Wednesday, April 30, 2008

WSPR - 160m spots analysis, or how to evaluate a station performance.

For a contester it is always interesting to know how good is the site he is using, what kind of handicap or bonus he has vs a direct competitor, or what is the real impact of an antenna improvement (ie adding more radials on a vertical). Everybody knows how difficult it is to base and antenna evaluation on a few subjective S-meter reports, affected by QSB. Providing a uniform and objective reporting system (Signal to Noise Ratio), WSPR is the ideal tool, capable of providing near to the dB accuracy when enough spots are averaged. Furthermore it allows to evaluate antenna and site, accross a full window from band opening to band closure, this function of the direction and distance.Following are a few examples. Above is the plot of the spots reported by DF1VB on 160m for F1VS and F6IRF stations. As usual all reports have been adjusted taking into account the TX-power (all reports are matched as if the TX was always 1W). Of course the distance being shorter for my station(896 vs 631 kms), it is quite normal to see such a difference. GI8HXY having provided a significative number of spots for F1VS and my station during two 160m recent nights, I repeated the experience. This time the distance difference is not that important, and in favour of F1VS, but the signal difference is slightly higher in my favour. This may be due to my house obstruction in DF1VB direction, while the path to GI8HXY is perfectly clear with just a 3 degrees horizon blocking due to the mountains (which has no effect on such a path on 160m). Of course there are many more parameters which may explain this offset (I don't even know what kind of antenna F1VS is using...)Looking at VE1VDM reports, the difference is this time minimal over time, but we see clearly the effect of the later sunrise at F1VS location. It would have been interesting to see earlier reports, but I was running too low power until 2:00 to get my signal decoded on VE1VDM side, and unfortunately Vern was not QRV last night.Back to GI8HXY reports, I plotted the most spotted 4 stations (data only for last night). On the above plots the SNR's are just corrected for power levels. Despite different data time-range, the difference between F1VS and my reports is still the same (about 6dB). OH3XR stoppped transmitting around 3:30, which explains the abrupt and early signal drop.On the above plot I have tried to correct the distance effect. For this I used VOACAP path-loss model as a function of the distance (the propagation mode being identical for the 4 stations: 1F2) and deducted a correction factor proportionnal to the distance from Transmitter to Receiver. It is amazing to see the performance of OH3XR station vs all others (to be honnest I did not remember who OH3XR was, when I did that... I just selected the 4 more spotted stations, and Marko was transmitting 500mW, so is signal level did not alert me before I plotted the above !).
I just looked at Google, and quickly found that OH3XR aka OH9XX is a famous topbander and contester and owner of an exceptionnal contest site. I have no idea what antenna Marko was using last night, but it is interesting, isn't it ?

WSPR: Evaluation of two160m RX-antennas using WSPR reports

The following examples focus on the 160m band, where the WSPR activity is increasing. Last night more than 30 stations were active on the band, which was in excellent shape, with very low level of statics. The EU map show the active stations between 21 and 22z, while the map covering EU and USA show the active stations between 2:30 and 3:30Z. Using the automatic antenna switching described in the previous posts, I took this opportunity to try to evaluate the performance of my two RX antennas. Although the number of spots considered is somewhat limited it already gives interesting indications. For the following plots, all reports have been normalized to 1W (deducting/adding the corresponding dB's when more or less power was used).The above graph shows the performance of my RX magnetic loop (in red) vs my K9AY (in blue) for a 850 kms circuit. Both antennas are pointing north (direction where my power line noise is minimal). Both antennas have a broad pattern, with a min E/W for the loop, and a min to South for the K9AY. For this direction/distance the magnetic loop is about 2dB better than the K9AY.As shown on the above, increasing the distance, tends to reduce the difference (the above plot is for a 2000 kms path) . Due to the limited number of spots from USA (especially on the loop), the plot would be meaningless, but it seems that the K9AY is working better at a longer distance, as more than 2/3 of the US spots were obtained using the K9AY. Will come back, on this when enough data is available. The above graph, shows the directivity of the loop, as F1VS is moreless located on the deep of the loop pattern and at a distance of 450kms. I was wondering if the directivity of the loop was still good for signals arriving quite high above the horizon... the answer is obviouly yes! (the K9AY has about 3dB front to side), in this direction the K9AY is obviously better (btw I have no idea how to explain the sinusoidal shape of the loop curve... but considering the limited number of spots, it may just be a random effect...). Those results are obviously only valid for my location, where my main problem is to eliminate the noise generated by the very High Voltage power line... The moon before moonset. The line is running East-West and located in my South. On 160m, the minimum noise is obtained with the magnetic-loop perpendicular to the power line (N/S). The noise floor increases by some 10dB when the loop is oriented E/W.

