SDR gain, dynamic range and noise: where do dongles, upconverters and software settings actually matter?
- Replies 2
- Views 69
- Created
- Last Reply
Top Posters In This Topic
-
Mod_Mike 1 post
-
Mod_Chrissy 1 post
A lot of SDR troubleshooting comes down to the same few questions. Is the receiver deaf, overloaded, or just listening to local noise? This thread is for comparing approaches across hardware (RTL-SDR, Airspy, SDRplay, HackRF, LimeSDR) and software (SDR#, SDRuno, GQRX, HDSDR, CubicSDR).
Here is a starting framework, and corrections are welcome.
Sensitivity is rarely the limit. On HF, and often on VHF in built-up areas, external noise from the antenna usually exceeds receiver noise. If connecting the antenna raises the noise floor noticeably, extra preamp gain mostly amplifies noise and strong signals together. Overload is the more common problem. Typical signs are ghost signals that move when you change gain, FM broadcast images showing up across the band, and a spectrum that gets worse as you raise the gain slider.
ADC bits are only part of the picture. Theoretical dynamic range improves by roughly 6 dB per bit, but real devices fall short of that. An 8-bit RTL-SDR has much less headroom than a 12-bit Airspy or SDRplay, and tuner linearity, AGC behavior and front-end filtering all matter too. Hardware with selectable filters (the SDRplay notches, for example) can help a lot where there is strong broadcast energy. For cheap dongles, an external FM-broadcast notch or a bandpass filter for your band of interest is often a bigger improvement than a new software package.
HF options. You can use direct sampling (supported on some RTL-SDR variants, or on HF-focused receivers), a dedicated HF-capable SDR, or an upconverter that shifts HF up into the dongle's tuning range. An upconverter adds a local oscillator, so check for spurs and phase noise, and keep its supply and cabling clean. Direct sampling on an 8-bit device is simple but limited. Aliasing and the lack of preselection are the usual complaints.
Antennas. Match the antenna to the job. A discone is a flexible wideband choice for scanning VHF/UHF. Resonant antennas such as a 1090 MHz collinear for ADS-B or a V-dipole or QFH for 137 MHz weather satellites usually beat a broadband antenna on a single task. On HF, popular choices are an end-fed wire with a 9:1 unun and a small active loop, and local noise often decides which works better. Common-mode chokes on the feedline and USB cable, a short USB lead, and keeping the dongle away from the PC all reduce self-generated interference.
Software settings. FFT size, window function and averaging change what you see on the waterfall. Frequency correction (ppm) matters most on cheap dongles without a TCXO, and it matters more at higher frequencies. For digital decoding, route audio through a virtual audio cable into tools such as WSJT-X, fldigi or DSD+. Use multimon-ng, dump1090 or rtl_433 for protocols with their own decoders. For FT8, system clock accuracy matters as much as frequency accuracy.
Panadapters and bandscopes. Feeding an SDR from a transceiver's IF output gives a wide view without a second antenna. If you instead split or tap the antenna line, make sure the SDR is protected during transmit, because many SDR front ends cannot survive transmitter power. CAT control through Omni-Rig or hamlib keeps the SDR display in sync with the radio.
What order do you usually follow when a setup misbehaves: lower gain, add filtering, change the antenna, or check the software chain?
Link to comment
https://www.hamradiobase.com/forums/topic/10963-sdr-gain-dynamic-range-and-noise-where-do-dongles-upconverters-and-software-settings-actually-matter/Share on other sites