the incoming signal

   3.  The wideband waterfall that lets me see the whole useful cw eme band
   
   4.  The graphically controlled filters.
   
   5.  The AFC that keeps the signal in the very narrow [usually 20-25 Hz]
       filter in spite of [the other guy's] drift

  I have also played with Linrad on HF, and for fun used it in the
  Frequency Measuring Test the first year it was resurrected by the ARRL. It of
  course did superbly, as I had all the oscillators GPS locked.  See
     
    ARRL Freq. test
  
  For that test I used 'The Time Machine" as my front end.
  
  I have also played with Linrad and my homebrew hardware as the 'IF rig'
  for 2304 MHz and up, but haven't done any serious investigations there yet.

  My Linrad computer software setup is pretty much as I detailed long ago
  on Rein's pages.  I use RedHat and never had any major software issues in
  any respect using it and the OSS drivers.  Those minor problems I had Leif
  helped me thru.  He has always been extremely instructive and patient,
  and I view him as an extraordinary 'natural resource' for us.

  There may be some other tidbits on 

      W3SZ Linrad tidbits that

  some will find interesting or useful.
  
  For me Linrad has meant the difference between being able to do a
  reasonable amount of EME from a noisy QTH, and not being able to do it at all.

  For all of this I thank Leif greatly.

  I hope the above has been of some interest and help.

  If it was boring and just a rehash too soon after my last post, please
  let me know [privately  ;)  ].

  And now to get back to getting that new USB printer working with Redhat
  8.0.

  73, Roger Rehr, W3SZ



                         -------------*********-----------------

  Sept 27 2004


  Hello, all,

  Here are my SSB parameters:

  First FFT bandwidth (Hz) [83]
  First FFT window (power of sin) [2]
  First forward FFT version [0]                     [SM5BSZ:  Change to 5 ( much faster )]
  First FFT storage time (s) [1]                    [SM5BSZ:  Change to 10 ( unless you have small memory, Allows 
  First FFT amplitude [1000]                                            longer averaging times )]
  Enable second FFT [1]                   
  First backward FFT version [0]                    [SM5BSZ:  Change to 1  is much faster but with Linrad-01.26 there will be             
  Sellim maxlevel [6000]                                 advantages of using floating point Linrad-01-25 and earlier, best use 1 here]
  First backward FFT att. N [6]
  Second FFT bandwidth factor in powers of 2 [2]    [SM5BSZ: Change to 0 or 1. Typicaly no need for narrow filtering in waterfalls for SSB]
  Second FFT window (power of sin) [0]              [SM5BSZ: Change to 2. Provides better filters] 
  Second forward FFT version [0]                    [SM5BSZ: 2 is much faster, but with Linrad-01-26 there will be advantages of using 
  Second forward FFT att. N [7]                                                floating point]
  Second FFT storage time (s) [5]  
  Enable AFC/SPUR/DECODE [1]                        [SM5BSZ: Use 1 here only if you actually want to use spur removal. AFC should not be
  AFC lock range Hz [150]                                                     used for SSB so make the window small and make it inactive on screen]
  AFC max drift Hz/minute [100]
  Enable Morse decoding [0]
  Max no of spurs to cancel [100]
  Spur timeconstant (0.1sek) [5]
  First mixer bandwidth reduction in powers of 2 [4]
  First mixer no of channels [1]
  Baseband storage time (s) [200]          [SM5BSZ: !!! Use much smaller time here. At high bandwidth this would require very much 
                                           memory. Many arrays are allocated in the baseband for CW decoding. The are useless in SSB mode
                                           but I have not removed memory allocation in ssb mode yet.
                                           In cw with a bandwidth of 50Hz or less, the sampling speed is very low and this is no problem. ]
 
  Output delay margin (0.1sek) [5]         [SM5BSZ: Maybe you can make this smaller. Check the margin with 'T' ]
  Output sampling speed (Hz) [10000]  
  Default output mode [1]
  Audio expander exponent [3]
  A/D speed [96000]
  Check [1110107]

