Review: v.3

An RTL-SDR dongle is a repurposed TV tuner, able to receive radio signals from about 24 MHz to 1.8 GHz. The latest generation is from, expands tunable frequency range into low MHz territory.
Previously unseen features in a sub-$30 receiver: software-selectable bias-T and HF via direct sampling using standard antenna port; and more, for a detailed breakdown and official point of view, please visit the manufacturer's page (link).
Comparison and references to its predecessor called v.2 or "previous Silver dongle" in this review, and to the Nooelec SMArt is inevitable; please read the SMArt review if unfamiliar with Nooelec's top dongle.
Reader note: New v.3 dongle has been extensively tested for over a month, on its own and against competitors, in two countries at multiple locations, with several antennas, so this is a long post with lots of images and videos, might take a while to load on a bad internet connection. Read the Manifesto for my approach, Youtube links provided separately, click or tap images for full-screen glory, underlined text are links, open up in a new page.
No cliffhanger: v.3 is the best RTL-SDR dongle available now.


Receiver only for $20, kit with two antennas $25 (plus shipping if applicable). Buy only from the manufacturer store, online marketplaces sell previous versions for $300. Yepp, read that right, as always, buy directly from maker.

What you get

Kit comes with small and large antenna, receiver with end caps, and magnetic mount with SMA connector.
Complete plug n' hear package, great as a present, anyone, especially radio nerds will be impressed.
Magnetic mount with short (39"/ 1m) and thin cable, two telescopic antennas, larger extends from 8" / 20 cm to 6 ft / 1.5 m and usable on HF, smaller collapses to 2 3/4" / 6 cm from 8" / 20 cm. Read: supplied antennas work for strong shortwave signals up to GHz range.

SMA connector standard, mount and antennas will be compatible with other SMA-equipped radio gear, such as SDRPlay and handheld transceivers and receivers.

What's inside

Same case and screws as previous v.2, printed circuit board (PCB) slides out easier, thermal pad different color, less sticky, SMA connector same length but with more thread:

PCB is full of components and a bit larger than v.2, everything surface mounted and properly labeled for soldering aficionados:

Clearly, there's barely any space on the PCB for any more improvements, and that list is long.

Cleaner waterfall

Less distracting "is that a signal?" streaks on screen.
Older v.2 on top, new v.3 on bottom, 25.4 dB gain, 135 MHz:

Not only in the middle of upconverter range, lowest visible noise of any dongle ever tested.
2 meter amateur band, same gain:

Noise floor graphs comparing latest v.3 (top) and previous v.2 (bottom) below, click or tap on image for full-screen view, or download with right-click and "Save As" for later:

Bias-T via software

Means that receiver is sending power down the coax cable to power external accessories, commonly a low-noise amplifier (LNA).
The best and recommended LNA out there is the LNA4ALL, which can be ordered with bias-T factory enabled for around $30 (manufacturer info and link).

Software selectable means that you won't have to do any physical work; two clicks and have power to a bias-T enabled device - not only to an LNA4ALL, a Spyverter upconverter can be also powered (not tested).
Instructions for Windows and Linux / Raspberry Pis provided in the user guide, bias-T worked straight away with LNA4ALL.

Don't forget to turn it off thought if changing antennas, because there's no feedback from the dongle that Bias-T is on.
As a safety feature, removing the dongle from the USB port  / disconnecting from power resets bias-T to OFF.

Direct sampling - access frequencies below 30 MHz

Main reason for buying new v.3.
Means that signal directly enters the brain of the dongle; concept is not new, possible with even the cheapest generic chinese dongles (walkthrough guide for other dongles link). Twist with new dongle is that standard antenna connector can be used: no need to open up the case and plug in wires.
Process is simple: stop front-end software (SDR# used, same principle in others), click on Cogwheel, then click on Direct sampling - Q branch. Roll down numbers on top with mouse wheel to a known frequency - commercial FM to shortwave with a few clicks.

Youtube link:

To get better performance, add an upconverter: v.3 dongle with Ham-It-Up v.3, direct sampling is less number by 125 MHz:

Note that 20 feet / 6.5 meter wire was used as antenna; direct sampling barely works, if at all, on a discone or on a physically smaller antenna. Furthermore, HF performance via direct sampling heavily depends on location, much more so than "regular" HF via upconverter. Far away from civilisation,v.3 was fantastic with a Balun One Nine and a monstrous T2FD, and as evident in the video, still usable in a city centre with 20 feet of wire.
It's easy: mount comes with a miniature silver protector, keep it, because it helps to hold a wire in place. Wrapped a length of wire around antenna mount, fixed it in place by tightening protector from top.
An upconverter is better? Yes, for $30-$50. In either case, better reception is possible with a a long piece of wire, and further signal enhancement is doable by building a 9:1 Unun, or order a balun One Nine (review hereorder here) if you're not the DIY type.

Holes for advanced users

For projects which were way beyond amateur's soldering skills before new v.3, such as passive radar and linking receivers together for a common clock.

Four mounting holes in the corner enable stacking dongles. Just like breast enlargement, it's good to look at features knowing they're there to play with - check out what passive radar is on this excellent webpage.

