The RTL-SDR blog recently announced release of their RTL-SDR v4 Dongle. This version has significant improvements over just about any other versions of the RTL-SDR. It is certainly better than the widely available dongles in the stock configuration. The blog's RTL-SDR v3 dongle was the prior leader in performance and extra features; this v4 design is better in multiple ways.
I tend to favor direct sampling for reception below 30 MHz, mostly because of a belief that it is better to use fewer oscillators and mixers in a radio. They add complexity and are pathways for things harmful to weak signal reception. However, the v4 dongle has a nifty upconverter design.
The upconverter uses the RTL-SDR's 28 MHz clock oscillator for mixer injection, which is a stable and accurate temperature compensates crystal oscillator (TCXO). It is a clean signal, with very little drift. For the actual mixing, a passive doubly balanced mixer circuit is used, which is a no-nonsense and effective design. To accurately tune the low bands, one simply needs to enter a proper frequency offset into the SDR interface. I hope every effen' WebSDR site with RTL dongles gets these, as reception should be much better than on dongles using direct sampling. Not because direct sampling is inherently bad, but most operators don't seem to install enough RF filtering or a decent external LNA.
No knowledgeable person would ever call an unenhanced RTL-SDR a sensitive receiver. With a good LNA (truly low noise, not just named "LNA" by a marketing team) and front end filtering, the dongle can perform on par with well designed commercial equipment. Most users will need to use higher internal gain settings in order to receive weaker signals, which raises the dongle's internal background noise and pushes all signals closer to a level where clipping and intermodulation happen. That causes spurious signals to be generated within the SDR, interfering and blocking real signals.
I live in a somewhat urban environment, with a few very strong FM broadcasters appearing on VHF, and one very strong AM mediumwave station stirring up things on the low bands. Being an RTL-SDR user with an interest in longwave and VHF airband sctivity, I must constantly deal with annoyances caused by strong out of band signals. I mostly solve overloading problems by using directional VHF antennas and a sharply tuned loop for the lowbands.
The RTL-SDR v4 dongle has RF filtering so much better than my stock dongle, that it is actually a sensitive receiver out of the box, and can be operated at higher gains without overloading. The bottom line is that v4 dongles have a far better dynamic range, therefore are better if you use an LNA or crank uo the internal device gain. Know that you should absolutely use a good LNA and directional antenna if you are serious about receiving weak signals.
The new V4 dongle, with its upconverter, can pick up signals more than 15 dB weaker anywhere below about 25 MHz. Above 30 MHz, and except for the FM broadcast band, the receivers are about equally sensitive. Or, you could say, they have similar noise figures over small VHF or UHF bands.
The RTL-SDR v4 dongle really shines in its ability to operate well in an environment with strong FM broadcasts and DAB signals. This is really a big deal. If you look at the RTL-SDR blog announcement, there is an important chart way down, not far from the bottom of the page: the Two Tone Sensitivity Drop Test. It shows that the new dongles score at least 30 dB better on frequencies over 300 MHz. On ADS-B and Inmarsat frequencies, the advantage is more like 37 to 40 dB better than a v3 dongle, which is already better than an off the shelf RTL-SDR.
With its much improved resistance to overload, the v4 dongles should work quite well for 137 MHz weather satellite downlinks, or even weather satellite downlinks on L band (on a suitable antenna).
The RTL-SDR Blog v4 Dongle is built up around the R828D tuner chip. Production of them has stopped, but there are batches available for use in the v4 units, but the supply will eventually run out. At that time, hopefully, there will be a different device which may be used to produce similar (or better) SDR dongles. The R828D is especially good because it allows the use of three different front end RF filters. Subsequent dongle designs, such as the more expensive R860, don't have multiple switchable inputs. WTF, might there be a future in which future tuner chips have lower noise figures and higher overload resistance? Performance is expensive, so that may not happen until competitive forces demand it...
The new RTL-SDR Blog v4 dongle has significant performance advantages over other RTL-SDR designs. Yes, you could build a system of similar or better performance with a mashup of external components. Buy why bother? The v4 dongle has most of the qualities users want, all packed nicely into one small case. Have fun with this, until you need a 14 or 16 bit SDR, or other advanced features or performance. The v4 units will make excellent airband receivers for voice, ACARS, or VDL Mode 2. It performs even better for Inmarsat downlinks or ADS-B.
There is an RTL-SDR v4 tutorial and user guide which is helpful in getting the dongle set up and running properly. See also our RTL-SDR info page for more about using these nifty SDR devices.