— John Pumford-Green 02/06/23 08:18

On recent 60m QSOs in the early morning with G4VLC / Hereford we've noted a substantial asymmetry in the observed S/N readings on signals in each direction.

Peter's signal arriving here in Shetland has been “good” with S/N of +10dB or more while my signal arriving in Herefordshire has been “poor” with S/N of -10dB or lower.

At other times of day the S/N reports have been more similar, with differences mainly due to different TX powers, but not such a gross discrepancy between the North → South and South → North paths as seen during these morning asymmetrical QSOs.

One thing I thought of that could explain the asymmetry is excessive feeder loss in my feeder between radio and antenna. This would decrease my ERP by the amount of the excess feeder loss, and the S/N would be degraded at the far end by the same amount. In the reverse direction a lossy feeder on receive might go un-noticed. The thinking for this is that the receiver's internal noise floor is likely to be substantially lower than the “atmospheric” noise collected be the antenna, along with the wanted signal. The S/N of the wanted signal is therefore set at the antenna and as long as the feeder loss doesn't reduce the “antenna noise” enough to approach the noise floor of the receiver, then the demodulated S/N will not change substantially despite the feeder loss.

To make sure that I wasn't causing the asymmetry myself, with a lossy feeder, I measured the power loss across the feeder from shack to Antenna feed point in 2 ways.

1. Inject a carrier at 10W at the shack and measure the power at the far end of the feeder
2. Make a swept return-loss measurement on the feeder, with an open circuit at the far end. The measured Return Loss will be 2x the feeder loss.

1. 10w input → 8.5W at the far end : 0.7dB feeder loss
2. Swept Return Loss → 1.6dB : 0.8dB feeder loss

I'm confident that the feeder isn't causing my TX signal to be unknowingly attenuated, and degrading my outward signal.

I have some patch-leads and coaxial switches in the shack between the transceiver and the feeder input, so I also did a swept through-loss measurement of these and found the loss to be around 0.4dB. That makes a total of 1.1dB. For 20W TX power I'd expect 15.5W to appear at the antenna feed point. I assume most other people have similar feeder losses (possibly worse?)

I decided to try and capture the effect on a North - South UK path between Shetland and G4ZFQ on the Isle of Wight, 1061km due south from here, so I set up a WSPR session overnight hoping to capture a asymmetry in the reported S/N ratios on the North → South versus South → North paths.

I ran WSPR overnight on the UK 60m WSPR slot, at 1W TX power and checked the resulting spots the next morning. Below is a graph of the reported S/N ratios for each path, separately.

It's clear that overnight the S/N was largely stable, with difference of around 10dB (favouring the South → North Path, and probably due to my lower noise floor?). At around 02:30z the S/N for the two paths started to diverge, as the S/N received here (South → North) improved while the S/N received at G4ZFQ (North → South) declined, and continued to drop through the morning.

As a first test I think it shows that there is something going on that affects North → South paths and South → North paths differently in the period following local sunrise.

What causes it?