just a hint first: this is an interesting question, but it’s very
USRP-centric; there’s the usrp-users mailing list especially for
USRP-related things, which I find quite interesting, most of the time
The N210 is a modular USRP, meaning that the mixing is usually done on a
daughterboard, which will usually actually take your input signal and
bring it down to baseband. Therefore, the ADC always sees -50MHz to
+50MHz of signal (it’s a complex sampling ADC, hence the negative
frequencies), and all daughterboards for the N2x0 are limited to a 40MHz
baseband bandwidth, so there is no aliasing (^1).
The N2x0 architecture always uses the 100MHz sampling rate on RX, and
the user is free to choose any integer fraction of that rate (^2). Using
digital filtering and decimation, the 100MHz Nyquist bandwidth is
reduced to the desired sampling rate.
Thus, in usual operation, your daughterboard will be tuned, to generate
a 300MHz (complex) tone, with which it mixes your signal down to -160kHz
and 160kHz, which you will see with any sampling rate >320kHz. You are
free to specify which frequency the daughterboard tunes to, so you can
also tune to let’s say 295MHz and see your signal at 5MHz ± 160kHz,
although you’ll need to use a much higher sampling rate, for example
100MHz/18~=5.56MHz, or your signal will be filtered out.
I hope that answered your question
^1 there’s the exception of the basicRX daughterboard, which neither
mixes nor has the 40MHz low pass characteristics; if you use that, you
will have to make sure that your signal fits into the ADC’s bandwidth,
otherwise you will see images.
^2 as long as the resulting rate fits through gigabit ethernet, ie.
25MS/s for 16bit, 50MS/s for 8bit max.