My mistakes with RTL SDR - Lessons learned

Buying an RTL-SDR dongle, and building a software-defined radio receiving system is easy and affordable; however, here are a few mistakes I did, and a few derived suggestions I learned since I started with an $8 generic dongle four years ago.
Subscribe on Twitter @rtlsdr4everyone or by clicking on blue button (top left) to get the latest updates.

Buying cheap unbranded (and ultimately expensive) equipment

The price is long forgotten when the quality remains, as the adage says, and that first dongle proved that truth when it failed after a few months. Why? Probably lack of static discharge protection, a common treat of Far East wonders at the time, so I ordered a Nooelec Mini. That dongle still functions flawlessly to this day, just like the Terratec I got after that, and mind you, those two have probably underwent more suffering than any dongle was ever designed to take.
Recent chinese generics doesn't seem to fail as often these days, but manufacturing quality leaves a lot to be desired. The price difference between a noname and branded dongle is 5-6 dollars now, and a premium bundle kit can be had for less than $30, a price I always advise to pay for longer warranty and peace of mind.

Play with software and especially with gain

Omitting this vital step was my biggest mistake after discovering software defined defined radio.
The game is called software defined radio for a reason: interpreting and processing the signal is done by specialized software.
Advanced SDRs, like the SDRplay RSP1 or RSP2, have so many options that it takes weeks to get a grip with software.
Even regular RTL-SDR dongles benefit from fiddling; noise reduction in SDRSharp makes a huge difference for shortwave signal reception.

Receiving is not transmitting

Most online and written resources focus on transmitting, and whilst some of the core principles are useful - transmission line improvements, HF propagation etc - receiving a signal is much easier.
I've read that a vertical quarter- and half-wave is "narrow-band". Yes, it is, for transmitting, but for reception, and an inch of error here or there is not the end of the world. A whip antenna, for say 150 MHz, will receive signals at 200 MHz and 500 Mhz - not as well as a whip antenna designed for 200 or 500 Mhz, but it will work.
As an example: the bottle antenna detailed on this blog was found by trial and error, seeing that it works wonderfully for Airband (as it's essentially a quarter-wave ground plane), I tried it for AIS, ADS-B and 70cm band. No bother.
Furthermore, a QFH, which is just 12 feet of wire wrapped in a cool shape around some wood, will also receive AIS, 70 cm and ADS-B signals. Not well, but it does.
Margin of error for receiving is greater than for transmitting.

It's a system, NOT just an antenna

Pressing the "Pay Now" button, then connecting your shiny new antenna will improve your reception, but better bang for the buck can be had by paying attention to the "transmission line" first: coax cable and noise reduction measures between antenna and receiver.
Remember this: until the signal is converted to digital, it can be influenced by external factors - after conversion, it's just ones and zeroes, but even then, USB cable quality matters.
Upgrading your coaxial cables will have an effect on reception and signal quality, and you don't necessarily have to spend a fortune: the difference between double-shielded and unshielded coax cable is an extra 20 percent.
Longer cable runs - over 30 feet / 10 meters - will have terrific losses unless low-loss coax is used. The effect is horrendous with higher frequency signals, for example: TV coax has 4 dB loss for a 10 metre length for ADSB - more than half of the signal is lost. I'm sad to see that many fellow ADS-B enthusiasts buy really expensive equipment and spend hours building a system, then connect $100 worth of antenna and preamp with low quality coax cable.

The importance of noise reduction

Signal is measured against electrical noise - similarly to measuring the height of a mountain from Mean Sea Level, called signal-to-noise ratio. Between two radio systems with the same antennas, the system with a lower noise floor / better noise reduction will be better.
Increasing the distance from the host laptop's electrical noise emission with an USB extension cable will have a significant effect, but it's a hit or miss.
As I call the second-largest city in Ireland home, with numerous antenna masts within a one-mile radius, reducing electrical noise was a necessity. Unfortunately, noise reduction info for receiving radio signals in a city was not available at one place, because most ham radio webpages (still) assume you live in the middle of nowhere and erecting a 120 foot / 40 metre mast is easy.
Starting small, I upgraded, tested and added noise reduction piece by piece, and was constantly amazed along the way by the incremental, yet when combined, significant improvements. The proliferation of smartphones, chargers, energy-saving light bulbs and Wifi in the last few years work against signal reception in urban areas, and whereas the latest gear will do wonders in a lab or at a remote location, the lack of noise reduction can render even the best receiver deaf in a city.