Friday, April 25, 2008

A statiscal method to evaluate TX antenna performance using WSPR - more...

The following graphs are based on a significative number of spots and show, what can be expected from the method described in the previous post. All have been done on 30m, as for the moment it is the band where the WSPR activity is maximum. The above graph, shows all the spots (about 2500) collected from 23 april 10:00 to 25 april 5:00 for the 2 antennas under test (5/8 vertical and dipole at 10m). Such a graph may look useless but similar plotting may be of interest for contesters, who want to reach a maximum of stations in all directions at all distances in a limited period of time. The slight superiority of the vertical can be easily explained by the fact that it is omnidirectionnal, thus allows a stronger signal in all directions, even at short distances. This would look probably different on lower bands, where high angle signals are more dominant for non DX stations. It would also probably look different with a 1/2 wave vertical which has a sharper pattern in the vertical plane and unlike the 5/8 no secondary lobe at high angle (see the article describing my vertical ).
For the above graph, I only used the data from DX stations (>5000kms - about 650 spots). It should be noted that the orientation of the dipole (310/130) is near to optimum for a huge majority of the spotters. It just confirms antenna simulation and what all experienced DX'ers and contesters know: For DX, a well designed verticaly polarized antenna is better than a horizontaly polarized antenna, if below 0.5wl AGL.
By the way an interesting phenomena can be observed on day 2 between 14 and 20z, when the dipole seem to exhibit slightly better performance than the vertical. It could be explained, by some higher angle mode of propagation (ie E's+F) during this particular period of time... but of course this is just an hypothesis...

Thursday, April 24, 2008

A statiscal method to evaluate TX antenna performance using WSPR

Failing to find a proper calibration for VOACAP to use it as reliable reference to evaluate antenna or site performance (*), I decided to try something else. The method I imagined consists in switching 2 antennas every 10 minutes, 1 second before minutes 00, 10, 20, 30, 40 and 50. Selecting a 33% TX'ing rate, in WSPR you have the guarantee to have a minimum of 1 transmission per 10 minute period (most of the time 2). Swiching is done using a little software available for free on the web and a simple interface to the LPT-port using a NPN transistor to drive a coaxial relay (similar to the popular CW interface used for K1EA's contest sofware "CT").
Following are a few examples, based on limited number of 30m spots, which however seem to validate the method. The antennas used are 1) a 5/8 ground-mounted full size vertical with about 30 burried radials with a simulated gain of 0.5dBi @16 degrees 2) An inverted V dipole, with appex at 10m oriented for AZ 130/310, with gain of approx 5 dBi @50degrees.
The first graph show the spots received by VK6DI on April 23. The distance is 14000 kms, the azimut from here, is about 100degrees (so about 30 degrees from the dipole main lobe). For this circuit, and at the time of the spots VOACAP indicates a vertical angle between 3 and 10 degrees (function of the time). All the spots have been corrected, to show a normalized TX-power of 1W (I deducted the corresponding 6dB's, when 4W was used). 4 spots received by VK6DI between 2140 and 00:48 were removed, as they were received only when I was TX'ing on the vertical (leaving those spots would have biaised the comparison). The superiority of the vertical is quite obvious, and peaks at 6dB just before 20z, with an average of 4dB across the window. This is quite normal, considering the gain of the dipole at angles below 10 degrees.
The second plot is from the spots received by K1JT. The distance is 6300 kms and the azimut from here is 296 degrees. VOACAP indicates a radiation angle around 10 degrees for the considered window. The difference peaks to 7dB in favor of the vertical, with an average 3dB accross the window.The 3rd plot is based OY3JE's spots. Distance is 2000kms and the AZ 340 degrees. Here the average received signal is about the same on the 2 TX antennas, but with higher peaks on the dipole.
Reducing further the distance gives a slight advantage to the dipole. Above are GM3AKF spots, at a distance of 1400 Kms and AZ of 338 degrees.
Finaly are the spots received by EA1FAQ at a distance of 1000kms and a bearing of 245 degrees. Here the vertical takes again the advantage, likely due to the directivity of the dipole.
As mentionned earlier, the above plots are based on a limited number of spots, the method being statiscal, more spots, more accurate results... The results may also be biased by the interaction of the 2 antennas only located 15meters away. Nevertheless, the results seem quite coherent with the antenna simulation, and show that WSPR and the network of stations that is growing every day is a fantastic tool, not only for propagation evaluation, but also for antenna and site evaluation (ie using 2 stations a few kms away from each others using the same antenna). The same method could also be used to evaluate RX antenna perfomance.
(*) important differences can be seen function of the selected circuit, so evaluating an antenna or a site with sufficient accuracy, even over a period of 3 weeks was a kind of utopy...

