AnalogTesting02
AnalogTesting02
Analog Testing
Objectives
This chapter contains information to help you:
• Understand how sources and detectors are used for analog in-circuit testing.
• Have a general understanding of analog in-circuit measurement techniques.
• Be able correct the following types of measurement errors: source voltage,
source loading, guard offset, and current splitting.
• Understand the process to discharge capacitors.
• Understand the structure and purpose of analog test blocks.
• Know which test statements can be used in analog in-circuit tests.
• Understand the purpose of the Measuring Operational Amplifier (MOA).
2 Analog In-Circuit Testing
Measuring Response
Stimulus
Sources Detectors
Operational (A/D)
I or E Amplifier AC / DC
AC or DC Circuit or
Phase Sync.
ASRU Card
The Module Control Card manages each in-circuit test by closing the proper
testhead relays to connect the device under test into the MOA circuit. Passive and
reactive devices, such as resistors and capacitors, are connected into the input path
of the MOA. Active devices, such as diodes and transistors, are connected into the
feedback loop of the MOA. A test for each type of device is shown in detail in Analog
Tests: Reference.
The stimulus sources and response detectors, selected by the test program, are also
connected by the Module Control Card. As the stimulus is applied to the MOA
circuit, the response detector measures the output of the MOA and sends the
results to the Module Control Card for evaluation. Depending on the results, the
Module Control Card sends either a pass or fail condition back to the test program.
Source Use
DC voltage sources Test resistors, channel resistance of FETs, fuses, jumpers,
potentiometers, switches.
AC voltage sources Test capacitors and inductors. Frequencies of the AC sources:
128/1024/8192 Hz. When selecting 128 Hz, always use the ed
option, which integrated the measurement over a complete line cycle
(normal integration time is insufficient for making valid
measurements).
DC current sources Test diodes, zeners, and npn or pnp transistors.
Even though the output voltage from the MOA is limited by its power supply, a wide
range of Rx values can be measured. This is accomplished by switching different
values of the reference resistor, Rref, into the feedback path of the MOA. Six
different values of precision reference resistors are available on the ASRU card.
These resistors are illustrated in Figure 2-2 on page 2-6.
By using an AC stimulus source, and an AC detector, reactive components
(inductors and capacitors) can also be measured by this technique. Reactive
component measurement, called Phase Synchronous Detection, is explained next.
Guarding (G Bus)
The device under test may have one or more parallel impedance paths due to the
circuit topology of the board under test. These parallel impedance paths cause
measurement errors by providing current paths around the component under test.
Zsg and Zig, in Figure 2-3, represent the components forming parallel impedance
paths around the component under test, Rx. When such parallel paths are formed,
parallel current, Ip, flows around Rx and through the MOA feedback path. The added
current through the feedback paths cause Rx to appear as a smaller impedance
than it actually is.
VS
Zsg Zig
Ip Vmoa
MOA
The G bus is used to guard the component under test by breaking parallel
impedance paths. Figure 2-4 shows where the G bus is connected in the circuit
when the component under test is shunted by a parallel impedance path. By
connecting the G bus as shown, the current that would flow through both Zsg and
Zig becomes insignificant.
When the non-inverting input to the MOA is grounded as shown in Figure 2-4, the
inverting input becomes a virtual ground due to characteristics of the op amp. This
also places the I bus connection at virtual ground. With the G bus also at ground
potential, no difference of potential exists across Zig, and no current flows through
the parallel path around Rx and through the MOA feedback path. Vs does, however,
supply current to Zsg. This current does not affect the measurement as long as the
Vs output impedance is very low compared to Zsg. Because there may be one or
more parallel paths around the device under test, there may be one or more G bus
connections.
ig
Z
VS
G Bus Relay
Vmoa
MOA
Gain
Wide
Band
Narrow
Band
Frequency
Sensing
Three-wire measurements, using the S, I, and G buses, provide the basic
configuration for analog in-circuit testing on the i3070 In-Circuit Test System.
