[hpsdr] HERMES: measurements of the MDS
Walter Schellenberg
w.r.schellenberg at bluewin.ch
Fri Aug 28 08:28:39 PDT 2015
Hi all,
I recently bought a Hermes EU board and measured the MDS. The significant
loss of sensitivity for Dither ON surprised me:
R = Random
D = Dither
1. Test with OpenHPSDR mRX PS v3.2.27 (5/14/2015)
Settings: 14.1MHz, CW, 500Hz, Att = 0dB,
FFT = Max, Audio Sample Rate 384'000
R&D OFF R ON D ON R&D ON
MDS = -136dBm -136dBm -119dBm -119dBm
2. Test with cuSDR64 BETA v0.3.2.14
Settings: 14.1MHz, CW, 500Hz, Att = 0dB, AGC Fast,
AGC-G 120dB, Sample Rate 384kHz
R&D OFF R ON D ON R&D ON
MDS = -134dBm -133dBm -117dBm -117dBm
3. Comparison with my old Mercury v3.3 and PowerSDR v KD5TFD 2 Feb 2011b
Settings: 14.1MHz, CW, 500Hz, Att = 0dB, AGC-T = 120,
Audio Sample Rate 192'000
R&D OFF R ON D ON R&D ON
MDS = -137dBm -136dBm -129dBm -129dBm
The slightly lower sensitivity of Hermes with R&D OFF is most probably due
to the digital step attenuator, which has an insertion loss of 1.3 to 1.9dB
(Mercury uses a relay switched attenuator).
But what happened to the sensitivity of Hermes with Dither ON? I think a
reduction of 16 to 17dB does call for an explanation.
What could be the reason?
- Except for the step attenuator the front ends of Hermes and
of Mercury are identical up to the FPGA.
- The datasheet of the LTC2208 ADC says “The dither DAC is
calibrated to result in less than 0.5dB elevation in the noise
floor of the ADC, as compared to the noise floor with dither off.”
- For my Mercury, I see an rise of the noise floor of about 3.5dB,
with Dither switched ON.
- For my Hermes with openHPSDR the rise is about 8dB, and with
cuSDR64 it is about 8 to 9dB.
Can anyone confirm my findings?
I like to switch both Dither and Random on, because only then I get a good
IMFDR (InterModulation Free Dynamic Range).
Your comments are highly appreciated,
Walter HB9AJG
More information about the Hpsdr
mailing list