Monday, April 21, 2008

VOACAP vs WSPR reports - part2

My previous post on the topic concluded by "I leave you judge...". In fact, there was a trap: VOACAP provides SNR in 1Hz bandwidth, while WSPR provides SNR in 2500Hz bandwidth, thus a difference of 10Log2500=34dB. Quite a difference ! Following are a few more plots, in order to try determining the correct "VOACAP" calibration for WSPR. For this purpose I have used the data collected on my 30m beacon, by W1BW (near to 400 spots). All reports have been normalized to 1W ERP (I used up to 100W at band closure / opening), which explains why a few reports are below -30dB. The first plot shows, on top all the spots collected by W1BW between 1st and 20 April (the dark blue dots). The light blue curve is a simple average on 4 spots and the blue curve the polynomial order 6 function. The other curves represent the VOACAP output for the parameter SNRxx (SNR at required reliability), adjusted for 2500Hz BW, for a circuit reliability of 90%, 50% and 10%. Other parameters: Power 1W, Isotropic Antennas, min angle 3deg, Ant gain 0dBi (both TX/RX assuming omni antennas are used at each end) , Noise -152dBW (CCIR quiet) , Method 30, Absorption "Normal", Coefficients CCIR (monthly) and SSN =15 (value suggested by G4ILO's VOAprop). The "required SNR" was set to 4dB (-30dB min required signal for WSPR to decode, plus 34dB for the bandwidth relation) but this parameter has no impact on the SNRxx output, and the Fprob. parameters were set to default. Last, but not least, the values near the edge (band opening/closure) and during the period 13 to 16Z should considered with care due to the limited number of spots available. Observations: 1) the SNR and SNRxx @50% provide identical results.2)Setting the required reliability to 10% generated earlier predicted band opening and a later band closure. 3) there is a litle time offset between the observed peaks and the predicted peaks which may be explained by the fact that VOACAP provides the value for the 9th hour, which is in fact 9z to 10z.
This has been corrected on the above plot, which shows basicaly the same things, but where all spots collected in the 9th hour were averaged at 9:00z. This time the required reliability for VOACAP was set to 5%, 10%, 25%, 50% and 75%. We can see the impact of the required reliability parameter: Reducing the xx below 20% has for consequence to predict an earlier band opening a later band closure and a higher level. In fact 50%or higher seems to provide more matching results (except in terms of level, but I'll come back on this). I also wanted to see the impact of the other parameters. First the coefficients URSI88 (daily) vs CCIR (monthly), then the absorption model (IONCAP vs Normal). as you can see, those parameters have little impact on this particular path and none on the maximum levels. I also tried the 3 different methods available: method 20 (complete System performance) method 22 (forced short path model) and method 30 (short long smoothing), but on this particular circuit, the 3 methods provided absolutely identical results.
Finaly I made the SSN to vary from 15 (suggested G4ILO's VOAprop value) to 10 (predicted monthly smoothed SSN) down to 4 (VE3NEA's Hamcap suggested value, from the latest table available on his site). At this stage we can see that, using the recommended / defaults values (method 30, absorption Normal, CCIR monthly coefficients) and a SSN=4 for the SNR parameter output, seems to provide the best match in terms of band opening/closure with the observed values. Remain the 16~20dB difference observed near the 2 maximums... Is VOACAP prediction too conservative, is noise level lower than estimated at W1BW location, does he have some RX antenna gain, is the WSPR reported level too optimistic? Might be a combination of all these factors, or did I do a big mistake somewhere... more to follow and of course comments and suggestions are welcome !

