Skip Excellent observations.
philw de ka1gmn On Fri, Jan 8, 2010 at 6:38 PM, KH6TY <[email protected]> wrote: > > > On UHF and VHF, Doppler spread, often found a very fast rate, causes the > "dots" of DominoEx to "smear" or elongate horizontally on the waterfall. > When there is a slow or large constant Doppler shift, such as when an > airplane is passing overhead, the dot pattern skews from one side or the > other of the main pattern. For PSK63, the parallel carriers slant starting > at one side of the "normal" carrier, and pass through the image of the > "normal" carrier, completely blocking printing when it coincides with the > main signal, even if the signal is strong (the multipath reflection is also > strong...). As it moves to the other other side, printing resumes. We often > see this effect on 144 Mhz when it is a gusty day and gusts blow by. This > makes PSK31 unusable on 144MHz, PSK63 is better, and DominoEX is best. > PSK63F does not seem to be as good as DominoEx, because the signal actually > shifts tone frequency, and is therefore mistuned, but DominoEx can tolerate > a shift of 50% of the signal width and still print. Our tests on 144Mhz and > 432 MHz suggest the PSK63F is about equal to DominoEX 8 in sensitivity, not > as good as PSK31, but that is on VHF/UHF. DominoEx 8 would print when PSK63 > failed under QSB with signals near the noise. > > Tony, a comparison of PSK63F to DominoEX 8 and DominoEX4 on your simulator > would be very appreciated! > > 73 > > Skip KH6TY > > > > > Tony wrote: > > > > > Phil, > > > It would be valuable to the community to be able to recognize the > presence of Doppler spread by some visual or > > aural means. > > There are times when the effect can be detected by sight and sound. Even > a moderate disturbance in the polar ionosphere can cause enough > Doppler spread to make HF voice communications sound hollow or as if the > signal was emanating from under water. The same applies to CW and digital > mode signals. > > Visually, narrow band modes like PSK31 will appear dispersed (spread > out) compared to others in the waterfall. The individual tones that make up > wide-band signals (MFSK) will also appear diffuse. Other visual > indications of Doppler induced frequency dispersions caused by signals > that propagate over the poles, would be the chaotic movement of the PSK31 > tuning indicator. > > Although there are times when Doppler spread is too small to be detected, > it can still cause throughput problems with certain modes. The amount of > throughput loss depends on the severity of the Doppler > spread and how resistant the mode is to that type of distortion. Narrow > modes like PSK10, for example, are especially sensitive to small frequency > spreads (1Hz) and can suffer throughput loss on a relatively quiet > mid-latitude path. > > The reflecting surface of the ionosphere is somewhat similar to the surface > of a pool of water where waves are traveling in all directions. As sunlight > hits the pools surface, the motion of the waves cause the light to scatter > at random. With RF waves, the motion of the > ionospheric irregularities cause random shifts in frequency creating > wavelets around the main signal (see attached: Normal vs. Doppler spread). > > Doppler spread is more pronounced in the active regions of the auroral and > equatorial zones than it is in the mid-latitudes. Its common to see an > increase in the effect on long distance contacts since the signal > will eventually end up near those regions. How much it effects digital mode > throughput depends on the state of the ionosphere and the mode in use. > > As for PSK63F, it combats this type of distortion much better than standard > PSK modes. > > Tony -K2MO > > ----- Original Message ----- > From: Phil Williams > To: [email protected] > Sent: Tuesday, January 05, 2010 8:49 AM > Subject: Re: [digitalradio] IZ8BLY's PSK63F > > > > > Demonstrating a suite of digital coding methods are vulnerable to Doppler > spread does not tell the whole story. What does the signal look like on the > a spectrogram when subjected to Doppler spread? Yes, you have incomplete or > scrambled text, but then the root cause of that could be anything. > > It would be valuable to the community to be able to recognize the presence > of Doppler spread by some visual or aural means. Armed with this > information then one begins to make choices of other modes that would be > less vulnerable to the effects of Doppler spread. > > philw de ka1gmn > > > On Tue, Jan 5, 2010 at 2:13 AM, Tony <[email protected]> wrote: > > > Phil, > > > > What about PSKFEC31 under the same test scenarios? > > Have a look: > > > Path Simulation: High Latitude (Moderate) > Path Delay: 3ms, Doppler spread 10Hz > Pangram Text: Quick Brown Fox > > > > PSK31FEC > > t e tio E ttaeH loo etee- e e e ˆyaooe n o > ao t aeepvede n neete ueeeu .tna0 o een > it=pctidr a ieae t e tio E ttaeH loo etee- e e > etˆyaooe on oe ne 6etnuEenoel o·b geogtee > > > > PSK63F > > the quick brown fox jumps over the lazy dog > the quick brown fox jumps over the lazy dog > the quick brown fox jumps over the lazy dog > > Tony -K2MO > > > > > ----- Original Message ----- > From: Phil Williams > To: [email protected] > > Sent: Monday, January 04, 2010 5:16 AM > Subject: Re: [digitalradio] IZ8BLY's PSK63F > > > > Very interesting. What about PSKFEC31 under the same test scenarios? > Certainly, there would be more a in throughput, but that is a matter of some > liberal use of CW shorthand. > > > philw de ka1gmn > > > > On Mon, Jan 4, 2010 at 2:48 AM, Tony <[email protected]> wrote: > > > All, > > Recent path simulation tests indicate that Nino Porcino's PSK63F offers > better performance over PSK31 and PSK63 in a couple of areas. The most > significant improvement is it's ability to endure Doppler spread found on > paths that cross the polar ionosphere. Both PSK31 and PSK63 fail miserably > in this area; see high-lat test samples below. > > Path Simulation: High Latitude (Moderate) Path Delay: 3ms, Doppler spread > 10Hz > Pangram Text: Quick Brown Fox > > PSK63F -- the quick brown fox jumps over the lazy dog > PSK63 -- mev roe tt#dtorl|f- bn ô mp e o ihe Fzy dg > PSK31 -- nls oSer Òe naAeta qlipM h nV o T rn agâ o > RTTY -- TH QACKH492, FOJUMP OR THTLAZY G > > Sensitivity-wise, it's quite a bit more sensitivity than PSK63, but only > marginally better than PSK31. Although it's speed is about 25% faster than > PSK31, it's about 40% slower than PSK63. Average wmp rate seems to be 63 > wpm > for PSK63F. > > Lowest S/N (sensitivity) > > PSK63F -12db > PSK63 -7db > PSK31 -11db > RTTY -5db > > Additional path tests indicate that PSK31 and PSK63F perform about the same > > under moderate mid-latitude conditions (CCIR fading channel). Tests show > that PSK31 and PSK63F will outperform PSK63 when signals are weak under > quiet conditions since they both have greater sensitivity. > > It would be interesting to hear from our HF digital friends up north who > experience the distorting effects of the polar ionosphere on a regular > basis; this is where the PSK63F mode can be put to the test. > > Available software: > > Nino Porcino's Stream -- http://xoomer.virgilio.it/aporcino/ > Patrick Lindeckers Multipsk -- http://f6cte.free.fr/index_anglais.htm > (thanks for including PSK63F Patrick) > > Tony, K2MO > > > > > > > > >