However, the addition of three more buses, A, B, and L, help overcome the
measurement problems associated with in-circuit testing. Each of the three
additional buses is paired with one of the three basic measurement buses, S, I, and
G. The buses are paired as follows:
S to A I to B G to L
Sensing the source requires enhancement measurements. Enhancement is
explained later in this chapter. To invoke enhancement you must specify the en
option in addition to the sa option.
The added buses provide the system with a six-wire, in-circuit measurement
technique. The extra buses are used only when the measurement situation requires
them. Figure 2-6 illustrates the location of these buses in the measurement circuit.
A Bus
Rs Ri Rref
Rx
B Bus
Z sg
ig
Z
VS
Vmoa
L Bus MOA
Rg
The example test block in Example 2-3 shows the connections and the
measurement options.
Sense Methods
This section describes the sense methods that can be used on the system to
generate more accurate measurements for some in-circuit tests. If you turn on
Remote Sensing, the development software determines the sensing needs of each
test. Depending on the device under test, topology of the circuitry, and the
accuracy of the test desired (according to the tolerance multiplier that you specified
when entering board data), any combination of the A, B, and L buses may be
required.
If sensing is required, the sense connections can be made in the scanner or they can
be made by seperate wires to the board under test. If the remote sensing flag is
turned on and the development software determines that they are necessary,
seperate sense wires will be used. If adequate sensing can be performed in the
sanner, then seperate wires are not required.
Example 2-4 shows the differenece between the two sense methods. Notice that
the only difference between the G and L sensing levels is in the way that the clear
connect statement is constructed.
Example 2-4
! This test uses seperate wire S bus sensing.
clear connect s to "R1_1"; a to "R1_1"; i to "R1_2"; g to "Q5e"
resistor "R1", 10.1k, 2.5, 2.5, sa, en
Enhancement
In difficult situations, you can use a measurement technique known as
enhancement. Enhancement makes use of a more complex system measurement
equation to compensate for:
• The measurement errors caused by the component topology of the board
under test.
• The non-ideal characteristics of the operational amplifier.
• The thermal offsets of bus and testhead relays.
Enhancement is invoked with the en option in the measurement statement.
Note that sense a (sa) requires enhancement (en), but enhancement can be used
by itself.
When enhancement is specified, voltages are measured at the device under test
and the reference resistor of the MOA. These measured values, instead of the
default values, are used in the formula that determines the value of the device
under test. This compensates for the voltage drop across the measurement buses,
for thermal offsets created by the system relays, and for DC offsets in an AC
measurement.
Figure 2-7 on page 2-13 shows the measurements taken when enhancement is
used. For DC measurements, voltages are measured at the four points shown with
the source set to zero, and again with the source set to the needed level (eight extra
measurements).
For AC measurements, voltages are measured at the four points shown with the
source signal at zero degrees, again with the source at +90 degrees, and again with
the source at -90 degrees (12 extra measurements).
The extra measurements of enhancement cause the device test to require
significantly more time. If test time is a concern, use enhancement only where
necessary.
VS
Vmoa
MOA
Figure 2-8 Bus impedances and thermal offsets cause measurement errors
Rs E DUT Ri Rref
Rx
I + I1 I + I2
I2
Z sg
ig
E
Z
VS
I1
Vmoa
E MOA
Rg Too Large
Three major types of measurement errors are caused by the problems illustrated in
Figure 2-8. These errors are: source loading (or source voltage error), guard offset
error, and current splitting. They are summarized in Table 2-3.
Error Cause
Source Voltage Error Is caused by the thermally induced offset voltages of relay contacts, impedances of the
measurement buses, and loading of the stimulus source by small parallel impedances on the
board under test. See Source Voltage Error and Solution on page 2-15 and Guard Offset Error
and Solution on page 2-17.
Guard Offset Error Is caused by the voltage drop across the impedance of the G bus. This voltage drop appears as
an offset voltage at the input terminals of the MOA. Guard offset error is sometimes called
guard gain. See Guard Offset Error and Solution on page 2-17.
Current Splitting Is caused by a small impedance path in parallel with the input to the MOA, and by an apparent
increase in the input impedance of the MOA when AC stimulus sources are used. MOA input
impedance increases as the frequency of the source is increased. A large value reference
resistor in the MOA feedback path also contributes to current splitting.
