I know what things to note when selecting an analog amateur radio, but SDR devices have completely different specs, e.g. sample rate and resolution.
What do I need to know about the device to make sure it is suitable for working Amateur Radio communications?
What you are really looking for in a SDR depends on your needs. Here are some factors to consider:
ADC/DAC resolution: this is measured in bits. It pretty much represents how finely the analogue to digital converter or digital to analogue converter can represent changes in the waveform. Usually higher is better (12-16 bits), but you can get away with lower with some DSP techniques.
Sampling rate: this is usually measured in MSPS (mega samples per second) or KSPS (kilo samples per second). It is a measurement of how much data the SDR can convert from analogue to digital or digital every second. Again,higher is usually better,as it will be more versatile,and versatility is one of the most important aspects of a SDR (in my opinion).
Tuning range: this is usually reported as a range of frequencies. Pretty much self explanatory,it is how much of the of spectrum the SDR can tune to.
Duplex: this is usually reported as one of the following: half-duplex, full-duplex, and 2x2. A half duplex SDR can only either transmit or receive at the same time. A full duplex SDR on the other hand, can do both at the same time,transmitting and receiving. A 2x2 (or higher) SDR will be able to transmit on two different frequencies AND receive on two different frequencies at the same time. A 3x3 can do it on three,and so on. There is no "best" option here per say, but it really depends on your needs. If you don't need to transmit and recieve at the same time, a half duplex is enough. However,if you do need to transmit and receive at the same time,such as a GSM base station, or something similar, you are going to need a full-duplex SDR. It really depends on your needs and your budget. A 2x2 is very versatile, but may be over budget and overkill in some situations.
Link type: usually reported as a name. The link type varies from SDR to SDR. The most common right now is Ethernet and USB. The advantage of Ethernet is that you can place the SDR in a farther away location, away from any emi generated by your computer. You are unable to do that with USB, as the spec does not allow you to stretch it for a long distance. However, I find USB easier to work with.
Price: usually reported as a multiple of a unit of currency. The price shouldn't be the main deciding factor when choosing a SDR. You should look at your current needs first, while also considering any future needs. Then, you should consider price. It is much better to spend a little more right now on a beefier SDR than to have to buy a new one down the road because it doesn't have the specs you need for a particular project.
SDR's typically use a sound card or sound card hardware to convert from quadrature at audio frequency to digital samples. This is due to the availability and low cost of audio-frequency components for computer soundcards. The remaining hardware is usually a direct-conversion receiver. Then the digital sample signal is operated on by DSP hardware, often the CPU of a general-purpose computer.
Analog to digital conversion of audio-frequency signals works by sampling the amplitude at a particular moment in time, over and over, and converting the amplitude to a binary number. This number is stored in a certain number of bits, such as 8 bits (values -128 to 127) or 16 bit (values -32768 to 32767). The highest frequency which can be received by a particular soundcard is determined by how frequently the amplitude is sampled. So, a 24 kHz sound card can only represent a frequency up to 12 kHz.
Since the dynamic range of the audio frequencies involved in amateur radio (voice mostly, or narrow bands of frequencies for digital modes) is small, the sample size (resolution) is is not particularly important. 16 bits is more than enough and 8 bits may even work. What is important is how fast the samples are taken, because this determines how much of the received signal can be converted to audio or digital signal. Even though frequencies above about 22khz are outside the range of human hearing, if they are digitized, they can be digitally converted to audible signals.
In this way, an SDR can process a wide swath of bandwidth, and the end user can select which small window of that bandwidth (say 2.5 khz) to listen to at a time. Thus, a 192 kHz sound card can convert 96 kHz either side of the received frequency to digital signals, which can be represented by a waterfall, and the user can select which portion of that digitized signal to convert into audio (or other modes).
