You are on page 1of 20

Watt-hour Meter

An electric meter or energy meter is a device that measures the amount of electrical energy
supplied to or produced by a residence, business or machine.
The most common type is more properly known as a kilowatt hour meter or a joule meter.
When used in electricity retailing, the utilities record the values measured by these meters to
generate an invoice for the electricity. They may also record other variables including the time
when the electricity was used.
BASIC SINGLE-PHASE METER. A single-phase watt-hour meter is essentially an induction
motor whose speed is directly proportional to the voltage applied and the amount of current
flowing through it. The phase displacement of the current, as well as the magnitude of the
current, is automatically taken into account by the meter. In other words, the power factor
influences the speed, and the moving element (disk) rotates with a speed proportional to true
power. The register is simply a means of registering revolutions, and by proper gearing is
arranged to read directly in kilowatt-hours. (Note: In some cases, the meter reading must be
multiplied by a factor called the "register constant" or "meter multiplier" to obtain total kilowatthours. See "Register constant (K),

W kN
Where:

W Pt

W reading of the watt-hour meter (watt-hour)


P power drawn (watt)
t time of usage (hour)
N number of revolutions made by the disk
k proportionality constant (meter/register constant)

Sample Problem 1:
The MERALCO test of a 10 A wattmeter having a constant of 0.4, the disk makes 40
revolutions in 53.6 seconds. The average volts and amperes during this period of test are 16
volts and 9.4 A. What is the percent accuracy of the meter at this load?
a. 97.45%
b. 98.58%
Wmeas

c. 98.07%
d. 96.44%

Solution:
kN 0.4 40 16 W hr
%accuracy

Wactual Elt

Wactual

1 hr
116 9.4 53.6
3600 sec
16.23 W hr

Wmeas
Wactual

16
16.23
%accuracy 98.58% answer

Sample Problem 2:
A given ampere-hour meter is under test by connecting it across a 230-V DC source.
For a duration of 180 minutes, a constant current of 20 A flows. This meter registers 570.5

and 585.8 kW-hr before and after the test respectively. Calculate the percentage accuracy of
the meter.
a. 92.31%
b. 90.02%

c. 91.44%
d. 89.20%
Solution:

Wactual Pt Elt

230 20 180 min 601hrmin

Wactual 13.8kW hr
Wread 585.8 570.5

13.8
100%
15.3
%Accuracy 90.02% answer
%Accuracy

Wread 15.3kW hr
Sample Problem 3:
A 15 A, 120-V watt-hour meter has a disk constant of 2. When tested on a unity power
factor load, 24 disk revolutions are counted in a period of 2 minutes. How many disk
revolutions would be counted per minute if the power were to change 0.5 lagging. Assume
the same line current and voltage in both conditions.
a. 6
b. 10

c. 8
d. 12
Solution:
Note: Voltage and Current are constant as given.

kN
; P Elpf
T
Elpft kN

E1l1pf1t1 kN1 1

E2l2pf2 t 2 kN2 2

Divide eq. 1 by eq. 2:


N1 pf1t1

N2 pf2 t 2
N2

N1pf2 t 2 24 0.5 1

pf1t1
1 2

N2 6 rev answer

WATTMETER READING USING A PT AND A CT


Pactual Wreading Multiplier
Multiplier CT ratio PT ratio
Where:

Pactual actual power consumption


Wreading wattmeter reading (watt)
CT ratio current ratio of the current transformer
PT ratio voltage ratio of the potential transformer

Sample Problem 1:

A 115-V, 10 A, three-phase watt-hour meter, having a basic meter constant of k = 2/3 is


connected to a three phase, three-wire circuit through a 100:5 ampere CT ratio and 2300:115
volt PT. A time check shows that the meter disk is making 15 revolutions in 50 seconds. What
is actual kW of the load?
a. 720 kW
b. 288 kW

524 kW
d. 380 kW
Solution:
P

kN

2 3 15
50

720 W

3600

Pactual Wreading Multiplier


720 400

Multiplier CT ratio PT ratio

Pactual 288 kW answer

100 2300

115
5

400

Sample Problem 2:
A power plant customer draws power at 220 volts from transformers on a pole. Current
transformers with ratio of 200/5 are used to meter the electrical usage. What is the multiplier
of the kW-hr and demand meters?
a. 40
b. 200

c. 100
d. 80
Solution:
Note: Since the metering did not use any potential transformer,
the PT ratio ratio in the above formula is 1.
Multiplier CTratio Pt ratio
200
1
5
Multiplier 40 answer
Multiplier