Thursday, April 17, 2008

VOACAP vs WSPR-reports

F6IRF à VE1VDM sur 30m: En couleur, la courbe de tendance obtenue en utilisant VOACAP. A l' intérieur, la courbe obtenue d'après les relevés réels de ma balise WSPR reportés par VE1VDM (les points bleus) pendant la première quinzaine d'avril, tous les reports étant normalisés pour 1W (en ajoutant ou soustrayant les dB's correspondants quand j'avais utilisé une puissance differente ). Ci dessous les paramètre utilisés pour le modèle. VOACAP a la réputation d'être le modèle de propagation ionosphérique le plus fiable disponible à ce jour... je vous laisse juge...
F6IRF to VE1VDM: In colors, the plot generated by VOACAP. Inside, the plot obtained from the WSPR real reports during the first 2 weeks of April (the blue dots; all have been normalized to 1W, adding or substracting the dB's difference to the reports, when a different power level had been used). Below are the parameters used for the modelization. VOACAP has the solid reputation of being the most accurate ionospheric model available today... I leave you judge...

Wednesday, April 16, 2008

WSPR - quelques essais de plus...

Voila le CR de 2 essais réalisés ces dernières 12 heures, et qui donnent une idée plus précise de ce que l'on peut ésperer de WSPR. Traverser l'Atlantique avec 4 mW sur 20m, oui c'est possible! Il en faut un peu plus sur 160, mais ça marche aussi !!!
Ci-dessus - cliquez pour agrandir - le relevé des spots reportés hier soir sur 20m en réduisant la puissance de
100mW jusqu'à 4mW.
Ci-dessous le setup utilisé pour contrôler la puissance délivrée à l'antenne (steppIR 2 elts @12m - Gain simulé 10dBi): le wattmètre (Daiwa CN801) est branché à l'envers pour béneficier de l'échelle la plus dilatée. L'émetteur est réglé pour 200mW , reste ensuite à insérer l'atténuation nécéssaire, jusqu'à 17dB hier soir (23dBm-17dB=6dBm soit 4mW !).

Ci-dessous résultats des relevés reportés par VE1VDM sur 160m cette nuit. La puissance était de 10W et l'antenne ma verticale 1/8 chargée à la base (Gain simulé -3dBi). Vern utilisait quant à lui un long fil de 400m, orienté E/W. Entre le coucher de soleil chez VE1VDM, et le lever de soleil, chez moi, l écart maximum du SNR n'est que de +/- 4dB par rapport à la valeur moyenne et pas une transmission n á été manquée. Après, un peak, la liaison s'est interrompue brutalement peu après le lever du soleil, et même un passage à 100W n'a pas permis de la rétablir.

Tuesday, April 15, 2008

A l 'écoute des rumeurs...