  
  -----end of SSB parameters
 


  Here are my wcw parameters:

  First FFT bandwidth (Hz) [100]
  First FFT window (power of sin) [3]
  First forward FFT version [2]
  First FFT storage time (s) [20]
  First FFT amplitude [1000]
  Enable second FFT [1]
  First backward FFT version [1]
  Sellim maxlevel [6000]
  First backward FFT att. N [6]
  Second FFT bandwidth factor in powers of 2 [4]
  Second FFT window (power of sin) [4]
  Second forward FFT version [2]
  Second forward FFT att. N [10]
  Second FFT storage time (s) [15]
  Enable AFC/SPUR/DECODE [1]
  AFC lock range Hz [150]
  AFC max drift Hz/minute [100]
  Enable Morse decoding [0]
  Max no of spurs to cancel [100]
  Spur timeconstant (0.1sek) [5]
  First mixer bandwidth reduction in powers of 2 [4]
  First mixer no of channels [1]
  Baseband storage time (s) [150]
  Output delay margin (0.1sek) [5]
  Output sampling speed (Hz) [6000]
  Default output mode [1]
  Audio expander exponent [3]
  A/D speed [96000]
  Check [1110107]

  Change only between brackets.
  
  If file has errors, Linrad will ignore file and prompt
  for a complete set of new parameters

  linrad-01.25

  With the SSB parameters I get a delay on the order of 0.5-0.7 seconds or  less.  I am not sure of 
  the delay with the CW parameters listed above, as  I didn't check it today, but generally I have 
  had delays of 2-5 seconds  [which are not a problem for my purposes].

  I post them with no representation as to their optimization.  Rather, they  are being posted to 
  provide an object for criticism, and to provide a  framework for the discussion of the tradeoffs 
  and consequences of specific  parameter choices.

  These are running on this system:

  CPU::  Pentium 3
  CPU Speed::  1 GHz
  RAM in MB::  770 MB
  SOUND CARD::  M-Audio Delta44
  SOUND DRIVER::  OSS/Linux OSS 3.99.1g
  ADDITIONAL CARD1::  SoundBlaster PCI
  VIDEO CARD::  ATI Radeon 7000
  VIDEO DRIVER::  Rage 128
  LINUX DISTRIBUTION::  Red Hat
  LINUX VERSION::  8.0
  KERNEL VERSION::  2.4.20-28.8
  GCC VERSION::  3.2 20020903
  SVGALIB VERSION::  1.9.19
  LINRAD VERSION::  01.25

  Hope that is of some interest, and

  73, Roger Rehr, W3SZ
  
  http://www.nitehawk.com/w3sz



                         -------------*********-----------------

   Oct 14 2004


  Hello, All!

  This is just a brief report on my experience with Linrad in the EME  contest this past weekend.  
  I have used Linrad each year beginning in 2000  for the annual ARRL EME contest.  For the last 
  several years it was my  primary receiver.  But this weekend, the first leg of the 2004 ARRL 
  EME  Contest, was the first time that I used it for EME with the  Antennspecialisten Linrad 
  hardware.  Like last year, I did not even have a  conventional receiver online.  Linrad was it 
  and nothing else was needed.

  I was very pleased and impressed.  I worked only CW for the contest, but I  did take some time 
  out to go up above 144.100 and listen to some EME JT65b  using Linrad and the Antennspecialisten 
  hardware as the front end for  WSJT.  What were the advantages of using Linrad in this manner? 

  [1] I could immediately see any JT65b station that came on the air over a 96 KHz  frequency span. 
  [2] I could take advantage of Linrad's noise reduction. 
  [3] The Linrad waterfall was  superior to Spectran's in my hands for seeing  the stations.
  [4] Using Linrad's click and point frequency control, I was  able to set, immediately and with no 
     trouble, the receive frequency so  that there was zero frequency offset for WSJT to deal with.
  [5] The receiver seemed very sensitive and I had no dynamic range problems.

