After trying a number of different orientations of earth-electrode pairs in my garden, I have come to the conclusion that this antenna structure does not work very well here in the new QTH. I am a little surprised as the earth-electrode antenna worked well just 350m away at the old QTH. Whatever orientation I try I seem to be always in the -22 to -28dB S/N region with G6AVK 78km away.
Why is it not so good? Well, the most likely answer is that the effective distance between the far and near ground rod electrodes is smaller, even with a similar physical separation. This could be because of buried water pipes and utilities criss-crossing the garden, although I wasn't aware of any, or different soil/rock conditions, although the chalk layer is about as far down here as at the old QTH and the soil is a similar alkaline sort. The fact that the 8.97kHz earth-mode results were so poor last week rather hinted at a problem that was not present at the old QTH.
I haven't quite exhausted every layout possible in my strange shaped back garden, so there are more things to try with earth-electrodes, but it is looking like I shall have to erect a short Marconi or a fully in the air loop antenna to get better ERP on 472kHz. On my previous tests at the old QTH, a full in the air loop with 1mm diameter wire and an enclosed area of around 80m sq was around 8dB better than the earth-electrode antenna, but far more critical to tune and match because of the hi-Q. If my earth-electrode antenna is several dB down on the performance of the same structure at the old QTH then I could be expecting >10dB improvement or even more.
Half the fun is the testing and finding out what does and does not work. The coming week we have visitors staying, so not too much time to experiment. I shall have to stick to WSPRing on 20 and 10m.