IPC-TM-650 EN 2022 试验方法-- - 第519页
mechanical p robing met hods. Operators and probing equip- ment should b e tes ted in ability to r epeat elect rica l probe con- tacts. 6. 3 . 6. 1 Pr o b es fo r S in gl e - En d ed Tr a ns m is s i on L in e Measuremen…
The probes used in TDR systems typically use spring-loaded
contacting mechanisms and these should be checked peri-
odically to ensure proper operation. Statistical process control
methods and control charts can provide useful information
regarding the condition of the TDR system and its associated
accessories.
6.1.6 Measured Values
The units of the delay values
computed using waveforms acquired by the TDR system are
in seconds. Propagation delay, which is in units of time per
unit distance (typically, s/m), is determined as described in
Section 5.
6.2 Calibration
6.2.1 Verification Field Check – Check Standards
The
mechanical tolerances of air line check standards should be
verified using mechanical gauges at each use. Damaged air
lines should be repaired and recalibrated before use. They
should always be handled with care. The air line should also
be calibrated and documented periodically (not less than once
every two years) by a qualified certification laboratory and kept
in an environment safe from mechanical shocks, dust and dirt.
Dust and dirt degrade the fine threads of the connection and
damage the electrical mating surfaces. Also, some TDR
equipment manufacturers have requirements for the minimum
length of the air line artifacts. The user should check with the
manufacturer regarding limits. For differential impedance of
100 Ω, each channel can be checked with a 50 Ω air line.
6.3 Measurement System
6.3.1 Bandwidth/Risetime Resolution
The frequency
components of the TDR step are approximately related to the
bandwidth by:
BW
−3dB
≈
0.35
t
d
,
where
BW
-3dB
is the 3 dB attenuation bandwidth and
t
d
is the 10 - 90% transition duration of the TDR
step response.
Note that this relationship may not accurately represent the
intended operational frequencies of the transmission line
being tested. The bandwidth and risetime characteristics must
be adequate to ensure the TDR can provide a waveform
epoch appropriate to accurately determine t
d
for a transmis-
sion line of a given length. This waveform epoch must provide
sufficient resolution (see 4.1.2) to accurately determine the
reference level instants (see 5.1.3) and be long enough to
ensure the TDR waveform has settled to a nominal value (nec-
essary for accurate computation of pulse amplitude.) Risetime
considerations, however, are not the best method for deter-
mining TDR resolution. It is better to consider the temporal/
spatial resolution of the TDR (see 4.1.2) than bandwidth/
risetime resolution when determining the performance of the
TDR measurement system.
6.3.2 Temporal/Spatial Resolution
The TDR unit may
not be the only limiting factor for temporal resolution. The
probe connecting the TDR unit to the test specimen may also
limit resolution and this needs to be considered. Because of
the nature of TDR, it is easy to include the effects of the TDR
unit and all of the probe devices collectively, by defining t
sys
as
the fall time of the TDR step that has reflected from a short
circuit placed at the end of the probe and returned to the TDR
head.
6.3.3 Amplitude Scale
If a coarse vertical scale is used,
quantization error can be significant in certain instruments.
Many instruments change accuracy when their scales are
changed, and this can result in significant but unknown errors
in t
d
.
6.3.4 Baseline and Amplitude Drift
The ability of the TDR
instrument to maintain a constant baseline voltage and con-
stant amplitude step pulse are critical to the repeatability of
the TDR measurement process. TDR step generators and
sampling units are sensitive to time and temperature drifts.
Drift should be minimized and have a value that corresponds
to less than one-tenth the desired t
d
uncertainty.
6.3.5 Electrostatic Discharge Damage
ESD damage to
TDR instrumentation is often not easily detected and may
unknowingly affect measurement accuracy. Therefore, system
calibration should be performed regularly to check for this (see
5.1.1). All cables should have a termination attached to one
end when not in use and while they are being connected to
the TDR instrumentation. The use of a static protection switch
helps eliminate ESD damage to the TDR. Operators should
have anti-static awareness training and should perform all
measurements in anti-static work areas while wearing anti-
static wrist straps.
6.3.6 Probes
Hand-held probing solutions are sensitive to
operator technique and may have a larger contribution to
uncertainty due to repeatability of connections compared to
Number
2.5.5.11
Subject
Propagation Delay of Lines on Printed Boards by TDR
Date
04/2009
Revision
IPC-TM-650
Page
14
of
16
mechanical probing methods. Operators and probing equip-
ment should be tested in ability to repeat electrical probe con-
tacts.
