[hpsdr] NPR of Mercury (was: Boat Anchors..PreAmps) - erratum

[Adam Farson]

In my previous post on this topic,  I made a calculation error. I used the band-limited noise bandwidth as the denominator in the process-gain equation rather than the non-band-limited value (35 MHz at 160m, otherwise 30 MHz.) Here is the corrected version.

With band-limiting filters switched out:

Correction factors: CF1 = 10 log (fs/2Brf) = 10 log (80/2*35) = 0.6 dB
                   
                    CF2 = 10 log (80/2*30) = 1.25 dB

Calculated NPR = observed NPR - CFx

1940 kHz notch: NPR = 63 dB observed/62.4 dB calculated (CF1)

3886 kHz notch: NPR = 68 dB observed/66.75 dB calculated. (CF2)

5340 kHz notch: NPR = 66 dB observed/64.75 dB calculated. (CF2)

7600 kHz notch: NPR = 68 dB observed/66.75 dB calculated. (CF2)

All four calculated NPR values were ~ 8 to 12 dB worse than the theoretical 74.01 dB maximum shown in ADI MT-005, Fig. 2, p.3. The noise floor of an ideal 14-bit ADC is -(14 * 6) + 1.76 = -85.76 dBFS. The LTC2206-14 has a -77 dBFS noise floor; this will account for the difference between the theoretical and "real-world" NPR values to a reasonable approximation. 

The NPR readings in the 72 - 76 dB range which I saw with band-limiting filters inserted appear to be consistent with the process gain derived from the bandwidth reduction due to these filters. They are also in line with values I measured when testing superhet receivers.

I apologise for the error in my previous message.

Cheers for now, 73,
Adam VA7OJ/AB4OJ




 1339217943.0