[Lowfer] WSPR-15 tonite

[Garry Hess]

Pat,

The fact that you have a "strong" converter ahead of your R75 does not 
protect you from IM degradation. The reciprocal (coherent) system input 
3rd order intercept (IIM3) equals the reciprocal of the first stage 
IIM3, plus the gain of the first stage divided by the second stage IIM3, 
plus ... The additional stages only degrade the overall IIM3 and if any 
stage isn't much better than the first stage when its IM3 is reflected 
back to the input the degradation can be substantial.

I believe you use a Clifton Labs e-probe like I do and it has a very 
good spec: output IM3=40 dBm and input IM3=46.5 dBm (I see conflicting 
numbers for the "gain" but Jack's testing method uses -6.5 dB). The rule 
of thumb for mixers is IIM3 equals the LO power + 10 dB. That means only 
about 27 dBm in your case and that's not nearly as good as the e-probe. 
Following up with a good Norton amplifier may be 50 dBm at its output, 
but at the input it's only about 40 dBm and going back through the lossy 
mixer and e-probe again gives an IIP3 less than the e-probe. The real 
killer may be the R-75. I can't find a clear IM3 number for it, but 
based on Eric's citation of 67 dB 3rd order dynamic range it may only be 
on the order of 0 dBm at its input and that's way worse than 40 dBm when 
reflected back to the e-probe input.

Having said the above I look at your capture and see 3rd order products 
in the waterfall that arean't much different than what I see with an 
SDR-IQ as the receiver. I'm not sure what the SDR-IQ IM3 number is 
either but I think it's in the area of 20 dBm. Much better than the R-75 
but still lousy compared to the e-probe.

Your attachment shows WE2XGR/2 being decoded at about +2 dB SNR. The 
signal at 475.823 kHz is WG2XJM and while stronger, it doesn't decode. I 
think I have seen decode failures like that too on occasion, even with 
no really big signal either outside or inside the detection band like 
WE2XGR/6. But one must be careful to let the WSPR code "warm" up with a 
partial decode window before it really starts working. That can take 
nearly 30 minutes with WSPRX-15. I have also received an occasional 
false decode. More often when there are two strong in-band signals if 
either the upper or lower 3rd order product falls in band I get an 
unwanted decode for it.

To resolve your decode failure issue I'd suggest you save the .wav files 
and send them to Joe Taylor. I'm sure he'd be interested in exploring 
what's going wrong.
-- 
73, Garry, K3SIW, EN52ta, Elgin, IL