Aliexpress images

I bought the black MSi.SDR (to complement the original blue dongle) out of curiosity after coming across a Reddit post mentioning it. There are a few versions but I went with the second which seems to be the most reliable one:

  • ‘simple’ - bare board and 5 antenna ports (manually switched - not very useful)
  • ‘full’ - metal case, single port, better filtering & TCXO (tested on 5MHz/10MHz WWVH signals)
  • RSP1C - see Github issue notes

More detailed info can be found in this Github README. There is also a neat comparison with the RSP1A. Important takeaway is it performs well on HF which is all I need.

HF spectrum comparison

Without filtering, the 50kW AM broadcast station near me overloads both receivers. However, the black SDR manages to filter out the interference above 12MHz. The following screenshots were taken using RSP spectrum analyser set to 2048kHz IF and total gain 55. The antenna was a 40m loop on ground. The cursor is on 9700kHz which is Radio New Zealand Pacific.

black vs blue no AM filter

With filtering the spectrum is much more usable. The noise floors are very similar at around -115dB, but the black SDR has peaks around 7dB higher.

with AM filter

Antenna selection

The first HF antenna I tested was an 84’ random wire antenna with 17’ counterpoise into a generic 9:1 balun. Note: I learned that if using this balun as an unun, you should cut the trace on the back of the PCB. This antenna, along with the EFHW (+49:1 balun) after it, was affected severely by suburban noise - things like solar inverters and Ethernet over power devices.

I then found the loop on ground (LoG) antenna (detailed analysis here). A 20m loop thrown on the yard miraculously outperformed the elevated horizontal antennas, in terms of lowering the noise floor. The screenshot below shows +20dB SNR for Radio NZ during the day.

LoG vs EFHW noise daytime

The noise ‘cancelling’ capability of the loop is clear if you unplug one side of it as shown below. A balun is also necessary for good results.

LoG balun comparison on 20m

The LoG is directional with highest gain perpendicular to the feed point. The downside of the LoG is the extremely low signal levels. For example, at an elevation of 15 degrees its gain is -13dBi at 14MHz. This necessitates a preamplifier particularly at lower frequencies e.g. 80m band. A larger loop helps but then at higher frequencies where length > 1λ, the lobes become very complex creating nulls in many directions.

It was not possible to quickly acquire a HF preamp so instead I set up a 40m loop around the backyard, approximately 2m in the air. This greatly increased the gain without increasing noise too much. To further reduce interference I positioned the antenna as far away from the house as possible. Due to a lack of coax it is fed using around 10m of twisted pair telephone jumper wire, which according to this calculator has a reasonable impedance of 40Ω. Additionally, at the balun connection I attached a clip-on ferrite.

Rough MS Paint diagram of setup

The setup performs well from 4MHz-20MHz. There seems to be a wide-band noise problem above 20MHz which makes 15-10m almost unusable. Topic for a future investigation…

24h 20m FT8 - 6000+ reports


<
Previous Post
OpenWebRX remote setup with RTL-SDR and MSi.SDR
>
Blog Archive
Archive of all previous blog posts