RSP1A vs. Airspy, 10m
30 Jan 2024
Another experiment, this time SDRplay RSP1A is compared to Airspy Mini. Just like in the previous test, both radios were connected to the same antenna through a splitter, ran for 24 hours on 10 m and decoded FT8.
First, the totals:
| SDR | Message Count |
|---|---|
| Airspy Mini | 126,415 |
| SDRplay | 126,286 |
The two counts are within 0.1% of each other, a difference smaller than the accuracy of the method.
Here is the plot that shows how the two radios performed in each decoding cycle. Most of the points lie around the red line, sometimes one radio decodes a few more messages than the other, and sometimes a few less. Just like on the chart for the RTL-SDR/Airspy comparison, there are outliers, but they appear at both sides of the red line. In other words, both receivers occasionally get overloaded, but not always at the same time, probably because their mirror channels are on different frequencies.
Two receivers together decoded 135,322 uniqe messages, 7% more than each of them alone.
Note that in yesterday’s RTL-SDR/Airspy test Airspy, in exactly the same setup, decoded 129,624 messages, which is 2.5% more than today. If I performed an A/B test (run Airspy for 24 hours, then SDRplay for the next 24 hours), I would come to a wrong conclusion that Airspy performs better than Sdrplay.
Added 5 Feb 2024:
A / B comparison by switching between two setups every few minutes is even less accurate than a day to day comparison. Here is a plot of the number of messages decoded with Airspy Mini in each cycle in the test described above. The counts vary from cycle to cycle by 20%-50%, making A/B comparison results meaningless.
Conclusions
- Airspy Mini and SDRplay RSP1A have the same FT8 decoding performance on 10m.
- Both radios sometimes get overloaded, a band filter might help.
- Running both radios in parallel increases the number of decoded messages.
- A / B tests may be misleading, only a side-by-side comparison produces reliable results.