The Weak Signal Propagation Reporter Network is a group of amateur radio operators using K1JT's MEPT_JT digital mode to probe radio frequency propagation conditions using very low power (QRP/QRPp) transmissions. The software is open source, and the data collected are available to the public through this site.

UPDATE Feb 26, 2009
===================

I've added a couple of additional features to the QRP WSPR transmitter.

1. Completed the switched attenuator so that I can send at 23 dBM, 17 dBm and 10 dBm under control of the PIC.
2. Added GPS support (see below)

I use the GPS to get time of day information and also an accurate 1 pulse per second clock. This is used to update the time of day and to calibrate the symbol clock for WSPR (generated from the internal soft clock generated from the 32.768 Khz (nominal) crystal.

Being paranoid, I also added a thermal fuse to my "oven" so that power to the heater is removed (permanently!!) if the temperature inside the enclosure gets to 72C.

I got the TX on the air just after sunset local time today so missed the bulk of the day's propagation - that said, my first spot was from WA2YUN!

Look forward to getting spots - especially at the lower power levels!

ORIGINAL STATUS
===============

With thanks to Gene W3PM and indirectly to Johan SM6LKM, I am now on the air with a homebrew WSPR transmitter running 0.5 watts (antenna = 160m inverted L with remote tuner) on 30m.

Just prior to 11.00Z my Multimode MEPT was off air for a few minutes for an Antenna change from the 6m Vertical whip to the low, sloping 10MHz 1/2Wave Dipole.

With apologies to DG0OPK if you were doing any tests.

I have changed my current set up for WSPR (other than my Multimode on 10.140073 MHz) in order to reduce power consumption, noise and to utilise Solar Power as much as possible. This will of course help considerably with Portable operation.

The Radio in use is the Yaesu FT-897. The computer the Samsung NC10 Netbook.

I can't believe it. Looking at the map. All that activity and i'm not even hearing anyone "beaconing". Yes i have a 30 meter dipole up and it works well. I hear the Propnet folks and other signals but i'm not getting anything with WSPR or WSJT 6. An occasional "funny" catch and i see the signals on WSJT 6. Only thing i get with that is an occasional "73", RRR or RO. Can propagation be that bad? In the past months i've collected pages of spots on WSJT 6, now i haven't spotted anything at all. Like waiting for good propagation to come around, i may not have any hair left by then.

Mmm!, thought of using the HB9CV for WSPR on 2, but decided to see what action there is on UI-View lately instead. Used to have it running years ago with the KW D-700, but had to flog it when financially embarrassed *sigh*. Anyway, seing as the IC-E2820 has only entertained one Digital Voice for over a year now (No repeaters in range), decided to use it on 144.800 with the old PK-12 tnc, plus the HB8CV of course. Since the new Kenwood has been around for a while now, it would appear that UI-View has had a resurgeance. Loads of peeps on the screen, i see now. This is running on a seperate computer to the WSPR set-up incidentally.

Also on that computer, using a switchable serial port, is Radiocom 6 driving an IC-817 for SWL'ng. Invariably i leave that decoding NAVTEX which is built in the Radiocom software. It also has SSTV and PSK capability, but honestly, it is not a patch on MMSSTV and Multipsk/DM 780.

The WSPR computer is also setup to drive a PCR-100 with Radiocom 6/Ham Radio Deluxe, BCTOOL driving a Uniden UBC3500XLT scanner, or SBS1 virtual radar.

Not connected to computers are UBC780XLT/UBC3300XLT scanners and a pristine condition FRG-7.

I think the concept of the Wednesday Special Activity Day (SAD) is excellent and up to now has been very successful in encouraging stations to try bands other than 30m. I do think however that it could be far more useful if it were extended into activity periods of say 3 days at a time over Friday, Saturday, Sunday for instance. This would allow operators who are at work during the week to take part and maximise activity.

Some stations are already doing extended transmit and receive antenna experiments, some are involved in propagation research, others in qrpp activity. All of these would benefit from extended periods of activity to obtain reliable results. It might also allow stations to perhaps try out different antennas during the activity period in the knowledge that the activity would be there again the next day/night.

The following list of dates, allows for activity on mainly darkness bands like 160m and daylight bands like 10m to take place during the same activity period.

Extended 3 day activity periods

6 – 8 February 160m/10m
13 – 15 February 80m/12m
20 – 22 February 40m/15m
27 Feb – 1 March 60m/17m
6 – 8 March 160m/20m
13 – 15 March 80m/15m

I'm back home and things are back to normal.
Sorry if some of my "unnormal" experiments with the WSPR and the WSJT7 did upset somebody.
Being without internet made me try to communicate directly in WSPR mode, probably not a sin, but still not normal behaviour!
I will not make it a habit.

