70cm WSPR Mobile Experiment

Recently I tried an experiment, I went mobile on 70cm WSPR. I used an Icom IC706mkIIg that I have in the car along with a rubidium 10 MHz frequency reference and tripler to replace the Icom's 30 MHz master oscillator. I ran WSJT-X on an old laptop. The antenna I used was a dual-band vertically polarized whip so it was cross polarized to the signals from the N6GN home station beacon running WSPR, JT65A, JT9 and CW on ~432.301500.

The picture attached shows the waterfall next to a map of the route that the display covered, roughly lined up. The bulk of it is California Hwy 12 on the south side of Santa Rosa, actually running West to East from bottom to top. As shown, North is to the left. My mobile crossed Hwy 101 during the quiet interval after the JT65 transmission.

The waterfall begins recording while waiting at a traffic signal at Fulton/Hwy12 shown, at the bottom of the picture. I then turned east (up on this map) and accelerated to free-way speeds. You can see the multipath components grow around what is the actual transmit frequency of 432.301500. You can also see a "main" path signal about 40 Hz low of this, most obvious at the end of the WSPR interval on the CW signal which is also transmitting at 1500.
This matches my speedometer indication and shows a relative velocity of: 40 Hz => .694m*40=27.8 m/s or 62 mph, away from N6GN.

In the next interval you can see KJ6MKI who is located about 5 miles off the map to the right, transmitting WSPR. His 432.501460 signal is showing up only ~10 Hz low because he is to the south and relative velocity is not nearly so high as for the N6GN signal, which has its QTH marked at the bottom left of the map. Even so, it looks like I might have been going just a little fast, following the traffic speed on Hwy 12. "Yes, I know officer, but it was a scientific study!"

In the last interval, JT9, you can see a nice ACS term cross the line intersecting N6GN and N6GN/mobile. I didn't look at Planefinder to identify the aircraft responsible but from previous experience I'm sure it would have been obvious. It could have been well East or West of Santa Rosa though, possibly 100 km or more away.

At the very end, the mobile is moving North and more or less at right angles to N6GN so the path length isn't changing much, at the very end I park to turn off the WSJT-X waterfall recording. Here you can see the CW ID at pretty much the correct frequency.

I find it particularly noteworthy that the average multipath signal, the center of the spreading, is on the actual rather than the Doppler shifted "main" signal. This makes sense since only the short path between the mobile antenna and any nearby reflector/scatterer is changing and this may as likely be in front of as behind the mobile. Most of the path length is fixed length.I suppose this fact means that it could be possible to reduce the energy associated with multi-path distortion by frequency domain filtering. Maybe this is commonly done on some systems somewhere, I don't know.

These are all cross-polarized signals so the sensitivity to multi-path is probably higher than if the mobile had a horizontally polarized antenna. Perhaps I'll try that some time for comparison. Listening to the tone by ear, the distortion is quite obvious and probably would be objectionable on SSB or for CW QSO.

Later I parked and spotted several other stations, though not to the database because I didn't have an Internet connection to the laptop. I also reversed the situation and transmitted from a parked location. For both of these stationary situations, all frequencies were correct and stable down to the last Hz.

Clearly decode of 70cm mobile WSPR is not going to work very well. The Doppler shift looks just like drift and beyond very slow speeds is going to cause standard decoding to fail. I did briefly try to decode JT9 and JT65A with no results but more investigation needs to be done here. It might be that one of these could work much better than WSPR.

Glenn n6gn