  On cw, I found [as before] that quite a number of stations that I heard  very clearly with Linrad 
  replied with 'QRZ' for a long time [sometimes  forever] to my calls.  That tells me that Linrad 
  with the  Antennspecialisten hardware hears very well.  I also found that when I  called CQ, 
  I was able to work the station calling me every time.  So I  believe I heard everyone who called me, 
  but not everyone whom I heard  clearly was able to copy me.  I was running 1500 watts, and 
  my 2 x 2  2mxp32 array has a theorerical gain of 21.8 dBd, so I had a reasonable  ERP.  I could see 
  and hear my own echoes with no problem.  Using Linrad, I  was clearly 'transmitter limited' in 
  terms of making contacts.  This is  where I want to be.

  I kept the 'smart' and 'dumb' noise blankers on [manual] nearly the whole  time, and I was not 
  bothered by noise.  When I turned them off I was  reminded that when my rotors are rotating they 
  generate a lot of noise,  totally obliterating the EME signals.  Their noise had been totally  
  eliminated by the noise blankersa and as I was never looking at the  wideband spectrum while 
  operating the rotors, I had been oblivious to the  noise.

  My cw parameters gave me a 5.3 second receive delay, which was not a  problem for me.  I learned 
  to watch the wideband spectrum as 'the end of  the minute' approached, and to start transmitting 
  when the spectral peak  of the station I was working disappeared.  

  These parameters were:

   First FFT bandwidth (Hz) [100]
   First FFT window (power of sin) [3]
   First forward FFT version [5]
   First FFT storage time (s) [20]
   First FFT amplitude [1000]
   Enable second FFT [1]
   First backward FFT version [1]
   Sellim maxlevel [6000]
   First backward FFT att. N [6]
   Second FFT bandwidth factor in powers of 2 [5]
   Second FFT window (power of sin) [2]
   Second forward FFT version [2]
   Second forward FFT att. N [10]
   Second FFT storage time (s) [20]
   Enable AFC/SPUR/DECODE [1]
   AFC lock range Hz [150]
   AFC max drift Hz/minute [100]
   Enable Morse decoding [0]
   Max no of spurs to cancel [0]
   Spur timeconstant (0.1sek) [5]
   First mixer bandwidth reduction in powers of 2 [4]
   First mixer no of channels [1]
   Baseband storage time (s) [100]
   Output delay margin (0.1sek) [5]
   Output sampling speed (Hz) [6000]
   Default output mode [1]
   Audio expander exponent [3]
   A/D speed [96000]
   Check [1110107]



  My parameters for JT65b [ssb mode] were:

   First FFT bandwidth (Hz) [83]
   First FFT window (power of sin) [2]
   First forward FFT version [5]
   First FFT storage time (s) [10]
   First FFT amplitude [1000]
   Enable second FFT [1]
   First backward FFT version [1]
   Sellim maxlevel [6000]
   First backward FFT att. N [6]
   Second FFT bandwidth factor in powers of 2 [2]
   Second FFT window (power of sin) [2]
   Second forward FFT version [2]
   Second forward FFT att. N [7]
   Second FFT storage time (s) [10]
   Enable AFC/SPUR/DECODE [1]
   AFC lock range Hz [150]
   AFC max drift Hz/minute [100]
   Enable Morse decoding [0]
   Max no of spurs to cancel [100]
   Spur timeconstant (0.1sek) [5]
   First mixer bandwidth reduction in powers of 2 [4]
   First mixer no of channels [1]
   Baseband storage time (s) [10]
   Output delay margin (0.1sek) [5]
   Output sampling speed (Hz) [10000]
   Default output mode [1]
   Audio expander exponent [3]
   A/D speed [96000]
   Check [1110107]

  The delay using these parameters was less than one second.