Raspberry Pi microcomputers

Wish I discovered Raspberries earlier. Not only an open platform to realize dormant electronics tinkering dreams, but ADS-B is so much easier, Power over Ethernet provides an easy way to create a remotely mounted station and access data over rtl_tcp. 1 MSPS is more than enough for browsing the bands.


I've spent a wonderful portion of my youth buried deep inside PC desktop cases overclocking processors, where less heat equals more performance. RTL-SDR dongles also generate heat, so why not try to apply the same principles? Cooling works for astrophotography as well, so maybe better reception is possible with RTL-SDR... I've tried various methods, from heatsinks to oil cooling, and seen improvements.
Basic heatsinking and fans added 3% more ADS-B reports in a recent test, which difference is inaudible, but I tried oil cooling in conjunction with Peltiers for sub-zero Celsius temperatures, which made an audible difference on weak signals.
Doubtful? Read this independent post for Outernet use and see for yourself.
Now I always cool non-review personal dongles.

A preamp will not solve issues

Unless transmission line, antenna, noise reduction and proper operating procedures are in place, buying a preamp will not transform your setup to a CIA listening station.
A preamp does not know what is a signal and what is noise: it will equally amplify both. When I opened the envelope with an LNA4ALL inside, I expected to hear everything. I was wrong, so wrong.
Granted, by having one, previously unheard voices might filter in - urging you to improve the system even more - but at this point I'm chasing my own tail, the message is: sort out everything else before ordering a preamp.

A preamp will help a lot

Dongles have a high noise figure, and low-noise amplifiers, as the name implies, have a much lower noise figure. Add the two together, and the overall noise figure will dramatically decrease.
Adding a preamp was the most dramatic improvement I could do to my system, but again, I've seen much better results after the rest of the system got sorted out.

Use a headphone

Whereas turning a dial with a huge headphone stuck over my head was the norm in the good ol' days, with the waterfall feature of SDRSharp, you actually see the signal. This is convenient and allows rapid signal acquisition, but for very weak signals, headphones and slow manual tuning is still king.

Receiver doesn't matter as much

Fretting over the receive performance difference between an $8 generic dongle and a $20 dongle and a $150 RSP2 is pointless; you get what you pay for.
More and more, I think that the affordable SDR revolution reached a point when the receiver matters less and less; the bottleneck is increasingly the antenna, man-made electrical noise and familiarity with software. In a city, it's hard to hear the difference between an $8 or $40 dollar device. The latest RSP2 incorporates broadcast FM and AM suppresssion, and if you don't have $150 to spare, an FM bandstop filter - the bane of radio reception with RTL-SDR dongles - is yours for less than $20.
Sit back, enjoy what you got. Stress is unnecessary.

Involve your significant other (and / or kids)

The body of my first weather satellite antenna - if we can call that - was a broomstick, with kebab skewers and chinese takeaway chopsticks taped together forming the arms. Coax cable used as receiving elements, 10 metres of unknown coax picked up at a car boot sale for 2 dollars.
Measurements were way off, and the whole contraption was held together with copious amounts of tape. I was deeply convinced it won't work, but the project gave me - and, more importantly, my partner as we "built" it together - a laugh.
It worked.
The resultant image was grainy and badly banded, yet she was involved in its genesis. So, when building successors, she not only put up with screws, wood and wires slowly invading her life, but actively participated.

Have fun

Radio enhances quality of life whatever you do: driving to work, finally catching a tune on shortwave, rendering a weather satellite image, or standing at the end of runway and hearing Tower clearance, expecting, then feeling the pressure waves shaking your internal organs from a departing Jumbo Jet.
Enjoy the ride. Technical details provide reassurance. Numbers and graphs are useful to justify a choice, and to fill your head with the warm fluffy feeling that you've got the best. The ultimate perceived difference between various receivers is down to you, be it an $8 RTL-SDR dongle, or a pro system costing thousand of dollars: it's up to you if you actually enjoy what you doing.

No comments:

Post a Comment