Sample Problem 3:
A current transformer with a turns ratio of 100:5 and a potential transformer with a
turns ratio of 10:1 are connected to the current and voltage coils of a single-phase wattmeter
measuring power delivered to a load. If the wattmeter reading is 240 W, what is the actual
power measured delivered to the load?
a. 50,000 Watts
b. 42,000 Watts
Solution:
Multiplier CT ratio PT ratio
100

10
5
200

c. 48,000 Watts
d. 45,000 Watts
Actual Power Wreading Multiplier
240 200
Actual Power 48, 000 Watts answer

Signal generator
An electronic test instrument that delivers an accurately calibrated signal at frequencies from
the audio to the microwave ranges. It is valuable in the development and testing of electronic
hardware. The signal generator provides a signal that can be adjusted according to
frequency, output voltage, impedance, waveform, and modulation.
A signal generator, also known variously as a test signal generator, function generator, tone
generator, arbitrary waveform generator, digital pattern generator or frequency generator is
an electronic device that generates repeating or non-repeating electronic signals (in either the
analog or digital domains). They are generally used in designing, testing, troubleshooting,
and repairing electronic or electro-acoustic devices; though they often have artistic uses as
well.
There are many different types of signal generators, with different purposes and applications
(and at varying levels of expense); in general, no device is suitable for all possible
applications.
Traditionally, signal generators have been embedded hardware units, but since the age of
multimedia-PCs, flexible, programmable software tone generators have also been available.
General purpose signal generators
Function generators
A function generator is a device which produces simple repetitive waveforms. Such devices
contain an electronic oscillator, a circuit that is capable of creating a repetitive waveform.
(Modern devices may use digital signal processing to synthesize waveforms, followed by a
digital to analog converter, or DAC, to produce an analog output). The most common
waveform is a sine wave, but sawtooth, step (pulse), square, and triangular waveform
oscillators are commonly available as are arbitrary waveform generators (AWGs). If the
oscillator operates above the audio frequency range (>20 kHz), the generator will often
include some sort of modulation function such as amplitude modulation (AM), frequency
modulation (FM), or phase modulation (PM) as well as a second oscillator that provides an
audio frequency modulation waveform.
function generators are typically used in simple electronics repair and design; where they are
used to stimulate a circuit under test. A device such as an oscilloscope is then used to
measure the circuit's output. Function generators vary in the number of outputs they feature,
frequency range, frequency accuracy and stability, and several other parameters.
Arbitrary waveform generators
Arbitrary waveform generators, or AWGs, are sophisticated signal generators which allow the
user to generate arbitrary waveforms, within published limits of frequency range, accuracy,
and output level. Unlike function generators, which are limited to a simple set of waveforms;
an AWG allows the user to specify a source waveform in a variety of different ways. AWGs
are generally more expensive than function generators, and are often more highly limited in
available bandwidth; as a result, they are generally limited to higher-end design and test
applications.

Special purpose signal generators


In addition to the above general-purpose devices, there are several classes of signal
generators designed for specific applications.
Tone generators and audio generators
A tone generator is a type of signal generator optimized for use in audio and acoustics
applications. Tone generators typically include sine waves over the audio frequency range
(20 Hz20 kHz). Sophisticated tone generators will also include sweep generators (a function
which varies the output frequency over a range, in order to make frequency-domain
measurements), multitone generators (which output several tones simultaneously, and are
used to check for intermodulation distortion and other non-linear effects), and tone bursts
(used to measure response to transients). Tone generators are typically used in conjunction
with sound level meters, when measuring the acoustics of a room or a sound reproduction
system, and/or with oscilloscopes or specialized audio analyzers.
Many tone generators operate in the digital domain, producing output in various digital audio
formats such as AES-3, or SPDIF. Such generators may include special signals to stimulate
various digital effects and problems, such as clipping, jitter, bit errors; they also often provide
ways to manipulate the metadata associated with digital audio formats.
The term synthesizer is used for a device that generates audio signals for music, or that uses
slightly more intricate methods.
Video signal generators
A video signal generator is a device which outputs predetermined video and/or television
waveforms, and other signals used to stimulate faults in, or aid in parametric measurements
of, television and video systems. There are several different types of video signal generators
in widespread use. Regardless of the specific type, the output of a video generator will
generally contain synchronization signals appropriate for television, including horizontal and
vertical sync pulses (in analog) or sync words (in digital). Generators of composite video
signals (such as NTSC and PAL) will also include a colorburst signal as part of the output.
Video signal generators are available for a wide variety of applications, and for a wide variety
of digital formats; many of these also include audio generation capability (as the audio track
is an important part of any video or television program or motion picture).
Application Notes

