Achieving and improving

Level & Attenuation

uncertainties in RF
calibration

Paul C. A. Roberts
Fluke Calibration

Co Author: Mark Ashcroft

©2011 Fluke Corporation.

 Abstract & Outline

Achieving and improving Level & Attenuation

uncertainties
i i in i RF calibration
lib i

How can a RF signal source produce low uncertainty

precision level and attenuation directly at its output
when traditionally signal generators, power meters,
p attenuators,, and complex
calibrated step p p
procedures Fluke 9640A RF Reference Source

are needed?

Presentation outline:
• Introduction – traditional methods
• Purpose-designed RF Reference source design architecture
• RF Reference Source calibration & traceabilityy
• Application example
– Attenuator calibration 2
Traditional precision level methods

• Why not use just a signal generator!

– Output can be un
un-leveled
leveled
– Output accuracy specs inadequate for precision applications
– High & unknown output VSWR
– Internal step attenuator either not fitted or accuracy inadequate
– Cable loss & mismatch

Output to UUT
Signal Generator

 Use p
power meter & splitter
p to improve
p level accuracy...
y

3
 Traditional precision level methods

 Use power meter & splitter to improve level accuracy

2-resistor splitter Levelled output to UUT

A UUT & Power sensor directly

B connected to Splitter ports
Signal Generator

Power Sensor
Remember to consider cable loss and
Power Meter
6dB splitter loss when considering
signal generator maximum output
level specs and required signal level
at UUT!

• Leveled output uncertainty contributions:

Splitter reversal often used to avoid
– Splitter port balance/tracking
port loss tracking contribution, but
– UUT / splitter port A mismatch adds extra mismatch contributions.
– Sensor / splitter port B mismatch
Second ‘standard’ power meter &
– Power sensor Cal Factor uncertainty
sensor often used in substitution with
– P
Power meter
t 1mW/0dBm
1 W/0dB reference
f uncertainty
t i t UUT. More complexity, potentially
– Sensor linearity, power meter instrumentation lower uncertainty, different mismatch
uncertainties, etc contributions, etc.
4
 RF Reference - System Partitioning

Output signals delivered direct to the load or Unit Under Test (UUT)
input minimize performance degradation due to cabling and
i t
interconnections.
ti Si
Signals
l generated
t d in
i the
th mainframe
i f are fed
f d to
t the
th
leveling head containing the level detector and attenuator circuits.

1mHz - 20kHz: Wavetable & DAC Fluke 9640A RF Reference Source

20kHz - 9MHz: Hi-Res DDS Multiple output

9MHz - 4GHz: PLL, Frequency Translation & Division amplifiers cover Filters provide very low
frequency range harmonic content

70dB of attenuation
10MHz in 10dB steps
OCXO

Frequency
Level O/P Level
Synthesis & Amplifiers Filters Attenuators
Control Atten Sensing
Division
Leveling Head

Floating
PSU
Mainframe (Base Unit) Additional attenuation for
lower level outputs,

Frequency Modulation Level Control circuits also With level sensing in

generated in FM DDS generate Amplitude Modulation mainframe when additional
attenuation in circuit.
5
 The Levelling Head

• Signals generated in the mainframe are fed to the

leveling head containing the level detector and
attenuator circuits.
• Output signals delivered direct to the load or Unit
Under Test (UUT) input
• Minimizes cable & interconnection attenuation, Fluke 9640A RF Reference Source

mismatch and repeatability effects

Level
Head O/P Attenuators
Sensing Signal & levelling
Leveling Head connections to
mainframe

Level control with splitter/sensor equivalent to “perfect” 50Ω resistor in series with zero impedance source

Specified level
Additional 60dB
delivered direct to
attenuation & similar
UUT from low
splitter/sensor in
VSWR output:
mainframe for lower
-130 to +24 dBm,
level outputs
1mHz to 4GHz,
1 1:1 to 500MHz
1.1:1 500MHz,
1.4:1 at 4GHz Levelling head with internal power splitter/diode sensor
and step attenuators (70dB)
6
RF Reference – architecture features
Space grade hermetically sealed relays
Leading RF switch repeatability, validated at every Calibration
Annual usage monitored, typically 100k operations.
0–70dB attenuator in the Head Rated 107 operations
p (preventative replacement
(p p at 5 million)).
Small signal only in N-Connector
Excellent interference and CM noise rejection
Output match defined close to load Power splitter and diode sensor in the Head
Output match defined close to load
Can typically neglect match errors out to 300MHz
Cable, flexing and connector to Head are sensed out

Purpose designed 50 & 75 ohm Heads

Output impedance conversion and losses included
in calibration and user interface.
Less chance damaging non-mating
non mating connectors

Heads stores ID & Calibration Data

Ensures Head has been calibrated with this base
Adjusts User Interface to 75/50 Ohm head fitted

0 & 60dB attenuator in Base

Minimum signal level in cable -56dBm
All Signals via a Single Connector
All metrology covered by one certificate Fl ti RF O
Floating Output
t t tto the
th Head
H d
- A comprehensive accredited Fluke certificate Breaks common mode current path if DUT is grounded
Reduced interconnects and connector wear Excellent CM noise rejection
7
Achieved performance summary

Key specifications summary for:

Accuracy, stability, resolution, dynamic range,

purity and low noise....