On voit ici en dessus du signal WSPR de OZ1PIF sur 30m, un signal PSK31. A droite mon émission, centrée à 1510Hz (Spectrum-Lab software)

Pour une fois il est intéressant d'écouter les rumeurs ! Non, pas celles concernant les derniers avatars de la vie personnelle de notre président, mais les chuchotements dûs au dernier soft de l'illustre Prix Nobel K1JT, WSPR (prononcez "whisper"et traduisez "Weak Signal Propagation Reporter"). Pour plus de détails le mieux est d'aller sur le site de Joe en suivant ce lien. Vous pourrez y trouver une présentation du programme ainsi que le lien pour charger la version expérimentale la plus aboutie à ce jour.
Comme l'indique l'acronyme, il ne s'agit pas d'un soft pour faire des QSO's, mais d'un soft pour tester les conditions d'une bande, des modes de propagation exotiques, une liaison particulière, voire d'évaluer les performances d'une nouvelle antenne son rapport A/R ou tout ce que vous pourrez imaginer comme application possible...
Le programme de K1JT - on peut voir sur le spectrogramme le signal recu de OY3JE à 10.194. le méssage transmis par ce dernier, inclue son locator (IP62) la puissance en dBm (30dBm soit 1 W) et le décodeur indique un SNR moyen de -24 dB dans une bande passante de 2.5kHz. On voit avec speclab (ci dessous), que ce signal serait juste suffisant pour être décodable en QRSS à 3s par point. On voit d'ailleurs un signal QRSS (FSS ?) entre les signaux de OZ1PIF et de G0HNW

Le signal géneré par WSPR est e
xtrêmement étroit (6.5 Hz de bande passante), la bande passante monitorée étant de 200Hz centrée à une fréquence audio de 1500Hz (avec un 756 pro, pro 2 ou 3, le mode USB-D et le filtre de 250Hz sont idéaux). Outre le TRX vous aurez besoin du même "setup" que pour le PSK31, à savoir un PC muni d'une carte son et une interface audio. (personellement j'utilise mon MK2R+ avec sa carte son USB intégrée), et comme pour le JT65 du même auteur, d' une bonne synchronisation de l 'heure de votre PC (temps internet, GPS, DCF77 ou autre).
Typiquement chaque station du Net, transmets pendant X périodes de 2mn, commencant à une minute paire ( 00-02-04,... X dépendant du T/R cycle sélectionné) et écoute le reste du temps. Si vous sélectionnez un T/R cycle de 20% vous allez transmettre pendant 6 des 30 périodes constituant l 'heure et écouter pendant les 24 autres, le choix des périodes d'émission étant "randomisée" afin que 2 stations données ne transmettent pas systématiquement sur la même période. De cette façon, même 2 stations ayant choisi la même fréquence à l'intérieur de la bande de 200Hz, ont toutes les chances de ne pas se gêner systématiquement pendant leur période démission, et chaque station du net a de fortes probabilités d'entendre toutes les autres.

Mais vous me direz, rien la de très nouveau... en effet j'avais moi-même mentionné ici le réseau propnet (suivre ce lien). La grosse différence, par rapport à ce réseau utilisant le mode PSK31, réside dans le niveau de performance obtenu grace au type de modulation utilisée, qui permets de décoder un signal jusqu`à un rapport Signal/Bruit de -30dB dans une bande passante de 2.5kHz soit presque 20dB de mieux que pour le PSK, et 16 bons dB's en dessous du plus petit signal CW décodable par l'oreille humaine, ce qui situe ce mode légèrement devant ce qu'il est possible d'obtenir avec les modes visuels tels que le QRSS à 3s/point qui étaient utilisés jusqu'à ce jour sur les "Grandes Ondes" (par ex. la bande 137KHz) pour repousser les limites du bruit.
Exemple de carte visualisable sur le site Ci dessous, le site dispose aussi d'une base donnée munie de filtres. On peut y voir mon activité récente sur 80m -cliquez pour agrandir - (5000kms avec 100mW, et il y avait encore plus de 10dB de marge !)

Enfin pour supporter le tout, un site internet permets de reporter les balises entendues (cela se fait automatiquement depuis WSPR depuis la version 0.6), et de visualiser tous les endroits de la planète ou votre propre signal a été entendu. A titre indicatif c'est sur la bande 30m que l'activité "naissante"est la plus soutenue et sur cette bande, une verticale et 1W sont largement suffisants pour se faire entendre aux antipodes tant au long-path le matin, qu'au short path le soir...
D'ores et déjà ce mode offre un large éventail d'explorations passionnantes... et nous n'en sommes qu'au début... (à suivre...).