EZ Terminal Software for Vista, XP and Win2000


Signal Forge EZ Terminal for Windows
Serial Communication Software for the Signal Forge Signal Generators
Signal Forge EZ Terminal program for Windows. EZ Terminal is and easy to use serial
communication software utility for the SF1000. EZ Terminal provides support for
macros, terminal window screen capture, downloading firmware (Wave Manager
software for the SF1000), and downloading arbitrary waveform descriptor files. EZ
Terminal is similar to the well known Windows HyperTerminal program.

3-in-1 Signal Generator

The Signal Forge 800/1000 Digitally Synthesized Signal Generator combines the
features of both a signal source, signal generator and function generator in a single,
low-cost device. In addition, the SF800/1000 incorporates features which make it ideal
for testing differential systems, such as high speed serial busses, analog and digital
circuits, and RF and telecommunication equipment.

Compensating for External Signal Loss


In any RF or analog test setup, there is likely to be signal loss due to cables,
attenuators, filters or switches between your source, the Signal Forge 1000 or 800,
and the device under test (DUT). The accuracy of the signal level that arrives at the
DUT is affected by the sum of these components. This application note describes how
to measure the signal loss and how to compensate for for it using the embedded Wave
Manager software.

Controlling The SF1000 With An External Program


In some cases it is necessary to control the Signal Forge 1000 (SF1000)
programmatically using an external application or controller. This document describes
how to create and use an external control file to operate the SF1000.

Controlling The SF1000 Externally - Sample File


This sample file may be used as a guide in developing a program to control the
SF1000 externally. See the "Controlling the SF1000 With An External Program" form
more information.

Converting LVPECL to LVDS and CML


The Signal Forge Digitally Synthesized Signal Generator provides a differential clock
output that conforms to the LVPECL standard. While LVPEC is a widely used standard,
there are other differential signaling standards in use today including LVDS and CML.
This application note addresses how to interface the LVPECL differential output of the
SF800/1000 to conform to the LVDS and CML standards.

External Clock Requirements for the "E" Model


The SF800E/1000E are designed to operate only when and an external 10 MHz digital
clock source is attached. (For applications where an external clock source is not
available, the SF800/1000 should be used.)
This document defines the requirements for the external 10 MHz reference clock
source.

Generating Spectrally Clean Output


The SF800 and SF1000 Signal Generators provides an AC coupled output that can
source a sinusoidal output in the range of 100 KHz to 1 GHz. This output provides a
signal with harmonics at least 20 dB down (from 300 MHz to 1 GHz) at programmable
output power levels. If a cleaner sinewave signal is needed, for example to drive the

LO, or a receiver mixer, an external filter may be attached as described in this


application note.

I & Q Output
While the Signal Forge 800/1000 does not produce I & Q outputs directly, an external
splitter may be used to derive the in-phase and quadrature components from the ACcoupled output. This application note describes how to design an external I & Q
splitter.

Increasing or Decreasing the Signal Generator's Output Power


The Signal Forge 800/1000 provides an AC coupled frequency source output with a
range of 11 dBm to +7 dBm. By adding an external amplifier or an attenuator to
increase or decrease power respectively, the amplitude range may be extended.

Power Conversion Table


This application note provides a conversion between dBm, mW and mV for the entire
power range of the Signal Forge 1000.

Signal Generators as Alternative to BER Testers


The traditional method of design margin testing in serial data communication systems
is to employ a specially designed bit error rate (BER) tester. While BER testers have
proven effective, they are typically expensive and they do not always exercise the
system under test with the same level of noise or using the same data patterns that
will be seen by the system in the field. This article discusses a test methodology, using
signal generators, that may be applied to a wide range of data communication systems
and devices which transmit data over a serial bus. This test methodology may be part
of the design verification process or it may be used to qualify substitute components
after the product has been released to production. In this case, the signal generators
uncovered a latent design flaw that had not been detected using a BER tester.

