On Sun, May 13, 2018 at 4:34 PM, Marcus D. Leech via USRP-users
> On 05/13/2018 04:25 PM, Zhongyuan Zhao wrote:
> Thank you for the response. 3.6dB NF is not impressive however I was
> considering a wideband LNA as an once-for-all solution for the entire
> frequency range 50-6000MHz of N310.
> Let me re-phrase the question in two different ways:
> 1) What is the minimal discernible signal (MDS) at the input of the N310
> receiver if no external LNA is used?
> I don't think there are measured specs out there yet, but my experience in
> the past with most of these radios is that the narrowband MDS is
> usually better than -120dBm at max gain over most of the operating
Part of the question is that MDS in these radios includes the digital
demodulation portion. In narrow band analog radios with an analog IF
stage, you could use
a voltmeter to measure the analog voltage present to determine an MDS.
With a Terminated input, you would measure the analog noise floor,
then add a known signal and measure the input signal needed to add 3
dB to the analog noise floor, which gave you a close measurement to
the radio noise floor, as well as a way to determine the MDS.
Digital radios make is more difficult to do this because of the
digital processing involved. MDS is heavily dependent on the type of
modulation as well as the performance of the demodulator. This is
something you're going to have to figure out. For example, spread
spectrum signals (GPS, for example) can be demodulated with signals
that are 15 to 20 dB below the noise floor, so something you couldn't
see on a spectrum analyzer.
Marcus's comment of a narrowband (25 kHz Channel type of narrowband)
matches well with the work I've done with analog radios with 2 to 3 dB
noise floor, so the quoted 5.8 dB noise floor for the N310 sounds
On a side note, the combination of wideband LNA's in front of wideband
radios is generally something to be careful of, as strong signals out
of band can cause problems. Bandpass filtering in front of LNA's can
be a very good idea to think about.
> 2) If only thermal noise is fed into the N310 receiver, (-103.9dBm for 10MHz
> bandwidth) what is the output power or power spectrum density (dBm/Hz) at
> the output of receiver? Suppose the calibration between dBFS and dBm is done
> correctly. In theory, a receiver with Noise Figure of 5.8dB adds 5.8dB to
> input noise, therefore output power of the receiver would be -98.1dBm.
> However I am not quite sure about if this is applicable to near thermal
> noise floor. A spectrum analyzer has a NF of 3dB but the noise floor
> displayed might be 20dB above the thermal noise floor.
> There is usually some non-linearity measuring very close to the noise floor.
> If you have an N310 and a decent lab (I assume that the UNL electronics lab
> has things like calibrated noise sources, etc), then you can answer these
> questions yourself.
> I am especially interested in the band of 500-1000MHz, but answers for other
> bands are helpful as well.
> Many thanks!
> Zhongyuan Zhao
> PhD Candidate,
> Department of Computer Science & Engineering,
> University of Nebraska-Lincoln
> Office Hour: WF 9:30-10:00am, Avery Hall 12,
> Suite 117, Schorr Center,
> Lincoln, Nebraska 68588-0115
> On Sun, May 13, 2018 at 2:36 PM, Marcus D. Leech via USRP-users
> <firstname.lastname@example.org> wrote:
>> On 05/13/2018 03:24 PM, David Bengtson via USRP-users wrote:
>>> The Receiver Noise figure should include sampling noise and ADC ENOB.
>>> Is that a free space path loss? You don't have any design margin for
>>> fades/blockages etc, so an External LNA might be a good idea. 3.6 dB
>>> NF isn't particularly low for a LNA, but it will help.
>> Technically, the equivalent noise figure of the ADC is "swamped" by the
>> gain stages ahead of the ADC.
>> System noise chain analysis usually (not always) gets increasingly boring
>> the further away from the antenna you get, because noise figures
>> of "deep" stages are diluted considerably by the aggregate gain of the
>> stages ahead of them...
>>> On Sun, May 13, 2018 at 3:49 AM, Zhongyuan Zhao via USRP-users
>>> <email@example.com> wrote:
>>>> I am trying to calculate the link budget for USRP N310.
>>>> Suppose I want a SNR of 20dB,
>>>> the signal bandwidth is 10MHz,
>>>> the thermal noise floor is -103.9dm.
>>>> Receiver noise figure is 5.8dB (N310 datasheet 5.8dB @1.8GHz)
>>>> Rx Antenna Gain 6dBi
>>>> Path loss: 120dB
>>>> Tx Antenna Gain 6dBi
>>>> RF cable Loss (RX+TX): 1dB
>>>> Then my TX power before TX antenna should be
>>>> -103.9dBm + 20dB + 5.8dB - 6dBi + 120dB -6dBi + 1dB = 30.9dBm
>>>> The receive signal strength at RX port of USRP = -78.1dBm
>>>> The noise strength at RX port of USRP = -103.9dBm
>>>> The SNR before receiver is 25.8dB, after receiver is 20dB.
>>>> However, due to the sampling noise and effective bits of ADC, I doubt
>>>> the noise floor after the receiver would be much higher than -103.9dBm +
>>>> 5.8dB = -98.1dBm, e.g. -90dBm?, in that case my receive SNR would be
>>>> than 20dB.
>>>> Does anyone know the output noise floor (or sensitivity) of USRP N310
>>>> receiver (e.g. RX2 port) in 500-1000MHz band? e.g. in the unit of
>>>> That would help me to decide if an external LNA is necessary. If so does
>>>> some typical wideband LNA with a gain of 20dB and noise figure of 3.6dB
>>>> would be enough.
>>>> Thank you!
>>>> Zhongyuan Zhao
>>>> PhD Candidate,
>>>> Department of Computer Science & Engineering,
>>>> University of Nebraska-Lincoln
>>>> USRP-users mailing list
>>> USRP-users mailing list
>> USRP-users mailing list
> USRP-users mailing list
USRP-users mailing list