Where dynamic range actually gets lost in an SDR chain: gain, filters, upconverters and antennas
- Replies 2
- Views 34
- Created
- Last Reply
Top Posters In This Topic
-
Admin_Terry 1 post
-
Admin_Lee 1 post
A better way to browse. Learn more.
A full-screen app on your home screen with push notifications, badges and more.
A lot of SDR troubleshooting threads boil down to the same symptom: the waterfall is full of signals that shouldn't be there. Ghost carriers, FM broadcast images in the airband, a noise floor that jumps when you touch the gain slider, or an HF band that looks like a comb. The usual advice is to turn the gain down, but that skips over why it works. So here is a framework for discussion: where in the chain do you run out of dynamic range, and what can you actually do about it?
Start with the ADC. An RTL-SDR dongle uses an 8-bit converter, which gives a theoretical ceiling of roughly 48 dB of dynamic range. That is before the tuner's own limits and before real-world ADC imperfections. Receivers like the SDRplay RSP series, Airspy R2/Mini, LimeSDR and the Airspy HF+ line use higher-resolution converters, and that gives you more room before a strong signal pushes the weak ones into the quantization noise. Bit depth is not the whole story, though. Linearity of the front end (the LNA, mixer and any AGC stages) usually determines whether strong out-of-band signals create intermodulation products inside your passband. A wideband receiver sees everything the antenna delivers, not only what is displayed in the waterfall.
This matters for the sample rate setting. Running 2.4 MS/s or 10 MS/s lets you see a wide slice of spectrum, but the ADC and front end are handling the total power of everything in that slice and often well beyond it. If a few strong FM broadcast stations or a pager transmitter are nearby, they can dominate the converter even when you are listening to a narrow signal. That is why a broadcast-FM notch filter, or a bandpass filter for the band you care about, often improves things more than a more expensive dongle would.
For setting gain, a practical method is to watch the noise floor, not the signal. Connect the antenna and raise the gain until the noise floor rises noticeably above what you see with the antenna disconnected or terminated. That tells you the receiver is no longer the limiting factor. Past that point extra gain mostly eats headroom. If a strong signal makes the whole waterfall breathe or creates symmetric copies of itself at other frequencies, you are overloading something and should reduce gain or add filtering ahead of the receiver. Also check whether the AGC in your software or driver is fighting you. Some drivers and tools let you set tuner gain stages individually, which helps when one stage is the problem.
HF is a separate case. Most RTL-SDR dongles don't cover HF in their normal tuner range. Options are direct sampling mode, available on some models and fairly limited in performance, or an upconverter that shifts HF up into the tuner's range. With an upconverter you have to account for the local oscillator offset in your software. SDR# can handle it with a frequency shift setting, and some SDR hardware and software combinations have built-in support. Watch for LO leakage and spurs, and keep the converter away from noisy USB hubs and switching power supplies. Hardware designed for HF, such as the Airspy HF+ series or SDRplay units with good HF filtering, avoids this extra stage but costs more.
On HF the antenna and local noise usually dominate before receiver sensitivity does. A noisy LED supply, a USB 3 cable or a computer monitor can raise the noise floor across several bands. Ferrite chokes on the USB and coax, a feedline that is physically separate from the computer, and a proper common-mode choke at the antenna feedpoint can change what you hear more than any receiver upgrade. For VHF and UHF, an LNA close to the antenna, preceded by an appropriate filter, usually beats an LNA at the receiver end of a long coax. Note that an LNA with no filter will often make overload worse.
Then there are the software tools and what each is good at. SDR# is popular on Windows and has a large plugin ecosystem. SDRuno is specific to SDRplay hardware and has sophisticated filtering and multi-receiver features. HDSDR is a Windows choice with an ExtIO driver interface, often used for HF and with CAT control. GQRX is the common option on Linux and macOS and works with SoapySDR and osmosdr-supported hardware. CubicSDR is cross-platform and also works through SoapySDR. For decoding, you usually pipe or route audio to another program: WSJT-X or fldigi for digital modes, multimon-ng for pagers, rtl_433 for ISM-band sensors, dump1090-type tools for ADS-B. A virtual audio cable is the most common connection method on Windows, with equivalent loopback devices on other systems.
Finally, panadapters and bandscopes. A panadapter takes a sample of a transceiver's IF or a separate receive path and shows the surrounding band on a waterfall, with CAT control keeping the SDR synchronized to the radio. The key rule is to protect the SDR from your own transmitter. Do not connect an SDR directly to an antenna that your transceiver transmits on. Use the radio's dedicated IF output, a proper RX-only antenna, or a switching and attenuation arrangement designed to survive TX power.
Questions for the thread: what is your approach to choosing between filtering and a better receiver? Where do you place the LNA? And for those running a panadapter, which method do you use for TX protection?
Link to comment
https://www.hamradiobase.com/forums/topic/11363-where-dynamic-range-actually-gets-lost-in-an-sdr-chain-gain-filters-upconverters-and-antennas/Share on other sites