Software Update Procedure


The embedded operating software of the Signal Forge 1000 and 800 Signal
Generators may be updated in the field by following the procedure described in this
document. Any available software updates may be on the Support page of the Signal
Forge web site www.signalforge.com.

Testing Amplifiers with the Signal Forge 1000


This application note describes a low-cost way to test amplifiers for gain as well as for
the 1dB compression point.

Testing Digital Systems


The Signal Forge 1000 / 800 may be used to test the design margins of a digital
system by varying the clock input to the system from min to max. This test ensures
that the design maintains the required setup and hold times under all conditions.

Testing High-Speed Serial Busses


Proper operation of high-speed serial busses requires that data integrity be maintained
between the two devices under test (eg. a disk controller and a disk drive) even though
these devices are not being driven from the same clock source. Since operation and
compatibility must be guaranteed across a range of frequency variations and
manufacturing variances, it is imperative to test many device samples under varying
conditions. This application note discusses how to configure a test serial bus device
using the Signal Forge 1000 or Signal Forge 800.

Testing IP3 with the Signal Forge 1000 and 800


This application note describes how to build a low-cost IP3 tester for amplifiers.

Testing Wireless Data Transmission Systems


Amplitude Shift Keying (ASK) and On/Off Keying (OOK) are two techniques used to
test and exercise digital data transmission systems such as wireless security systems,
keyless entry systems and garage door openers. This application note describes how
to use the ASK and OOK waveform modulation functions of the Signal Forge 1000
Signal Generator for testing wireless digital data transmission systems.

USB to Serial Adapter


Keyspan's USB - Serial adapter makes the Signal Forge 1000 or Signal Forge 800
Signal Generator accessible from a USB port.
The Keyspan USB Serial adapter model 19HS may be purchased online directly from
the Keyspan web site:

Oscilloscope
An oscilloscope (commonly abbreviated to scope or O-scope) is a type of electronic test
equipment that allows signal voltages to be viewed, usually as a two-dimensional graph of
one or more electrical potential differences (vertical axis) plotted as a function of time or of
some other voltage (horizontal axis). The oscilloscope is one of the most versatile and widelyused electronic instruments. [1]
Oscilloscopes are widely used when it is desired to observe the exact wave shape of an
electrical signal. In addition to the amplitude of the signal, an oscilloscope can measure the
frequency, show distortion, and show the relative timing of two related signals. Oscilloscopes
are used in the sciences, medicine, engineering, telecommunications, and industry. Generalpurpose instruments are used for maintenance of electronic equipment and laboratory work.
Special-purpose oscilloscopes may be used for such purposes as adjusting an automotive
ignition system, or to display the waveform of the heartbeat.
Originally all oscilloscopes used cathode ray tubes as their display element, but modern
digital oscilloscopes use high-speed analog-to-digital converters and computer-like display
screens and processing of signals. Oscilloscope peripheral modules for general purpose
laptop or desktop personal computers can turn them into useful and flexible test instruments.

What does an oscilloscope do?


An oscilloscope is easily the most useful instrument available for testing circuits because it
allows you to see the signals at different points in the circuit. The best way of investigating an
electronic system is to monitor signals at the input and output of each system block, checking
that each block is operating as expected and is correctly linked to the next. With a little
practice, you will be able to find and correct faults quickly and accurately.
An oscilloscope is an impressive piece of kit:

The diagram shows a Hameg HM 203-6 oscilloscope, a popular instrument in UK schools.


Your oscilloscope may look different but will have similar controls.
Faced with an instrument like this, students typically respond either by twiddling every knob
and pressing every button in sight, or by adopting a glazed expression. Neither approach is
especially helpful. Following the systematic description below will give you a clear idea of
what an oscilloscope is and what it can do.
The function of an oscilloscope is extremely simple: it draws a V/t graph, a graph of voltage
against time, voltage on the vertical or Y-axis, and time on the horizontal or X-axis.
As you can see, the screen of this oscilloscope has 8 squares or divisions on the vertical
axis, and 10 squares or divisions on the horizontal axis. Usually, these squares are 1 cm in
each direction:

