I’ve been using the RTL-SDR dongle and a Nooeelc HamItUp+ Upconverter for the better part of a year for various projects and shortwave listening. I think they’re great devices, especially for the price point. Recently, I purchased an Airspy HF+ Discovery in an attempt to upgrade my listening experience, particularly on the HF / shortwave bands. The following is one of many tests that I’ll be conducting to compare the three devices.
- RTL-SDR v3
- Nooelec HamItUp+ UpConverter
- Airspy HF+ Discovery
- 9:1 balun
- Fencetenna connected to balun with wires as follows:
- -1x 100ft 16 AWG wire that is 2ft off the ground
- -1x 100ft 16 AWG 6ft off the ground with another 100ft 22 AWG magnet wire extension that is attached to the second story of my house at about 20ft
- SDR-Console V3
One caveat is that my tests are in no way scientific and it’s mostly based on my experience with signal, noise, and audio. However, I will provide supporting evidence to support my conclusions. I hope that you’re able to find this information useful.
In the first test below I’m testing 9690 kHz - Radio Exterior de España, coming from Noblejas, Spain. My QTH is Austin, Texas (5100 mi DX) and the test was conducted at about 1:40 UTC time on June 4th (June 3rd at 8:40 pm locally) while the 31m band was very active. I chose this station because SIO was 555. With this kind of performance I think that it will be obvious if one device can produce better audio than another and the difference in signal/noise.
The conditions that evening were favorable to the 31m band as you can see below.
The first test was a visual test on the waterfall to see what was visible on the band. Remember that the Airspy bandwidth is 768 kHz. On the RTL-SDR I’m using the bandwidth was set to 2.4 MHz. My friend Rose kindly reminded me that this is the sweet spot for the RTL-SDR so I think it makes sense to test it at this setting. I kept the floor and ceiling the same but I did move the waterfall intensity slider. More on that below.
In each test I’ll use the visuals to call out the S-meter readings for the signal and noise. I will be using this to understand the SNR. In addition, I’ll also be using the visuals to see relative selectivity. I have a fair amount of residential RFI in my neighborhood so this location will put the devices to the test.
RTL-SDR in Direct Sampling Mode
With the RTL-SDR in direct sampling mode we see a lot of noise mixed in with signals. The difficulty here is that it is hard to determine which are signals and what is noise. I found my self spending a lot of time selecting each one and trying to see if there was a DX station or if it was just noise. The S-meter shows S9 and the spectrum shows about the same. The noise is below S1.
In the audio, the noise is noticeable as a consistent hum but the signal is pretty solid. You can hear a little bit of doubling so this means I turned up the gain a little bit too much. To improve the audio, we could enable a noise reducer to get rid of some of the noise.
RTL-SDR with Nooelec HamItUp+ UpConverter
I paired the RTL-SDR with an Upconverter in an attempt to improve the signal to the dongle. The RTL-SDR works down to about 50 MHz by default so you can add an Upconverter to shift the frequency up to a sweet spot for the dongle. I left the visual gain as is to give an indication of the difference in gain between direct sampling and using an upconverter. What I observed here is that the noise seems to still be visible on the waterfall but would disappear once I turn down the visual gain. This improves the selectivity when visually looking at the signals on the waterfall. As you can see, the up converter increases the gain of the signal. The S-meter is at S9+18 which is a considerable improvement over direct sampling mode. The noise floor continues to be low at under S1.
In the audio, the noise seems to have been smoothed out a bit but is still noticeable. The signal seems solid and I can hear the voices in the interview well. I don’t hear any doubling present.
Overall I think that this is an improvement over direct sampling mode.
Airspy HF+ Discovery
What I noticed here is a pretty solid set of signals across the waterfall with very little noise. The S-meter shows S9+47 for 9690 kHz. One thing to note is the noise floor, which is considerably higher at about S3. However, given the signal strength and relative increase in gain compared to the RTL-SDR this is much better SNR.
In the audio, the signal is improved over direct sampling mode. The signal seems to be locking on the carrier a nicely and the fading is not as noticeable. The noise is controlled much better and there seems to be some more noise reduction in comparison with the other two devices.
I would say that this is a noticeable improvement in signal and audio vs the RTL-SDR in direct sampling mode but marginal improvement over the upconverter.
I definitely think that the Airspy HF+ Discovery has better performance, noise filtering, and overall sensitivity. It seems to be able to lock onto the carrier frequency and reduce fading much better than the RTL-SDR device or coupled with the Nooelec up-converter. The noise floor may be higher on the Airspy, however, I think that is indicative of better signal reception overall. The SNR is better and the noise much less noticeable in the audio. The signal came in clear and I wasn’t able to observe as much interference on the waterfall. The audio was clearly much better in my opinion. To add to this, the device is considerably small, especially compared to the up-converter
However, this isn’t a smoking gun. The RTL-SDR held its own in direct sampling mode and with the up-converter. Given the price, you get a lot of value out of the RTL-SDR and you are not missing a whole lot with strong signals. I think that you could probably reduce some of the noise and get better selectivity with improved grounding and more chokes. The upconverter does provide an improved listening experience but I still think the Airspy HF+ edges it out.
In upcoming tests I'll visit different bands and weaker signals to get a sense for selectivity of signal for a weak signal vs quality of audio with a strong signal like we see here.
In the meantime, drop me a line and let me know what you think. You can email me at alonsovargas at gmail dot com.