|
|
# Measurements
|
|
|
|
|
|
When designing, implementing, and verifying real-world communication systems,
|
|
|
there are numerous helpful measurements that can be taken. This page explains
|
|
|
some of these measurements and details how they can be taken using the Keysight
|
|
|
FieldFox RF Analyzer.
|
|
|
|
|
|
## Contents
|
|
|
|
|
|
1. [Keysight FieldFox](#keysight-fieldfox)
|
|
|
2. [Spectrum Analyzer](#spectrum-analyzer)
|
|
|
3. [Oscilloscope](#oscilloscope)
|
|
|
4. [Frequency Offset](#frequency-offset)
|
|
|
5. [Transmit Power](#transmit-power)
|
|
|
6. [Occupied Bandwidth](#occupied-bandwidth)
|
|
|
7. [DC Offset](#dc-offset)
|
|
|
8. [IQ Imbalance](#iq-imbalance)
|
|
|
|
|
|
## Keysight FieldFox
|
|
|
|
|
|
The Keysight FieldFox RF Analyzer is a convenient tool for making a number of
|
|
|
usesful measurements on digital communication systems. Its small size,
|
|
|
ruggedized form, and battery make it ideal for portable applications, although
|
|
|
the FieldFoxes will be limited to desktop usage for ELEN90089. The FieldFox can
|
|
|
be used in one of three modes:
|
|
|
|
|
|
- Cable and Antenna Analyzer (CAT) - typically used to to test an entire
|
|
|
transmission system, from transmitter to the antenna. Similar to VNA mode.
|
|
|
- Vector Network Analyzer (VNA) - used to characterize two-port RF devices by
|
|
|
measuring both amplitude and phase responses to an applied stimulus.
|
|
|
- Spectrum Analyzer (SA) - used to observe and measure the characteristics of
|
|
|
a signal generated by some device or occurring over-the-air.
|
|
|
|
|
|
In ELEN90089 we will use the FieldFox almost exclusively in spectrum analyzer
|
|
|
mode. The following are detailed references on the performance characteristics
|
|
|
and measurement setup for the FieldFox:
|
|
|
|
|
|
- [Keysight FieldFox N9914A Datasheet](datasheets/5990-9783.pdf)
|
|
|
- [Keysight FieldFox N9914A User's Guide (Abridged)](datasheets/N9927-90001.pdf)
|
|
|
|
|
|
> **Note:** It is important to note the accuracy of a given measurement made
|
|
|
> with the FieldFox will be limited by its relevant performance specifications.
|
|
|
> See the [datasheet](datasheets/5990-9783.pdf) for full performance
|
|
|
> specifications.
|
|
|
|
|
|
## Spectrum Analyzer
|
|
|
|
|
|
Using the FieldFox as a basic spectrum analyzer is straightforward. Either
|
|
|
attach an antenna or cable your transmitted signal into the *RF IN* port of
|
|
|
the analyzer. An SMA to N-type adaptor should be present on all FieldFoxes in
|
|
|
the lab. The FieldFox can be put into spectrum analyzer mode be selecting:
|
|
|
`Mode > SA`.
|
|
|
|
|
|
At a minimum, to observe the frequency spectrum of a signal at RF you should set
|
|
|
the following parameters:
|
|
|
|
|
|
- Center Frequency: `Freq/Dist > Center`
|
|
|
- Bandwith: `Freq/Dist > Freq Span`
|
|
|
|
|
|
The FieldFox will likely set the amplitude (Y-axis) scale automatically, but you
|
|
|
may wish to adjust this to better investigate the signal of interest:
|
|
|
|
|
|
- To use autoscale: `Scale/Amptd > More > Autoscale`
|
|
|
- To set the units per division (gridlines): `Scale/Amptd > Scale`
|
|
|
- To set the reference level (top of graph): `Scale/Amptd > Ref Level`
|
|
|
|
|
|
Markers can be used to investigate specific points on the output spectrum. For
|
|
|
example, to find the frequency and power of the highest peak displayed you can
|
|
|
choose: `Marker > Peak`.
|
|
|
|
|
|
Refer to the [user's guide](datasheets/N9927-90001.pdf) for full details on
|
|
|
configuring the FieldFox.
|
|
|
|
|
|
## Oscilloscope
|
|
|
|
|
|
The FieldFox can be used as an oscilloscope by setting the frequency span to
|
|
|
zero: `Freq > Freq Span`. You can then:
|
|
|
|
|
|
- Set the center frequency as desired: `Freq/Dist > Center`
|
|
|
- Set the sweep time to change the length of the time plot:
|
|
|
`Sweep 3 > Sweep Time`
|
|
|
- Set the resolution bandwidth (ResBW) to change the captured bandwidth:
|
|
|
`BW 2 > ResBW > Man`
|
|
|
|
|
|
## Frequency Offset
|
|
|
|
|
|
The frequency offset of a device (e.g., the bladeRF) can be measured by
|
|
|
configuring the device to generate a tone at a known frequency and then using
|
|
|
the marker function to observe the location on the FieldFox display. Frequency
|
|
|
offset is then the difference between these two values.