Many of the controls of the oscilloscope allow you to change the vertical or horizontal scales
of the V/t graph, so that you can display a clear picture of the signal you want to investigate.
'Dual trace' oscilloscopes display two V/t graphs at the same time, so that simultaneous
signals from different parts of an electronic system can be compared.
Function of an oscilloscope
The function of an oscilloscope is to be able to display waveforms on some form of display. In
the normal mode of operation time is displayed along the X-axis (horizontal axis) and
amplitude is displayed along the Y axis (vertical axis). In this way it is possible to see an
electronic waveform on an oscilloscope as it may be envisaged. The waveform could be
likened to that of the ripples on traveling along the surface of a pond when a stone is dropped
into it.
Features and uses
By seeing a waveform in this manner it is possible to see analyze the operation of the circuit
and discover why any problems may exist.
Oscilloscope exterior
An oscilloscope will normally have a large array of items on the exterior of the case. The font
panel will typically have a number of items on it:
1. Display The first things that is noticed on an oscilloscope is the large display that is
used for displaying the waveform. This typically may take around a quarter of the
space on the front panel or even a little more. It is often good to have a reasonably
large display then it is easier to see the various elements of the waveform.
2. Connectors There is a variety of different connectors on the front panel. Typically
there is an input for each of the channels to be displayed - often an oscilloscope will
have more than one channel. Many oscilloscopes are dual channel and can therefore
display two signals at the same time, allowing waveforms to be compared. Other
inputs may include a trigger input that will enable the trace on the oscilloscope to be
triggered according to this signal.
3. Controls There is a variety of controls on the oscilloscope:

Vertical gain / signal input sensitivity: This is generally calibrated in V/cm, i.e.
each vertical division on the scale represents a given number of volts.
o Timebase: This alters the speed at which the trace crosses the screen
horizontally on the oscilloscope. It is calibrated in terms of time / division, e.g.
1ms / cm, assuming the divisions are at one centimetre intervals.
o Trigger: The controls that are associated with the trigger enable the timebase of
the oscilloscope to be triggered in various ways. This enables a still or stable
picture to be obtained on the screen of the oscilloscope.
o

In order to be able to operate the oscilloscope correctly it is necessary to connect the right
signals into the inputs, and also to use the controls correctly.

Operating an oscilloscope
Like any other piece of complicated test equipment, an oscilloscope can take a few minutes
to get used to if one has not been used before. However once familiar with it, the controls
soon become second nature and it becomes very easy to use.
It is obviously necessary to turn the oscilloscope power on, and then once it is running it may
be necessary to adjust the intensity of the trace so that it is easily visible - often oscilloscopes
will free run when no signal is present. If the oscilloscope does not free run, then no trace will
be seen yet.
Then the next control to set is the vertical gain control. Set is so that the anticipated
waveform will fill a reasonable amount of the screen. Leave some margin so that if it is bigger
than expected, it will not go wildly off the screen.
Next set the time base of the oscilloscope. This is often set so that a period of the waveform
will fill most of the horizontal axis of the screen. If it is initially set to this then it can be
adjusted to suit later.
Connect the signal to be viewed. The oscilloscopes will posses a connector for the input - this
is normally a BNC connector. In most cases where a connection to a circuit board is required,
a scope probe will be used, so that they are easy to connect to pins or connection points on
the board.
With the signal now present it is necessary to adjust the trigger control to gain a stable trace
of the signal.
With a trace of the signal now visible, the vertical gain and timebase controls can be readjusted to produce the best picture of the signal.
Although these instructions do not give an exhaustive description of how to use an
oscilloscope, the exact number of controls and operation will depend upon the particular
scope in use. However they should enable the scope to be used in a basic but reasonable
manner.
SIGNAL GENERATORS and
OSCILLOSCOPE CALIBRATION