Two problems that I would like to mention:

1. When using WSPR only p/ calls like EA8/LA3JJ will not work.
Solution A: To use the home call and in addition send complete CW ID.
First days I used MixW and WSPR simultaneously, sending CW ID from MixW at higest rate in the last seconds of each timeslot.
This was a procedure that required manual timing so I looked for other means.

Solution B: To use WSPR and WSJT7 simultaneously sending DE EA8/LA3JJ in some of the WSPR Rx slots.
This required less timing effort and as an extra bonus WSJT7 does have a built in CW ID generator (at 800Hz).
I also tried using WSJT7 as a Tx robot sending all the required WSPR information which worked fine except from uploading of received messages.

Dear friends - men and robots,

The development of our hobby is amazing, some new things will catch on very fast.
I was thrown into this WSPR thing late November last year and must admit that I have not had many “real QSO’s” since then. And my wife also appreciates this as I can in a greater extent be a part of the real world while I am enjoying my hobby. She noticed the change, so I explained some of the changes. “Oh, so you are just talking to robots now, well that’s a strange development!”
It made me think, would be better to know if there is actually a person somewhere in the other end.
May be WSPR QSO mode is something for me
Tried it for some days now and it is not easy. Only correct answer to my CQ EA8/LA3JJ call was from VE9DX and it was hash coded as it should be. But conds changed and I took to long…
It seems that everybody else are in the same learning curve as myself.
To you all: read the manuals carefully and understand about hash coding and limitations of messages before you go on. I have learned a lot through trial and error the last days and hope I will soon be able to make WSPR QSO’s efficiently.

When I first discovered WSPR I had an initial burst of great enthusiasm for it. I still think it is a great mode. It is fun to see how far signals usng low power or a very small antenna can be received.

But I have really lost most of my enthusiasm for actually using WSPR these days. Why? Because virtually all of the activity is concentrated on 30m, to the extent that there often isn't a sufficiently large number of stations monitoring any other band to guarantee the chance of a report.

I supported, indeed was involved in the inception of the idea of special activity days on other bands. These have been pretty successful in encouraging more stations to use other bands. But the rest of the time everyone goes back to 30m. Why? Personally I'm bored with getting the same old reports from the same old places on the same old band. There is nothing new to do on 30m as far as I am concerned.

On Saturday, Jan 10th, I had my transmit-only beacon that I wrote sending both a visual MEPT indicator (the MV that you can see in the attached image) and the WSPR beacon, and tried out a simple program that simultaneously records, and later makes an FFT display of that portion of the band.

It seems to work!

Recently installed the ICOM FL53 (250HZ) CW filter in the IC703. This works very well
and it's passband sits nicely centered at 1500 hz with zero IF Shift. This is very
convenient for WSPR QSO mode in USB with WSJT 7 as this is the default center for this mode.

I will be operating from my QTH on the island of Fuerteventura from January 13 - January 25.

This is a long overdue project.

Using the downloadable CSV data, it's pretty trivial to load the data into R for some decent analysis. I was able to load all 3+ million spots into a data frame, and I'm going to start to try to learn some things. My first pictures are trivial histograms of Received SNR, Transmitted Power, and SNR-Power (un-normalized path loss). These are across all 30m spots to date.

Documentation and software downloads of WSPR programs and other WSJT-related modes of Joe Taylor, K1JT, are available from the WSJT web site. The entire database of millions of WSPR spots is available for public download in comma-separated value (CSV) format compressed with the gzip and zip utilities. Each file contains the data for one calendar month of spots. The data begins with the initial release of WSPR in March, 2008. The file for the current month is updated automatically once per day in the early morning UTC, and the monthly historical files are re-generated once every month. Compressed file sizes range from 1-20MB.