|
|
|
|
|
|
> **Note:** The accuracy of frequency measurements are limited by the accuracy
|
|
|
> of the LO used in the FieldFox, which is +/- 0.7 ppm + aging. See the
|
|
|
> [datasheet](datasheets/5990-9783.pdf) for full performance specifications.
|
|
|
|
|
|
## Transmit Power
|
|
|
|
|
|
Any external attenuator used (e.g., the 30 dB SMA attenuator we use for hardware
|
|
|
loopback) should be accounted for in power measurements. This can be done
|
|
|
automatically in the value displayed by the FieldFox by setting
|
|
|
`Scale/Ampltd > More > External Gain`.
|
|
|
|
|
|
The power level of a single frequency tone is best measured using a marker:
|
|
|
`Marker > Peak`.
|
|
|
|
|
|
For signals with non-zero bandwidth, e.g., a modulated signal, the channel
|
|
|
power (CHP) within a frequency range should be measured using:
|
|
|
`Measure 1 > Channel Measurements > Channel Power`.
|
|
|
|
|
|
- To set the integrated bandwidth used in the CHP measurement:
|
|
|
`Meas Setup 4 > Integrating BW`
|
|
|
|
|
|
> **Note:** Power measurements with the FieldFox generally have an accuracy of
|
|
|
> +/- 1 dBm. See the [datasheet](datasheets/5990-9783.pdf) for full performance
|
|
|
> specifications.
|
|
|
|
|
|
## Occupied Bandwidth
|
|
|
|
|
|
The occupied bandwidth measurement looks over a frequency span and determines
|
|
|
the start and stop frequencies between which is contained a specific percentage
|
|
|
of the observed power, e.g., 95% of the observed power. These frequencies are
|
|
|
marked with vertical bars and both occupied bandwidth and occupied power
|
|
|
measurements are displayed. To measure occupied bandwidth:
|
|
|
|
|
|
- `Measure 1 > Channel Measurements > Occupied Bandwidth`
|
|
|
|
|
|
## DC Offset
|
|
|
|
|
|
The RF front-ends of the SDRs we employ in ELEN90089 all employ direct
|
|
|
conversion architectures (zero IF or homodyne) for both their transmit and
|
|
|
receive chains. Direct conversion greatly reduces the number of components in
|
|
|
the RF front-end and simplifies the design, but results in a number of specific
|
|
|
impairments that need to be addressed. DC offset, partially as a result of LO
|
|
|
leakage, is one such impairment.
|
|
|
|
|
|
DC offset of a transmitter can be observed by configuring the device to
|
|
|
generate a signal with no energy component at dc and then observing the output
|
|
|
on a spectrum analyzer. As an example, we can generate a complex exponential
|
|
|
rotating at a frequency f_0 at baseband, which after upconversion by our RF
|
|
|
front-end should be present f_0 Hz above our carrier frequency (f_c + f_0).
|
|
|
The ratio between the power at our desired frequency (f_c + f_0) and that at
|
|
|
the carrier (f_c) is a quantification of the severity of the DC offset
|
|
|
present. This is typically expressed in units of dBs below carrier (dBc).
|
|
|
|
|
|
```
|
|
|
P_dc (dBc) = P_{f_c + f_0} (dBm) - P_{f_c} (dBm)
|
|
|
```
|
|
|
|
|
|
Larger values of `P_dc` are obviously better. DC offset of a receiver can be
|
|
|
measured in a similar way by feeding a passband signal with no energy at the
|
|
|
carrier frequency (f_c) into the RF front-end downconverter and observing the
|
|
|
resulting baseband spectrum. Care must be taken that the DC offset of the device
|
|
|
generating the waveform is negligible so as not to distort the measurement.
|
|
|
|
|
|
> **Important Note:** DC offset is both gain and frequency dependent.
|
|
|
|
|
|
## IQ Imbalance
|
|
|
|
|
|
IQ imbalance is the other main impairment of direct conversion receivers that we
|
|
|
discussed. It is the result of a mismatch in the gains and phases of the
|
|
|
in-phase and quadrature LOs generated for use in up- or downconversion. It is
|
|
|
present in all architectures that use quadrature modulation or demodulation,
|
|
|
but is more pronounced in direct conversion architectures.
|
|
|
|
|
|
IQ imbalance of a transmitter can also be measured by generating a tone at a
|
|
|
known frequency offset (f_0) from the carrier (f_c) and observing the output
|
|
|
spectrum. When present, in addition to the desired tone at f_c + f_0 (Hz), IQ
|
|
|
imbalance results in a complementary tone at the same offset below the carrier,
|
|
|
f_c - f_0 (Hz). The ratio between the power of these two tones is a measure of
|
|
|
the severity of the IQ imbalance.
|
|
|
|
|
|
```
|
|
|
P_{iq} (dBc) = P_{f_c + f_0} (dBm) - P_{f_c - f_0} (dBm)
|
|
|
```
|
|
|
|
|
|
Receiver IQ imbalance can also be measured using the a single-tone test but,
|
|
|
again, care should be taken to ensure the measurement is not capturing any IQ
|
|
|
imbalance in the device generating the waveform.
|
|
|
|
|
|
> **Important Note:** IQ imbalance is both gain and frequency dependent. |