This paper shows how standard signal generators can be used as leveled sine wave sources
for calibrating oscilloscopes. To do so, it is necessary to show that signal generator
specifications match, or can be made to match, oscilloscope calibration requirements.
Assume that oscilloscope calibration equipment is currently in place and performs
adequately, including leveled sine wave performance up to 500 Mhz. Also assume a need to
increase the frequency capability of the leveled sine wave source for new, higher frequency,
oscilloscopes. A good place to start looking at performance requirements for a signal
generator is the performance of existing calibration equipment. The Tektronix SG 5030 is a
commonly used leveled sine wave generator.
SG 5030 specifications
Frequency
0.1 Hz to 550 MHz
Frequency accuracy
+/- 3 ppm
Amplitude Range
4.5 mvpp to 5.5 Vpp
Amplitude accuracy
from 0.1 Hz to 50 kHz --- +/- 1.5 % of setting
Flatness (relative to 50 kHz amplitude)
+/- 1.5 % from 50 kHz to 100 Mhz
+/- 3 % from 100 Mhz to 250 Mhz
+/- 4 % from 250 Mhz to 550 Mhz
VSWR
< 1.2:1
Harmonic distortion
all </= -50 dBc from 0.1 Hz to 49.999 kHz
2nd Harmonic </= -30 dBc from 50 kHz to 550 Mhz
all other Harmonics </= -35 dBc from 50 kHz to 550 Mhz
all Non-Harmonics </= -40 dBc from 50 kHz to 550 Mhz
On the other side of the coin are the capabilities of available signal generators. HP, Rhode &
Schwarz, Anritsu, and many others are happy to provide signal generators. If we limit the
search to the lower price range ($5K to $15K), there are available frequency and level
characteristics as below:
Maximum Level @ Frequency
+ 10 dBm to + 20 dBm @ 1 GHz (or 1.9 to 6.26 Vpp )
+7
+ 19
2
(or 1.25 to 5.57 Vpp)
Frequency stability
</= +/- 3ppm
Level accuracy @ Frequency
+/- 0.5 dB to +/- 1.5 dB to 1 GHz (or 5 to 20 %)
0.9
1.5
2
(or 11 to 20 %)
Harmonic distortion
</= -30 dBc (all) specs are below levels from +4 to +10 dBm or 1 to 1.9 Vpp
Sub-Harmonic distortion
</= - 40 dBc
Non-Harmonic distortion
</= - 50 dBc

SWR
<1.5:1 for frequency <2 GHz
A quick look says that the signal generators may be marginal in maximum output capability,
VSWR, and harmonics and are certainly not accurate enough in level. At least this is the
case in comparison to the SG 5030. We need to go a step further and look at calibration
procedure requirements. If we look in the procedures for digitizing oscilloscopes with high
frequency capabilities in the 1 to 1.5 GHz range, we see another set of leveled sine wave
requirements.
Maximum Level @ Frequency
600 mvpp @ 1 GHz
1.2 Vpp @ 1.1 GHz
500 mvpp @ 1.5 GHz
1.2 Vpp @ 1.5 GHz
(plus optional tests to 5 Vpp @ 1 GHz )
Accuracy -- Accuracy requirements are not specified in procedures but test
equipment is. We need to look at the test equipment specs to determine the accuracy
requirements.
Power meter with +/- 3% accuracy (to set levels)
Signal generator
Level available - + 19 dBm @ 1 GHz ( 5.6 Vpp )
Level accuracy - +/- 1.5 dB ( +/- 20 % )
Harmonics - </= -30 dBc for levels </= + 10 dBm
Source impedance - nominal 50 ohms
VSWR -- </= 1.5:1 for frequency </= 2 GHz
In fact, these procedures call out a signal generator as the leveled sine wave source together
with a power meter to provide the required level accuracy. Since the leveled sine wave
source is used to determine the bandwidth ( and flatness ) of the oscilloscope, we are
concerned with the signal generator specifications as they relate to level accuracy. We need
to relate harmonics and source / load match effects to level accuracy
Before continuing this path, we should look to see if there are other elements in the system
which will contribute problems. The obvious one is that there needs to be a connection
between the signal generator and the oscilloscope. This can be a very short cable in a
simple test setup or can be considerable cable length together with switches in an automated
factory, or calibration lab, test set. Fig. 1 shows a typical interconnection.

Good quality Co-axial cable will show attenuation of approximately 0.3 dB / meter at 1 GHz.
In the ideal case, we have no other losses and no VSWR problems. Fig. 2 shows the loss vs
frequency characteristic of this example.

Add in some typical loss, impedance, and VSWR numbers.


switch -- 0.3 dB loss and 1.3:1 VSWR @ 2 GHz
Co-ax -- 0.3 dB per meter loss @ 1 GHz and impedance from 49 to 51 ohms
Signal generator -- Output SWR of 1.5:1
This yields the loss vs frequency characteristic shown in Fig. 3.

This error, maximum of -1.4 dB at 2 GHz, is correctable if our chosen signal generator
has sufficient extra power available. In order to examine some realistic numbers we use the
HP 8648B in standard configuration. This unit provides + 13 dBm ( 2.8 Vpp ) minimum. The
highest requirement noted has been 1.2 Vpp (~ +6 dBm ) at 1.5 GHz. To make a worst-case
example, assume that the 1.2 Vpp requirement is at 2 GHz. In order to correct for the 1.4 dB
loss, the signal generator needs to provide + 7.4 dBm ( ~ 1.5 Vpp ). This is still within the

capability of the HP 8648B. Now we can look at level correcting our test set in order to obtain
the necessary level accuracy.
There are at least three methods available to accomplish automatic level correction.
The oldest method is the analog leveling loop.