See Current-Splitting Error and Solution on page 2-19.
See also:
• Discharge Capacitors on page 2-21
Figure 2-9 Bus impedances and thermal offsets cause measurement errors
Rs and Ri are the S and I bus impedances
Rs Ri Rref
Rx
I DUT
VS Vmoa
Vapplied MOA
Too Small
Source Voltage Error = Vs – Vapplied
In Figure 2-9, resistors RS and RI represent the impedances of the S and the I bus,
respectively. The voltage drops across these resistances subtract from VS and the
remaining voltage is applied across the component under test. This smaller source
voltage, Vapplied, across the device under test, results in less feedback current, I,
through the feedback resistor, Rref. With less feedback current, the output voltage,
Vmoa, is also smaller than expected. This causes the measured value of the
component under test to be higher than its actual value.
Source voltage error can be reduced by sensing the source with the A bus. Sensing
the source also requires enhancement (en). When the A bus is connected and the
measurement options sa and en are specified, the value of the source is measured
at the device under test. The measured value, instead of the default value, is used
in the formula that determines the value of the device under test. This compensates
for the voltage drop across the S bus, for thermal offsets created by the system
relays, and for DC offsets in an AC measurement.
Example 2-6
clear connect s to "R1-1"; a to "R1-1"; i to "R1-2"
resistor 2k, 1.4, 1.4, wb, sa, en
Note that the use of the sa option alone does not improve the accuracy of the
measurement. The sa option must be accompanied by the enhancement option, en.
Source voltage error correction can also be improved by sensing the I bus with the
B bus. The B bus connection, shown in Figure 2-10, moves the virtual ground of the
MOA to the component under test. This places the impedance of the I bus in series
with the reference resistor, Rref. The inverting (-) input to the MOA can no longer
sense the voltage drop across the I bus. To incorporate the B bus in an in-circuit
test, it must be included in the clear connect or connect statement and the option
sb must be specified in the options field of the analog in-circuit test statement.
Figure 2-10 Use of the B Bus (sb) reduces source voltage error
A/D1
Rref
S Bus A Bus I Bus
Rx
B Bus Relay
VS
B Bus
Vmoa
MOA
page 2-15. The resulting MOA output causes a larger measured resistance value
than the actual resistance value for the component under test.
Because source loading is a source voltage error, you can compensate for source
loading the same way as for source voltage error.
Z sg
VS
Rg Vmoa
MOA
Too Small
ig
Z
VS
I1
Verror – Vmoa
Vg MOA
+ – +
Vg = Voltage Drop Across Rg Rg
The voltage drop across Rg causes most of current I2 to flow through Zig as shown.
The voltage is multiplied by the ratio of Rref to Zig and appears as an error voltage
at the output of the MOA. The equivalent circuit showing how this error voltage
appears to the MOA input is illustrated in Figure 2-13. The increase in current
through Rref makes the measured value of Rx smaller than its actual value (Vmoa too
large).
VS Zig
I2
Vmoa
Ve MOA
Too Large
Vs = Source Voltage
Ve = Error Voltage
The L bus is used to correct for guard offset error. The L bus connects the
non-inverting input of the MOA to the G bus connection. This eliminates the offset
voltage (the voltage drop across the G bus) normally sensed by the non-inverting
input of the MOA. The development software uses the L bus, when it is needed, with
the sl option in the statement's measurement option field. Figure 2-14 illustrates
the electrical connection of the L bus. If you turned Remote Sensing on, the
software determines which tests require sensing of the G bus. The fixture reports
include the extra L bus wires.
Figure 2-14 The L Bus (sl) corrects for guard offset error
S Bus Rref
Rx I Bus
Z sg
ig
Z
VS
Vmoa
MOA
L Bus
G Bus
Unlike the A and B buses, you can use testhead level sensing with the L bus. That
is, connect gl to <node> connects the G and L buses together inside the testhead
and there is no additional wiring required in the fixture. You can still connect the L
bus to the board under test with:
"connect g to <node>; l to <node>"
The development software determines the level of sensing necessary based on
device values and topology of the circuit. See the examples below.
