Choosing an SDR for HF through UHF: dynamic range, upconverters, and where the noise really comes from
- Replies 2
- Views 38
- Created
- Last Reply
A better way to browse. Learn more.
A full-screen app on your home screen with push notifications, badges and more.
This thread is meant as a working discussion of how to match an SDR to the signals you want, and how to avoid the usual gain-staging and noise mistakes. Spec sheets tend to emphasise frequency range and maximum bandwidth. In practice, the ADC resolution, front-end filtering and your antenna system decide how a receiver behaves on a crowded band.
Here is the rough landscape as I understand it. Please correct anything that is off.
RTL-SDR dongles (RTL2832U with an R820T2-class tuner) are cheap and good for VHF/UHF monitoring, ADS-B, rtl_433, POCSAG and similar work. They use an 8-bit ADC, are usually run around 2.4 MS/s to avoid dropped samples, and can overload easily near strong FM broadcast or paging transmitters. Some versions have a direct sampling mode for HF, and the dongle then behaves very differently: the ADC sees the whole spectrum with little filtering, so aliasing and broadcast overload become real concerns. A TCXO model and a ppm correction setting help with frequency accuracy.
HackRF One covers roughly 1 MHz to 6 GHz, has a wide sample rate and can transmit, but it is half-duplex and also has an 8-bit ADC. It is a flexible experimentation tool rather than a high-dynamic-range receiver. LimeSDR hardware offers 12-bit conversion, a very wide tuning range and full-duplex operation, which suits experimenters working with SoapySDR and GNU Radio. SDRplay RSP units and Airspy models target receive performance: higher-resolution ADCs, better filtering, and in some cases built-in notches or preselection. The Airspy HF+ line focuses on HF and VHF with strong large-signal handling at a narrow bandwidth. Check each model's datasheet for exact ranges and bandwidth before buying.
On the software side, SDR# and HDSDR are popular on Windows, and SDRuno is the Windows application for SDRplay hardware. GQRX is the common choice on Linux and macOS, and CubicSDR runs on several platforms via SoapySDR. Most of them can pipe audio to a virtual audio cable or a loopback device, which is how people feed WSJT-X (FT8/FT4), fldigi, DSD-type decoders or multimon-ng. Sample rate, AGC, filter bandwidth and audio level all need setting. Overdriving the decoder's audio input is a frequent cause of failed decodes.
I would like to discuss three things:
1. For HF on an RTL-SDR, is an upconverter better than direct sampling? If so, how do you handle the local oscillator offset in each program? Is it a frequency shift setting, an ExtIO option or something else?
2. How do you decide on gain? A rule I have seen is to lower the gain until the noise floor just starts to rise when the antenna is connected, then check for images or intermod products by changing the gain by a few dB. Is that sound, or are there better tests?
3. What antenna and feedline practices make the biggest difference to the noise floor? Examples would be a common-mode choke, an active loop away from the house, a discone for scanning, and a 137 MHz V-dipole.
Link to comment
https://www.hamradiobase.com/forums/topic/10821-choosing-an-sdr-for-hf-through-uhf-dynamic-range-upconverters-and-where-the-noise-really-comes-from/Share on other sites