This has the advantages of acceptance by tradition. The accuracy is as good as that of the
detector used.
The disadvantages include hardware (some custom), difficulty in applying calibration
corrections, power loss (6 dB when using a 50 ohm splitter), and low level limitations due to
detector and loop problems. In fact, our example will not work with this method. The 6 dB
loss due to the splitter requires that the signal generator be capable of a + 14.4 dBm output
level. An unequal power splitter or a signal generator with higher output level capability
would be required.
Another method is to use the internal level correction capability of the signal generator. The
signal generator service manual should detail the procedure. The advantages are that the
end system requires no extra equipment, the power meter corrections (as from NIST) can be
incorporated, and no additional loss is added to the system. The disadvantages are that the
available internal correction range may not be adequate and that the correction points may
be too few or not optimally placed for the application. Another possible problem would be if
the signal generator manufacturer required a special test set to do the level correction. This
method could likely be made to work with our example.
The most versatile method uses a computer to store and apply the corrections as external
level commands.
This allows corrections to be applied in whatever manner gives the best accuracy and allows
the use of a factory standard signal generator. The disadvantage is that a computer is then
required in order to run tests and oscilloscope calibration routines. In any automated system
a computer is a part of the system and no penalty is accrued. The most simple and accurate
way to use this capability is to make a correction list, or table, with all of the frequency/level
points of interest and have test software use this table to provide the level corrected signals.

A more general, but less accurate, method is to make a frequency/level array that covers all
possible points of interest and use an interpolation routine to determine the correction values
for points that fall between the calibration points in the array. This requires more knowledge
of the signal generator architecture and performance. Internal switch points for attenuators
and other circuit elements will affect the choice of calibration points for the correction array.
The best starting point for this array is the internal level correction array. Calibration points
can be added, changed, and subtracted as accuracy requires.
The accuracy which can be achieved by this technique depends on the accuracy of the
power sensor calibration, the match between the source and the power sensor, and
temperature sensitivity.
Power measurement uncertainties example
Sensor calibration
+/- 1 %
Source / power sensor match (source p = .08, sensor p = .1)
+/- .8 %
Temperature
This will be added later as an overall measurement effect.
The RSS sum of these components gives an uncertainty of +/- 1.3 %. We can expect the
level characterization of the source to be accurate within 1.3% at the measured points.
NOTE: The RSS sum is defined as the square root of the sum of the squares of the
individual components.
The power level errors of the signal generator and the loss contributed by interconnecting
hardware are both correctable errors. Considering only these factors, we can get
performance as good as traceability and uncertainty allows. There are some factors that are
not correctable, at least in a practical sense. These may be factors which are not directly
specified in terms of level effects. In these cases, the effect on level must be derived. Some
factors which may not be correctable include:
Harmonic distortion (can have large effect)
Temperature (effect should be small in calibration environment)
Amplitude modulation (should be small with a good signal generator)
Source and load match interactions (may be difficult to quantify and may be large)
Harmonic distortion can be a problem because it is a phase dependent contributor to
peak/peak voltage and a phase independent contributor to power measurements. This
means that a signal with a particular harmonic component can be viewed on a spectrum
analyzer while the phase of the harmonic is changed and no change will be observed on the
spectrum analyzer display. Performing this same experiment while applying this signal to a
power sensor, will likewise indicate no changes in power level. If the signal is applied to an
oscilloscope and the phase of the harmonic is changed, there will be changes in the
observed signal. The output level of the signal generator is being measured, and corrections
determined, by a power meter. The measurements that are made using the signal generator
are done with an oscilloscope. This leaves an error term which is due to harmonic
components on the signal.
A general expression for a sinusoidal signal is:
A0 + A1 sin (wt + z1) + A2 sin (2wt + z2) + A3 sin (3wt + z3) + . . . .
dc
fundamental
2nd harmonic
3rd harmonic

The effect of adding harmonic components to the fundamental can be seen by doing this
mathematically and rotating the relative phase of the harmonic.

Notice that the peak/peak level does not change but a dc component is added.

