G3XBM (Roger): New ERP calculations on 472kHz with earth-electrode antenna in 8cm snow

With several repeatable ground wave reports on 472kHz WSPR using the earth-electrode antenna, it is time for a recalculation of my ERP based on a formula (for 500kHz) supplied some years ago by M0BMU.  The assumption (from M0BMU, that I am not at all sure is right) is that the noise floor is around 3uV/m in a 2500Hz (WSPR) bandwidth. So, based on the WSPR S/N ratio the field strength when using the earth-electrodes can be calculated.

• G3ZJO at 79km gets me at around -19dB giving my field strength around 355nV/m. 
• G3XDV at 61km he gets me around -12dB S/N giving my field strength around 810nV/m.  
• M0BMU at 69km he gets me around -30dB S/N giving a field strength around 100nV/m

ERP Calculations

• ERP using G3ZJO's results = (E*d)^2 / 49; with E = 355nV, d = 79km   ERP = 16uW
• ERP using G3XDV's results = (E*d)^2 / 49; with E = 810nV, d = 61km  ERP  = 49uW
• ERP using M0BMU's results =  (E*d)^2 / 49; with E =100nV, d = 69km  ERP = 1uW

Now, to me these figures seem far too low, so perhaps the noise level is much higher than 3uV/m. Certainly that is the case here where I have an S4- S8 noise floor due to SMPSUs and other noise sources. I think these numbers can be doubled based on ground wave losses but, even so, the figures appear to be extremely low.  My Marconi antenna is some 8dB (average) better than the earth-electrode antenna and I am sure the ERP is well above 1mW using this antenna.

Conclusions

I am totally puzzled: measuring antenna current on the Marconi I get ERPs that are an order of magnitude higher yet measured results with the earth-electrode antenna average just 8dB down on the Marconi. This suggests ERPs for the earth-electrode antenna in the low mW region, which I find much more believable.

Any thoughts pleased readers?Image may be NSFW.
Clik here to view.Image may be NSFW.
Clik here to view.