  I played around with the CW parameters to get good performance with  reasonable delay times, but 
  I did not play around with the SSB parameters  as I didn't need to do so.

  For the second leg of the contest I am going to actually hook up a  conventional receiver, so that 
  I can again directly compare the Linrad  software and hardware to the conventional receiver, much 
  as I did in 2000  and 2001 when using my own front end.  I am doing this out of curiosity,  not 
  because I need anything besides Linrad.  I will be using an SSB  Electronics LT2S-Mk V along with 
  either my Elecraft K2 or my FT1000MP.  If  I get the narrow roofing filter for the MP in time, 
  it will be the one I  use.

  I can't wait for the second leg of the contest.  The combination of the  Linrad hardware and 
  software made this EME contest the most enjoyable one  I've had since my first getting on EME in 
  1998, and I can't wait to spend  another EME contest weekend with the combination.

  The Lin-2-1000MP add-on to make a Linrad transceiver out of my FT1000MP  and Linrad worked fine 
  except for those occasions when I forgot to type  'q' to bring the MP on-frequency before 
  transmitting.

  73,

  Roger Rehr,  W3SZ


  On Feb 1 2005,  Roger put out the following:

  Hello, All,

  Those of you who have read my webpages over the years know that I had  previously used Linrad a bit 
  as a sort of panoramic adapter for microwave  work.  [I think there is still an old picture from 2 
  or 3 years ago of the  W3CCX beacon at 2.3 GHz on my website somewhere].
  But when I moved things to the new QTH this fell by the wayside.
  My setup for this was not ideal, as my receive chain at that time  sufferred from too little signal 
  at the input to my then homebrew Linrad  converter, which was just a TUF-1H mixer.

  I had planned to get Linrad running for monitoring the microwaves for the  VHF Contest just ended, 
  but I ran out of time to do so.

  This weekend I finally got around to doing it.  I did it in a sort of  jury-rigged way, but it works 
  GREAT, and relaying that positive message is  the purpose of this note.
  I have all of the electronics for 2.3 GHz thru 24 GHz inclusive on the  tower, and I downconvert 
  to 28 MHz on the tower and just run a single 28  MHZ IF cable down the tower to the shack.
  I wanted to use the Antennspecialisten Hardware that I have in place, so I  made a simple 28 MHz
  to 144 MHz upverter using a TUF-1H.

  I put a Cougar amplifier in the 28 MHz portion of the microwave receive  chain in the shack just
  before the TUF-1H, to give me a 10 dB signal boost.

  I put a Minicircuits splitter in the receive line after the Cougar  amplifier, and fed half the
  signal to the FT1000 and the other half to the  Upverter and Linrad.

  I of course calibrated the Linrad system again after installing the  upverter, using a pulser
  etc. as outlined on Leif's webpages, and feeding  the pulser etc. into the input of my Upverter.

  My shack is on a nice hill, but in the middle of a forest, and with the  towers down I can never
  hear [or see on Spectran, etc.] any of the W3CCX  beacons which are about 50 miles distant,  due
  to the attenuation of  looking thru the forest.

  However, with Linrad I was able to easily see the distant and otherwise  invisible 3456 MHz
  beacon on the waterfall, and to then copy its CW with a  20 Hz filter.  By using the Lin2FT
  program to bring the FT1000 on  frequency, I could select the narrowest CW filters on the FT1000
  and turn  on its DSP and with the DSP 120 Hz or 60 Hz filters also copy the beacon  using the
  FT1000. But I never would have been able to find it without  Linrad.  I came back and did the
  same thing the next day as it was so  unbelievable.  I had done this experiment probably 100
  times without  Linrad and NEVER detected anything!

  I was really amazed by this demonstration and wanted to relay it to the  group.

  I can't wait for the next Microwave contest to use Linrad to see 90 KHz of  microwave spectrum at
  a time.