Notice the variation in the peak/peak level.


Some facts about harmonics:
1) Third harmonic is the most important when peak-to-peak voltage is the quantity of interest.
The uncertainty due to the third harmonic is:+/- [ 10 exp (-dBc/20) ] % example ---30dBc > +/- 3 %
2) The second harmonic adds a similar dc component (uncertainty?) which is usually not
important.
3) Higher order harmonics and, in some cases sub-harmonics, have similar effects but are
usually at much lower levels.

The effect of temperature on level accuracy is usually lumped into the overall level accuracy
specification by the signal generator manufacturer and is not easy to extract. It is, luckily, a
small contributor to the specification. Flatness and attenuator errors are the large
contributors. Much of the signal generator error budget is tied up with very low level ( 1 mv
and below ) signals which are of no concern in an oscilloscope calibration application. Of
course, the better the environment is controlled, the less this is of concern.
Some reasonable numbers might be:
< +/- 1 % if the temperature is held to < +/- 2 deg C
< +/- 2 %
< +/- 5 deg C
A poor ( or broken ) signal generator might have power line related amplitude modulation.
This would likely first appear as a noisy signal until identified as power line frequency related.
Normally, there should be no amplitude modulation related level errors.
Look at a picture to better understand the mismatch error problem.

The error is +/- p1*p2 % example --- p1 = p2 = .1 (VSWR = 1.22) then error is +/- 1 %
Since we do not know actual numbers only max allowable limits, we must use the max
numbers to calculate the possible error and use this in our error budget. Our example scope
has a VSWR of </= 1.5:1 from 500 Mhz to 1 GHz, or p = .2. Our example signal generator
has a VSWR of </= 1.5:1 for frequency < 2 GHz, or p = .2. Our error from this source is then
+/- 4 % maximum. We have no control over the oscilloscope input VSWR. The signal
generator output match can be improved if there is excess signal power available after level
correction. In that case, we can insert an attenuator pad in the system. This must be done
prior to any level correction, of course. In our example we do have an excess and can insert
a 4 dB pad and still maintain some headroom. This improves the signal generator output
match to </= 1.17:1, or p = .08. The error is reduced to +/- 1.6 %. Improving the signal
generator output match also reduces the mismatch error contribution in the level correction
process.
Let us put all of these errors and uncertainties together and see the limits of our calibration
process when we use a signal generator as the leveled sine wave source. First, use pointby-point corrections and a good source/load match (obtained by adding a pad to the system).

Level correction error


Source / power meter mismatch
Certification uncertainty
Source resolution error (0.1 dB steps)
Harmonic error (use spec value of - 30 dBc 3rd harmonic)
Temperature effect error
Source / load (source / oscilloscope) mismatch error

+/- 0.8 %
+/- 1 %
+/- 0.5 %
+/- 3 %
+/- 1.5 %
+/- 1.6 %

The RSS error in this case is +/- 4 %.


As a second example, repeat this calculation but add an interpolation error of +/- 1 % (array
method of level correction).
Level correction error
Source / power meter mismatch
+/- 0.8 %
Certification uncertainty
+/- 1 %
Source resolution error (0.1 dB steps)
+/- 0.5 %
Harmonic error (use spec value of - 30 dBc 3rd harmonic)
+/- 3 %
Temperature effect error
+/- 1.5 %
Source / load (source / oscilloscope) mismatch error
+/- 1.6 %
Interpolation error
+/- 1 %
Linearity error
+/- 1 %
The RSS error in this case is +/- 4.2 %
As a final example, repeat the second calculation but increase the source/load mismatch
error to 4 % by removing the pad.
Level correction error
Source / power meter mismatch
+/- 0.8 %
Certification uncertainty
+/- 1 %
Source resolution error (0.1 dB steps)
+/- 0.5 %
Harmonic error (use spec value of - 30 dBc 3rd harmonic)
+/- 3 %
Temperature effect error
+/- 1.5 %
Source / load (source / oscilloscope) mismatch error
+/- 4 %
Interpolation error
+/- 1 %
Linearity error
+/- 1 %
The RSS error in this case is +/- 5.6 %
As a point of comparison, the SG5030 yields an RSS error of 4.9% at 550 Mhz. The errors
are distributed somewhat differently but they add up about the same. There appears to be no
fundamental problem with using a signal generator as the leveled sine wave source for
oscilloscope calibration.