As with the A and B buses, you can use one probe for both wires or separate probes
with the L bus. The software always assumes a single probe, and automatically
assigns the resource and fixture wire.
To use the L bus in a component test, you must include it in the clear connect, or
connect statement, and you must add the sl option to the component test
statement:
Example 2-8
! This test uses testhead level sensing.
! This test uses separate wires to sense at the device under test.
I2
ig
Z
VS
Rg Vmoa
MOA
Figure 2-16 The B Bus (sb) and maximum moa input-to-ground impedance (Zig)
reduces current splitting
S Bus I Bus Rref
Rx
I
VS Zig
B Bus
Maximize Zig Vmoa
MOA
G Bus
Other steps that you can take to reduce current splitting error are: use an AC source,
lower the frequency of the source, and change the bandwidth of the MOA.
In summary, refer to Figure 2-17. To reduce the effects of current splitting:
• minimize Ierr by placing the S and I buses such that maximum circuit
impedance is between the I and G buses.
• minimize the voltage drop across the I bus by using the B bus.
• increase the gain of the MOA by changing the bandwidth.
• lower the frequency of the AC source.
Ierr
RI Bus
Vs
Rref
A+1
Discharge Capacitors
The discharge capacitors feature lets you discharge capacitors that can affect
analog in-circuit tests. Capacitors may already be charged when the board is
placed on the fixture, or they may charge as a result of applying sources during
in-circuit tests, or applying the DUT power supplies during powered testing.
The i3070 In-Circuit Test Software determines which capacitors need discharging
and writes an analog test block that discharges them with the S and G buses. The
discharge algorithm attempts to discharge the board to 0.1 volt or less. The
maximum voltage that can be discharged is 100 volts.
The discharge algorithm returns a variable to indicate its status as shown in
Table 2-4.
Value Status
0 No discharge needed
1 Discharge successful
2 Voltage too high to discharge (> 100 volts)
3 Charge not decreasing
4 Charge decreasing but did not reach the exit level
5 Unable to make measurement: hardware problem
6 Discharge not executed: over-voltage error
The dps statement executes the discharge algorithm once to remove excess
trapped charge from the board. The board might not be discharged all the way to
0.1 volt.
These statements execute a dps which includes a single discharge:
• All statements that turn on vacuum, such as faon, fbon, and fabon, when
executed in the unpowered mode.
• scratch board and load board
• fixture lock and fixture unlock
• cps and dps
• powered and unpowered
• autoadjust, confirm, and learn fixture timing
• Any statement during which an overvoltage error occurs.
• testhead is, testhead power on, and testhead power off
The discharge requires 0.1 to 0.5 seconds if no discharge is necessary. The time
required if a discharge is necessary depends on the number of nodes that need
discharging, the RC time constant of the discharge circuit, and the level of the
charge on the board.
For tests developed for multiple-board panels, all boards of one type on a panel are
discharged as one board.
You can add discharge blocks for problem areas. The discharge block is very similar
to an analog test block. It is executed by the test (BT-Basic) statement in the
testplan:
test "analog/discharge_C1"; Return
As shown in Example 2-10, the discharge block contains a clear connect
statement to connect the S and G buses across the capacitor to be discharged, and
the discharge statement. The discharge statement includes an optional name, an
entry voltage level, an exit voltage level, and an optional return variable.
Note that the block name must be unique and that if you want to evaluate the result
of the discharge test, you must include a return variable. To pass the return variable
from the discharge block to the testplan, you must use an explicit block. An explicit
block is delimited by the test analog and end test statements.
The entry voltage level specifies the voltage above which to start discharging
(0.1 minimum). If the voltage detected on the capacitor is not above this level, no
discharging takes place. The exit voltage level specifies the voltage below which to
stop discharging (0.005 minimum). If you do not specify the entry and exit voltage
levels, the system uses an entry level of 0.1 volt and an exit level of 0.05 volt.