Drift, TCXO, heat

Drift means that tuned frequency changes as dongle heats up, TCXO eliminates this problem: enter a number, hear signal on that number. No guessing.
TCXO works, less than 1 ppm as promised by manufacturer.
New dongle gets really hot, significantly more than previous generation, so raised my concerns with the manufacturer, who replied the following:

"The heat produced is normal, and the V3 is expected to get a little hotter as there is an extra amp in there for the HF circuit, drawing about 20-30mA extra. However, the V3 has much better heat dissipation because we redesigned the PCB to have much larger copper sink areas (in internal PCB layers, which other dongles don't have), and to allow the heat to be extracted via the thermal pad much better. The result may be a hotter case, but cooler insides."

Ran two v.3s to check, one bare and one with heatsinks in identical ModMyPi ADS-B back-end setup, and difference was 1.6 percent over three days, well within statistical error.
Cooling is not necessary per se, stock dongle performed without hiccup left under the sun in temperatures exceeding 100 Fahrenheit / 40 Celsius degrees for 7 hours, and its brother worked flawlessly for a week on max gain ADS-B duty.

Antenna base and cable

Wide antenna base with strong magnet, same as supplied with a Nooelec SMArt, addition of golden-looking copper-based texture (feels and shines like an anti-snail tape), supposed to increase signal strength.
Couldn't hear or see a difference versus a non-copper / standard black antenna base (both on the same metal plane), but it looks cool and it's super-sticky, won't peel off:

Magnet is same strength as in SMArt's magnetic mount, meaning it will pick up a baby Iron Man, small household objects, and I could leave the large telescopic fully extended on the balcony railing without worrying that it will fall off.

Teething problems and manufacturing quality

Had no problems at all. sent two new v.3 receivers, indistinguishable by touch, waterfall image, gain setting, noise profile or performance. These were one of the first shipped units, and discovered online that minuscule issues plague first-batch dongles, like some functions not working as they should, for a full list, manufacturer posted problems at the bottom of the user guide. All of these issues have been rectified in units that ship now. I commend the maker for publishing faults, openly and publicly, even when they won't affect 99.9 percent of buyers.

Improved ESD protection and toughness

Don't try the following at home.
ESD means electrostatic discharge, electricity in the air flying around finds your antenna, nests there, then kills your receiver. Had this issue with generic chinese RTL-SDRs, not a problem with recent dongles, but went ahead and supposedly increased dongle's ESD protection.
Does it work? Forgot antenna on car connected to dongle, powered up, torrential downpour with spectacular thurderstorm, nothing amiss. Left in the car for hours, interior got so hot that couldn't touch steering wheel, Silver dongle still happily provided ADS-B data.
Case is the same as previous generation's, tough as nails.
I'm confident new v.3 will survive anything you throw at it, deliberately or by mistake.

Front-end software

Worked with SDRConsole, SDRuno 1.04, and SDR# Will probably work with others, basics (RTL2832U with R820T2) are the same as other dongles.
If you experience installation issues, the process is the same as with every RTL-SDR dongle: install proper driver with Zadig first.


Read the list what can be received with RTL-SDRs, new dongle does most of them, out of the box, without the need to spend extra on accessories. Received signals from 900 kHz to 1090 MHz, and if you're new to the game, that range encompasses the majority of human radio communication which can be received without a satellite dish or complicated radio nerd antennas, from a device the size of a lighter.
For $25, anyone can listen to HF straight away with supplied accessories, a feat no other software defined receiver can pull off. Surely, five-letter people will complain that audio quality is terrible on HF with direct sampling, and they'll be right, but for a beginner who never experienced a shortwave station coming in from thousands of miles away, the sense of accomplishment will be priceless.
Even if you take bias-T and HF access out from the equation, less visible and audible noise makes a huge difference during daily use. Software Defined Receivers are tuned in software, not with buttons and dials, cleaner waterfall = easier tuning.
For better reception, place mount outdoors on something made of metal, like a metal chocolate tin, then adjust telescopic to frequency of interest.
On my workhorse discone, swapping between previous v.2 and latest v.3 Silver dongle, or switching to the SMArt wasn't a life-changing experience, all three dongles (v.2, v.3, SMArt) had plus-minus 1 dB SNR ratio displayed, reacted to gain changes similarly, and sounded much alike.
I'm not gonna dwelve into the vs SMArt or vs my grandma's wooden radio or vs <insert receiver here> debate (at least not here and now), because no other radio device in this universe offers this much for this little money. No, not even $8 chinese R820T2 generics.

Best buy

Twenty-five dollars. For two antennas, cable with magnetic mount, and a receiver usable from kHz to Ghz range.
More features than any other dongle available at the moment.
Peace of mind that v.3 Silver dongle will work, below 30 MHz, with bias-T, with less visible noise.
Consequently, v.3 is the dongle I recommend to anyone.

1 comment:

  1. Great review, thank you!

    At 40dB and 50dB the v3 has less noise, but at gains of 0-35dB the RTL-SDR v2 seems to have a lower noise floor than v3, is that right?

    What was the gain setting for the Nooelec vs. RTL-SDR v3 waterfall plots of the 2m band?