Indicated areas of particular

interest in level & attenuation
calibration applications...

Note: output match (VSWR)

critical to achieving level &
attenuation uncertainties

8
Detailed Level Accuracies

Highlighted areas also show recent performance improvements....

... relevant to AC rms voltmeter and hi-performance spectrum analyzer calibration
Detailed Level Accuracies - % power

10V p p-p
p
3.5V rms
250mW ±0.7% (W) ±1.2% (W)
II ±0.35% (V) ±0.6% (V)
II
II
II
II
200nV p-p
70nV rms
0.1fW rms

Highlighted areas also show recent performance improvements....

... relevant to power sensor reference level and linearity calibration 10
 Detailed Attenuation performance

Hi hli ht d areas also

Highlighted l show
h recentt performance
f iimprovements....
t
... relevant to power sensor linearity calibration and for fixed & switched attenuators
11
RF Source Level Calibration
• Measurement range
– 10Hz to 4GHz
– -130dBm to +24dBm (50Ω), +18dBm (75Ω)

• Method based on well established approach

– AC Measurement Standard measures low frequencies
q directly
y ((Hi-Z
I/P + precision 50/75Ω termination), sets LF reference point
 Traceable through Fluke primary standards lab to the national lab

– Power meter & sensors measure higher frequencies relative to LF

reference
 Traceable directly to national lab

• Spectrum analyzer extends measurement range to much

lower levels
– Power sensors not useable at low levels
– Spectrum analyser also extends range above maximum power
sensor level
– M
Measured d analyzer
l linearity
li it errors generally
ll much
h lless than
th
the estimated uncertainty of the test (< 0.02dB for a 70dB
range).
12
 Traceable through ratio (attenuation substitution) measurements
 Match conditions at the transfer points
Low Level Measurement

• Match conditions must be maintained

– Avoid mis-match error changing between the normalization and subsequent
measurements

• 9640A output attenuator & level ranging specifically designed to

accommodate this requirement.

• Points for analyzer normalisation and subsequent level measurement

chosen to ensure match conditions are maintained.

50Ω
20dB 10dB 20dB 20dB

Attenuation switching sequence presents one of these two pads at

the output when attenuation is required.
13
 RF Level Measurement Process
1): Up to 100kHz an AC Std
measures the RMS voltage
developed across a precision 4): High performance spectrum 2): The 100kHz point provides an absolute
50Ω termination,
termination and the power y
analyser linearity
y extends level range
g reference for flatness measurements with
is calculated. above maximum for power sensor. power meter & sensors up to 4GHz at
levels down to -47dBm.

+24dBm

Thermal power
+13dBm
sensor for levels
down to -7dBm.

-7dBm Diode power sensor

for levels from -7dBm
to -47dBm.
-47dBm

-85dBm

3): High performance spectrum

y
analyser linearity
y referenced to p
power
-130dBm
130dB
sensor & AC Std at -47dBm extends
range to lower levels, re-referencing
10Hz 1kHz 100kHz 125MHz 4GHz
at -85dBm.
14
Factory Calibration System

Product Test & Calibration Systems in

manufacturing cell.
Identical systems deployed at service
centres. T
Temp Control
C t l
Power Meter
Function Generator
Spectrum
p Analyser
y
Phase Noise Mixer
AC Measure Std
Power Sensors
VSWR Bridge
UUT
Frequency Counter
Freq Ref Buffers
Signal Switching
Phase Noise Ref

Rack enclosure provides temperature

controlled environment for calibration Fluke fully documents the whole calibration
standards and Unit Under Test (UUT) process and equipment in the 9640A Instruction
Manual (download from www.flukecal.com). 15
Calibration Certification
• All 9640A models provided with calibration ISO17025 and ANSI-Z540 compliant calibration certificate
– UKAS accredited cal available from Norwich UK factory & service center
– A2LA accredited cal available from Everett US service center

• Certificates contain extensive Level, Attenuation and VSWR results data

– Provide traceability import for customers using 9640A as their standard
– Provides actual VSWR data for mismatch uncertainty calculation rather than relying on datasheet specification
values
– Other parameters also included (Eg: harmonics, spurious, phase noise, modulation, etc)
 Attenuator Calibration Example
The precision, continuously variable attenuator within the 9640A can be used as a
reference in the calibration of other attenuators, or in the calibration of gain or loss of
other RF components.
• Example: Calibrating a 10dB coaxial attenuator (N-Type connectors) at 50MHz

Step 1:
With UUT attenuator in circuit,,
set 9640A output to +16 dBm,
Power Meter note power meter reading
9640A (approx +6dBm for 10dB UUT)

F/F UUT
Power Sensor
Adapter Attenuator

Step 2:
With UUT attenuator
removed, reduce 9640A
Power Meter output* to reproduce the same
9640A power meter reading.