A capacitor may discharge, and then, due to dielectric absorption or the circuitry
around the capacitor, become charged again. To make sure the capacitor is
discharged, you can loop the discharge block until the return variable of the
discharge statement is less than or equal to 1:
loop
test "analog/discharge_C1"; Return
exit if Return <= 1
end loop
To be sure this does not cause an infinite loop, limit the number of times the loop
can be executed. If the discharge is still not successful after the maximum number
of loops, stop the test:
Count = 1
loop
test "analog/discharge_C1"; Return
exit if Return <= 1
exit if Count > 10
Count = Count + 1
end loop
if Count > 10 then goto Cleanup
Note that the entry voltage level must be greater than or equal to the exit voltage
level.
The i3070 In-Circuit Test Software writes only implicit test blocks.
Passing Parameters
Note that implicit test blocks can only use local variables. If you want to pass
parameters either into, or out of, the test block you must use an explicit test block.
You might want to pass a parameter into a test block to programmatically control
some of the measurement options, or pass a parameter from the test block to the
testplan to evaluate the result of a measurement. See Example 2-13.
Testplan:
Ed_Flag = 0
test "analog/R1"; Ed_Flag
Test Block:
test analog; Ed_Flag
clear connect s to "R1-1"; i to "R1-2"
resistor 10K, 1.4, 1.4, re5, wb, ed Ed_Flag
end test
Passing a parameter to the testplan:
Testplan:
test "analog/R1"; R
if R < 7000 then print "R1 Failed"
Test Block:
test analog; Return
clear connect s to "R1-1"; i to "R1-2"
resistor 10K, 1.4, 1.4, re5, wb, Return
end test
the nominal value. The second measurement wider tolerances to accommodate any
inaccuracy in the adjustment.
Example 2-16
Notice that the one-line on failure statement does not use the end on failure
statement. You can have one or more print and report statements, each one
cancels the previous one. This is also true for on failure loops.
Example 2-17
on failure
report "Device failed."
exit test
end on failure
clear connect s to "Node_1"; i to "Node_2"
resistor "Div_1", 5k, 13.2, 12.5, wb, R
clear connect s to "Node_2"; i to "Node_3"
resistor "Div_2", 10k-R, 26.3, 24.4, wb, ed
The second measurement uses the value of R from the first measurement. If the first
measurement fails, the second one also fails. In this example the test exits after the
first failure. Use the off failure statement to disable an on failure loop.
Command Resul t
capacitor Measures fixed or variable capacitors.
connect-unpowered Connects specified buses to nodes or brc's.
clear Disconnects all buses then connects specified buses to nodes or brc's.
connect-unpowered
disconnect-unpowered Disconnects specified nodes or brc's.
discharge Discharges capacitors.
diode Measures forward bias diode or zener voltage.
end on failure Marks the end of an on failure loop.
end test Delimits the end of a test block.
exit test Exits a test block. Usually used with an on failure loop.
fuse Verifies the presence of fuses.
gpconnect Closes the specified GP relay. This statement coincides with the
BT-Basic gpconnect statement. See the syntax descriptions of both
statements for details.
gpdisconnect Opens the specified GP relay. This statement coincides with the
BT-Basic gpdisconnect statement. See the syntax descriptions of
both statements for details.
inductor Measures fixed or variable inductors.
jumper Verifies the presence of jumpers.
nfetr/pfetr Measures Ron of N-channel and P-channel FETs.
npn/pnp Calculates the gain of npn and pnp transistors by measuring two values
of DC base current.
off failure Turns off an active on failure statement.
on failure Marks the start of an on failure loop.
potentiometer Measures potentiometer resistance.
print Outputs a message to the printer is device.
report Outputs a message to the report is device.
resistor Measures fixed or variable resistors.
Command Resul t
switch Verifies the contact position of a switch.
test analog Delimits the start of a test block.
zener Measures zener reverse breakdown voltage.