  PS  I found that most of my system birdies are 'fixed' and don't move when  I change band segments
  with Linrad.  I could therefore cycle up or down 25  KHz in Linrad to see
  which signals to ignore.

  I still haven't installed the STAR ground at the new QTH.  I know I should  do that but just haven't
  had time yet.  Even so, my center discontinuity  is not a significant problem.

  73,

  Roger Rehr, W3SZ 

  On March 20 2005, Roger added the following:


  Hello, All,

  I was able this weekend to get on 144 MHz EME outside of contest time for  the first time since 2002 
  [though I have made the ARRL contest every  year], except that I was able to get on very briefly on 
  2 occasions at  Holiday Time on 12/19/04 and 1/2/2005.

  After being 'gone' for quite a while outside of contest periods, I was  really pleased [and pleasantly
  surprised] with the large amount of  activity I found on 144 MHz EME. I certainly worked only a small 
  fraction  of the stations on the air and workable this weekend, but even so was able  to garner 14 
  Initials and 24 contacts total.

 I think all of this activity speaks well for 144 MHz EME.  Seeing all of  this activity certainly had 
 a positive effect on me in terms of making me  want to work hard to try to get some more free time to 
 do it again before  another 2.5 years pass  ;)

  Here is a list of stations I worked, all JT65b:

 Initials:

 DL8YHR
 SM5CUI
 EA5SE
 F8DO
 OZ1LPR
 ZS6WAB
 KD3UY
 5B8AD
 IK1UWL
 UA4AQL
 DL8GP
 DL7UAE
 K7MAC
 VE1RG  300 watts, single yagi

 Worked but not initials:
 W5UN
 EA3DXU
 WA8CLT
 S57TW
 S52LM
 AA7A
 KB8RQ
 RU1AA
 RK3FG
 N5BLZ

 My station was as before 2 x 2 2mxp32 with both H and V receive channels  simultaneously
 fed into the Linrad receiver the output of which was fed to  JT65b, LZ2US 2 x GU74b
 amplifier, loafing along at 600-1200 watts out  depending upon the capability of
 the station at the other end.

 Initially I had some problems with the Linrad / JT65b interface only  because of
 my stupidity;  on the very weak LT65b signals Linrad cannot  figure out the
 receive polarization angle, and so although there may be a  waterfall trace
 on Linrad, no audio will get to JT65b if the polarization  angle set by Linrad
 is orthogonal to the actual angle.  Before I realized  this I of course got
 very inconsistent results with JT65b.  After I figured this out the receive
 system was VERY SOLID with consistently good  decodes when I couldn't see
 the signal even with Linrad.

 So after I realized I needed to go to manual receive polarization angle
 control all was well.  On those signals I could see, I merely adjusted the
 receive polarization angle using the high resolution spectrum in Linrad so
 that the received signal was green and as large as possible, and the  purple
 [orthogonal] signal was minimized.  When I couldn't see or hear the  signal
 at all, I first set the angle to 0 and if no signal was received  went to 90
 degrees.

 I use a Linrad setup that gives me less than 800 msec receive delay, and I
 had no occasion where timing issues prevented QSO, although several  stations
 had dTs right at or close to 5 seconds.

 I very much liked exchanging grid squares with every station [as is  standard
 with jt65] in addition to the usual 'O', and also appreciated the  'over the
 air' signal reports in WSJT dB units I received in some cases, although not
 being at all experienced in JT65 I really worried the first time I saw from
 the waterfall that the signal sent had gone from an  obvious ''R'' the last
 transmission to a text transmission currently. But then when a signal report
 in WSJT db units popped up in the JT65 window  along with a text message I
 was very happy ;)


 I had meant to do some CW also, but I was so busy in the upper part of the
 band with the many JT65b stations that were New Stations for me that I  didn't
 get the chance.

 Thanks everyone for all the activity and Keep up the Good Work!

 73,

 Roger Rehr
 W3SZ