2008-03 gz zip
2008-04 gz zip
2008-05 gz zip
2008-06 gz zip
2008-07 gz zip
2008-08 gz zip
2008-09 gz zip
2008-10 gz zip
2008-11 gz zip
2008-12 gz zip
2009-01 gz zip
2009-02 gz zip
2009-03 gz zip
2009-04 gz zip
2009-05 gz zip
2009-06 gz zip
2009-07 gz zip
2009-08 gz zip
2009-09 gz zip
2009-10 gz zip
2009-11 gz zip
2009-12 gz zip
2010-01 gz zip
2010-02 gz zip
2010-03 gz zip
2010-04 gz zip
2010-05 gz zip
2010-06 gz zip
2010-07 gz zip
2010-08 gz zip
2010-09 gz zip
2010-10 gz zip
2010-11 gz zip
2010-12 gz zip
2011-01 gz zip
2011-02 gz zip
2011-03 gz zip
2011-04 gz zip
2011-05 gz zip
2011-06 gz zip
2011-07 gz zip
2011-08 gz zip
2011-09 gz zip
2011-10 gz zip
2011-11 gz zip
2011-12 gz zip
2012-01 gz zip
2012-02 gz zip
2012-03 gz zip
2012-04 gz zip
2012-05 gz zip
2012-06 gz zip
2012-07 gz zip
2012-08 gz zip
2012-09 gz zip
2012-10 gz zip
2012-11 gz zip
2012-12 gz zip
2013-01 gz zip
2013-02 gz zip
2013-03 gz zip
2013-04 gz zip
2013-05 gz zip
2013-06 gz zip
2013-07 gz zip
2013-08 gz zip
2013-09 gz zip
2013-10 gz zip
2013-11 gz zip
2013-12 gz zip
2014-01 gz zip
2014-02 gz zip
2014-03 gz zip
2014-04 gz zip
2014-05 gz zip
2014-06 gz zip
2014-07 gz zip
2014-08 gz zip
2014-09 gz zip
2014-10 gz zip
2014-11 gz zip
2014-12 gz zip
2015-01 gz zip
2015-02 gz zip
2015-03 gz zip
2015-04 gz zip
2015-05 gz zip
2015-06 gz zip
2015-07 gz zip
2015-08 gz zip
2015-09 gz zip
2015-10 gz zip
2015-11 gz zip
2015-12 gz zip
2016-01 gz zip
2016-02 gz zip
2016-03 gz zip
2016-04 gz zip
2016-05 gz zip
2016-06 gz zip
2016-07 gz zip
2016-08 gz zip
2016-09 gz zip
2016-10 gz zip
2016-11 gz zip
2016-12 gz zip
2017-01 gz zip
2017-02 gz zip
2017-03 gz zip
2017-04 gz zip
2017-05 gz zip
2017-06 gz zip
2017-07 gz zip
2017-08 gz zip
2017-09 gz zip


CSV File Format The fields in the CSV file are as follows:
Spot ID
A unique integer identifying the spot which otherwise carries no information. Used as primary key in the database table. Not all spot numbers exist, and the files may not be in spot number order
Timestamp
The time of the spot in unix time() format (seconds since 1970-01-01 00:00 UTC). To convert to an excel date value, use =time_cell/86400+"1/1/70" and then format it as a date/time.
Reporter
The station reporting the spot. Usually an amateur call sign, but several SWLs have participated using other identifiers. Maximum of 10 characters.
Reporter's Grid
Maidenhead grid locator of the reporting station, in 4- or 6-character format.
SNR
Signal to noise ratio in dB as reported by the receiving software. WSPR reports SNR referenced to a 2500 Hz bandwidth; typical values are -30 to +20dB.
Frequency
Frequency of the received signal in MHz
Call Sign
Call sign of the transmitting station. WSPR encoding of callsigns does not encode portable or other qualifying (slash) designators, so the call may not represent the true location of the transmitting station. Maximum of 6 characters.
Grid
Maidenhead grid locator of transmitting station, in 4- or 6-character format.
Power
Power, as reported by transmitting station in the transmission. Units are dBm (decibels relative to 1 milliwatt; 30dBm=1W). Typical values are 0-50dBm, though a few are negative (< 1 mW).
Drift
The measured drift of the transmitted signal as seen by the receiver, in Hz/minute. Mostly of use to make the transmitting station aware of systematic drift of the transmitter. Typical values are -3 to 3.
Distance
Approximate distance between transmitter and receiver along the great circle (short) path, in kilometers. Computed form the reported grid squares.
Azimuth
Approximate direction, in degrees, from transmitting station to receiving station along the great circle (short) path.
Band
Band of operation, computed from frequency as an index for faster retrieval. This may change in the future, but at the moment, it is just an integer representing the MHz component of the frequency with a special case for LF (-1: LF, 0: MF, 1: 160m, 3: 80m, 5: 60m, 7: 40m, 10: 30m, ...).
Version
Version string of the WSPR software in use by the receiving station. May be bank, as versions were not reported until version 0.6 or 0.7, and version reporting is only done through the realtime upload interface (not the bulk upload).
Code
Archives generated after 22 Dec 2010 have an additional integer Code field. Non-zero values will indicate that the spot is likely to be erroneous (bogus callsign, appears to be wrong band, appears to be an in-band mixing product, etc. When implemented, the specific codes will be documented here.

Hi all,

I looked in the database today and found sometimes a strong difference in receiving and transmitting level of some stations over the last months.

Can it be that the spotted TX power level in the database is not correct....

I am using ERP Scale ...5W ERP means around ...37dBm

73 de michael DG0OPK

Pages