Command Resul t
analog Invokes the analog mode.
gpconnect Closes specified general purpose relays.
gpdisconnect Opens specified general purpose relays.
learn capacitance off Turns off the capacitance learning.
learn capacitance on Turns on the capacitance learning. The capacitance compensation is
activated for the tests that are executed from the testplan between
this statement and the following learn capacitance off
statement.
minimum wait Specifies a wait interval, in seconds, between applying a stimulus and
taking a reading for all analog in-circuit tests.
printer is Specifies the printer device.
report is Specifies the report device.
test Executes the specified test block.
test cont Continues a test after a pause.
tolerance margin Changes the test limits specified in the in-circuit test statements. The
tolerance margin statement can be used to increase or decrease
the test limits. If no value is specified, the value used is 0.
unpowered Initializes the system to execute in-circuit test statements, and
executes the capacitor discharge block.
PC Board
Test Fixture
ASRU Relay
Vs Source
MOA Detector
The component under test is represented by resistor Rx. Under control of the test
program, Pin Card and ASRU Card relays switch the component to be tested into
the MOA circuit. When stimulus source voltage, Vs, is applied, the MOA output is
measured by a detector to determine the value of the component, Rx.
For more details, see the following topics:
• Understanding the Operational Amplifier
• Basic Op-Amp Equations
where Vmoa is the output voltage of the op-amp, Vs is the input (stimulus) voltage,
Rx is the component under test, and Rref is the feedback resistor.
However, the voltage gain equation does not take into account some of the
non-ideal circuit conditions. Figure 2-19 illustrates a basic op-amp circuit, and
some non-ideal circuit conditions in terms of currents and voltages.
Ib
Vs Source
Voltage
Vmoa
Vin MOA
The basic circuit components consist of the op-amp, an input resistor, Rx, and a
feedback resistor, Rref. If a source voltage, Vs, is applied at the circuit input, current
I flows through Rx. This current also flows into the summing node, N. If the op-amp
exhibits the ideal characteristic of an infinite input impedance, all of the current, I,
also flows through the feedback resistor, Rref.
However, because no op-amp is perfect, a very small bias current, Ib, flows into the
inverting input. Ib is considered an error current because ideally, all of the current
should flow through the feedback resistor. The remaining current, (I – Ib), flows
through the feedback resistor, Rref.
Figure 2-19 also illustrates the voltages in the op-amp circuit. The voltage drop Vin
is an error voltage across the inputs of the op-amp caused by the error current Ib.
Vmoa is the output voltage of the op-amp.
(2) I = Ib + ( I – Ib )
( V s – V in )
(3) --------------------------- = I
Rx
( V in – V moa )
(4) ------------------------------------ = ( I – I b )
R ref
Equations (3) and (4) are derivations of Ohm's Law (I = E/R). Equation (2) can be
rewritten using equations (3) and (4) by substituting values for I and (I – Ib).
( V s – V in ) ( V in – V moa )
(5) --------------------------- = I b + -----------------------------------
-
Rx R ref
If both sides of equation (5) are divided by (Vs – Vin), the new equation is:
b + ( V in – V moa -)
1- = I----------------------------------------------
------
(6)
Rx ( V s – V in )R
ref
The value of Rx can be found by inverting both sides of equation (6) which gives the
equation:
( V s – V in )R
(7) ref
R x = -----------------------------------------------
I b + ( V in – V moa )
To further simplify equation (7), some assumptions can be made concerning the
MOA circuit. The first of these assumptions is that the error current Ib is so small
that it is insignificant, and this value in the equations can be set to zero. If Ib is small,
then the voltage drop it causes, Vin, is also small enough to be insignificant, so it
too can be set to zero. These assumptions are shown in equation (8).
( V s R ref )
(8) R x = ----------------------
-
– V moa
Equation (8) is the one most often used by the system to calculate the value of Rx,
and is merely a transposed version of the op-amp voltage gain equation:
( – V s R ref )
(9) V moa = --------------------------
Rx
Equation (9), however, does not always work for all in-circuit measurements. The
component topology of the board under test can sometimes exaggerate the
non-ideal characteristics of the op-amp, and the MOA circuit in general. When the
system encounters a difficult measurement situation, an expanded equation
equivalent to equation (7) is used. Use of the expanded measurement equation is
referred to as Enhancement.
PC Board
Test Fixture
Detector
Rref
ASRU Relay
Vs DAC Measurement
Amplifier
Source
Detector
Digitized Measurement Circuitry