Sunday, August 19, 2018

Comparing Real-Life Noise Levels on Some SDRs

The coming DX season is only a week away here in the frozen north - if in any way a maximum summer temperature of 32.8 degrees Celsius is compatible with the term "frozen".

The 500-metre, 50-degrees Asia/Pacific beverage was set up, or rather "laid out" to avoid complications with remaining reindeer, yesterday. Noise levels seem to be fine, and today I decided to compare three of my 1.5 MHz+ sampling SDRs on a basically signal-free MW band.

I used both the Cloud-IQ and the RSP1A with SDR Console V3.0.2. I didn't want to go through the hassle of setting up Perseus for SDR-C only for this test (I use Jaguar at all times), so I decided to use its native Perseus software instead.

I used a reference signal on 468 kHz, and adjusted gain levels so they read the same signal level, in this case -68 dBm in AM mode, 6 kHz bandwidth. In the case of the RSP1A, this meant RF Gain 9, IF Gain -35, Visual Gain 15. The Perseus however measured -63 dBm. I used the same sample rate on the Cloud-IQ and the RSP1A (1536 kHz, around 1230 kHz alias-free), and the 2000 kHz (1600 kHz alias-free) sample rate of the Perseus. Center frequencies were 1100 kHz.

All SDRs were connected to the 500-metre beverage via a DX-Engineering preamp and a 1:8 Mini-Circuits splitter.

So, how did it go? First off, the Cloud-IQ:
RF Space Cloud-IQ Noise Level, 500-metre beverage antenna
Then, we take a look at the RSP1A:
SDRPlay RSP1A Noise Level, 500-metre beverage antenna
And finally, the Perseus:
Microtelecom Perseus Noise Level, 500-metre beverage antenna
Taking into account the different scale on the Perseus, there isn't much variation. The RSP1A is the most noisy of the three, but not by much.  A few spikes may indicate USB noise. I used ferrite clamps on the USB cable. Although the pics above indicate otherwise on 1224 kHz, the RSP1A's noise level was generally 1-3 dB higher than the Cloud-IQ.

Mind you, this is an empty band, so any sensitive radio with even the most fragile front end would do well. As I experienced last winter, the problem with the RSP1A is with high signal levels. You can always reduce RF Gain, but not when you're doing automated recordings. So, you either risk saturation, or not hearing the weakest signals.

I will have all three SDRs connected to the beverage and do some recordings in the evenings when there are signals in the band. Weak signal readability is what we're after.

And finally, we could take a look at how the KongSDR with a 70-metre longwire compares:
KiwiSDR Noise Level, 70-metre longwire antenna

And I saved the best for last. Too bad the Airspy HF+ has only 660 kHz bandwidth. A performance like this with full MW coverage would be stellar. Roughly 7 dB better than the Cloud-IQ! (Visual Gain reduced to show same signal level as the others on 468 kHz). But the trade-off compared to one SDR covering and recording the entire band is too huge.
Airspy HF+ Noise Level, 500-metre beverage antenna


Fumiaki Minematsu said...

Hello, I would like to know the antenna gain of 500m Bevaraage antenna at dBi unit. I guess the gain of preamp made by DX Engineering is +16dB. My guessing is correct?
Happy MW DXing!

jos_nqn said...

Another low cost SDR can you tesr are the afedri, if performs at full 2 Mha bandwith when use by ethernet cable

Unknown said...

You may want to review the You Tube videos on the effect RX bandwidth has on noise floor.