Reduction in 9640A output

F/F gives actual attenuator value.
Power Sensor
Adapter
17
* Reduce 9640A output by UUT attenuator nominal value to avoid excessive/damaging power sensor input
User interface design objectives & impact
User interface design objectives:
Include application-oriented features…
Simplify
p yp procedures…
Reduce opportunity for mistakes…

Step & Offset/Error features allow direct

readout off UUT attenuator error from
f nominal

• Very quick & convenient for fixed and step attenuator calibration

18
 Achievable uncertainty example
Using an example DUT of 10dB at a match of 1.05 : 1, at 99% confidence we have:
(converting to and combining as power uncertainty and converting back to dB)

9640A Specification (+16dBm Reference) +/-0.02

+/-0 02 dB +/-0.46%
+/-0 46% power
Mismatch Uncertainty +/-0.011 dB +/-0.25% power
Comparison Uncertainty +/-0.013 dB +/-0.30% power

RSS Total (99%CL) +/-0.026 dB +/-0.61% power

RSS Total
T l (95%CL)
(9 %CL) +/-0.021
/ 0 021 dB +/-0.47%
/ 0 4 % power

19
 Optimising match
Adding a ‘masking pad’ (attenuator) can improve source match contribution
….Pad inserted at point of most benefit i.e. worst/most variable match

• Adding a 3dB masking pad to the previous example

– Pad inserted to ‘mask’ source output (attenuator switching) VSWR changes
 3dB pad gives 6dB improvement (reflections travel through pad twice)
– Caution!... large masking pads improve mismatch at expense of signal level at sensor

Step 1:
With UUT attenuator in
circuit, set 9640A output
to +16 dBm, note power
Power Meter meter reading (approx
9640A +6dBm for 10dB UUT))

F/F Masking UUT

Power Sensor
Adapter Pad Attenuator

Step 2:
St 2
With UUT attenuator
Pad ‘masks’ adapter removed, reduce 9640A
and source match output to reproduce the
same power meter
Power Meter
reading.
9640A
Reduction in 9640A
output gives actual
F/F Masking attenuator value.)
Power Sensor
Adapter Pad 20
 Uncertainty improvement
DUT of 10dB at a match of 1.05 : 1, with 3dB masking pad Without masking pad:
(also 1.05:1), at 99% CL:

9640A Specification (+16dBm Reference) +/ 0.02

+/-0 02 dB +/ 0.02
+/-0 02 dB
Mismatch Uncertainty +/-0.0075 dB +/-0.011 dB
Comparison Uncertainty +/-0.013 dB +/-0.013 dB

RSS Total (99%CL) +/-0.025 dB +/-0.026 dB

RSS Total
T t l (95%CL) +/-0.020
/ 0 020 dB +/-0.021
/ 0 021 dB

• In this case, improvement is quite small

– Mismatch contribution reduces from 0.011dB to 0.075dB
– T l uncertainty
Total i reduces
d ffrom 0
0.026dB
026dB to 0
0.025dB
025dB

• 9640A output VSWR is already low, so masking pad impact is minimal!

– However, the masking pad technique is extremely useful in many other cases

• In this application, a masking pad could improve poor sensor/detector match if placed at the
sensor/detector input.

21
Conclusions
• The design architecture and calibration of a purpose-designed RF
Reference has been explained
– Enabling precision level and attenuation without the need for additional
equipment
– distinguishing it from general purpose signal generators
• The application example described illustrates how the RF Reference
simplifies and streamlines the calibration of attenuators
– Achieved measurement are uncertainties equal to or better than those
obtained with traditional measurements techniques
• Optimizing matching conditions with masking pads can improve
attenuation measurement uncertainties
• Additional techniques and considerations relevant to attenuation
measurements are discussed in more detail in the written paper
– Coping with large value attenuators and detector choice
– Test signal
g spectral
p content
– Danger of using high input power
©2011 Fluke Corporation. 22
 References

• Rhode & Schwarz application note “RF Level Test System +20dBm to -130dBm” .
• P. C. A. Roberts “Measuring Output VSWR For An Active Leveled Source”,
Measurement Science Conference, 2008.
• P. C. A. Roberts “The Verification and Calibration of an RF and Microwave
Calibration Source”, NCSL Conference & Symposium, 2006.
• P. C. A. Roberts “The Challenges of Precision Analog Modulation Measurement”,
Measurement Science Conference, 2007.
• P. C. A. Roberts “Implementing and validating state-of-the-art phase noise
measurement.”, NCSL Conference & Symposium, 2010.
• Fluke 9640A/9640A-LPNX RF Reference Source Instruction Manual,, Chapter
p 5
“Calibration”, available from www.flukecal.com.
• Fluke Application Note “Calibrating attenuators using the 9640A RF Reference”,
document 4015975A A-EN-N, available from www.flukecal.com.

©2011 Fluke Corporation. 23