6.3.6.1 Probes for Single-Ended Transmission Line
Measurements
The probe assembly impedance is often
chosen to be 50 Ω to match the impedance of the TDR sys-
tem. Impedance matching minimizes reflections at the inter-
face between the probe and the transmission line under test.
These reflections, which appear at and around the transition
region in the TDR pulse and can extend for some time after
this transition, are perturbations in the TDR waveform and are
undesirable because they may affect the computation of the
reference level instant, thereby increasing measurement
uncertainty. When the characteristic impedance of the trans-
mission line under test is nominally 50 Ω, these perturbations
will normally decay rapidly. If the impedance of the transmis-
sion line under test is significantly different from 50 Ω, the
magnitude of the perturbations can be large and their duration
long enough to affect the computation of the reference level
instant. The effect of these perturbations must be taken into
account when determining the appropriate waveform epoch
(see 4.1.2). The design and quality of manufacture of the
probe has a large effect on the magnitude and duration of
reflections generated between the TDR system and the trans-
mission line under test.
When probing non-50 Ω lines, it is possible to separate, in the
TDR waveform, the large signal perturbations caused by the
TDR/probe interface from those caused by the probe/
transmission line interface. To do this, a specially designed
probe is required that is impedance matched to the transmis-
sion line under test and that also has a long propagation delay
between the TDR/probe connection and the probe tip. The
long propagation delay can effectively move the large pertur-
bations at the TDR/probe interface out of the waveform
epoch.
6.3.6.2 Probes for Coupled-Signal-Line (Differential)
Transmission Line Measurements
The probe consider-
ations described in 4.3.3 apply for probes used in differential
transmission line measurements. However, the necessity to
simultaneously probe two signal lines and one or two refer-
ence plane contacts makes differential probing more difficult
than probing single signal line structures. In a PB manufactur-
ing environment, the use of two probes that were previously
used for single-ended measurements may not be possible.
This is because the operator is required to use both hands for
probing, which leaves them unable to operate the instrument.
Contact your instrument manufacturer for their probing solu-
tions and advice. Probes from one manufacturer can also be
used with another manufacturer’s TDR if the impedance val-
ues and connectors are compatible.
6.4 Adjustable Measurement Parameters
6.4.1 Sampling Interval (Point Spacing)
The temporal
resolution of the TDR unit is an issue only if it affects the dura-
tion of the transitions in the TDR waveforms (see 4.1.2) that
are used to compute t
d
. The temporal resolution of the TDR is
affected by the transition duration of the TDR step response,
the transition duration of the step response of all intervening
electrical components (connectors, cables, adapters), mea-
surement jitter, the interval between sampling instances, and
timebase errors. For typical TDR measurements, timebase
errors and sampling intervals should not be an issue (both are
or can be made to be less than 10 ps). The effect of measure-
ment jitter can be modeled by convolving the jitter distribution
with the TDR step response to yield an effective TDR step
response. The effect of jitter on the bandwidth of the TDR
measurement can be assessed from the jitter spectrum,
which can be described by:
J(,) = e
−2(πσ,)
2
,
where
J is the jitter spectrum,
f is frequency, and
σ is the rms jitter value.
If the effective jitter step response differentially impacts the
duration of the two or more waveform transitions used to
compute t
d
, then jitter must be reduced. More than likely, jit-
ter will be nearly identically distributed for each transition. But
if the jitter is so great as to affect the accuracy of computing
the transition instants, then the user must reduce the duration
of the waveform period or reduce the system jitter. Reduction
in the duration of the waveform period may introduce a bias in
the voltage values and this may affect the computed value of
t
d
. If the rms jitter value is less than 20% of the transition
duration of the TDR step response, then the jitter is small and
can be ignored. For typical TDR systems, however, rms jitter
is less than 10 ps and will not affect the t
d
measurements.
Similarly, the effect of cables, connectors, and adapters on
the measurement can be modeled by convolving their step
responses with that of the TDR unit. If the transition duration
of this new step response meets the requirements of 4.1.2,
then the performance of the cables, connectors, and adapters
is adequate.
Number
2.5.5.11
Subject
Propagation Delay of Lines on Printed Boards by TDR
Date
04/2009
Revision
IPC-TM-650
Page
15
of
16
6.4.2 Waveform Averaging and Number of Samples in
the Measurement Zone
Waveform averaging reduces the
effective noise level of the measurement by M
-1/2
, where M is
the number of acquired waveforms (typically, 8 ≤ M ≤ 256).
Consequently, averaging can reduce measurement noise.
This reduction is limited by the number of bits of the analog-
to-digital converter of the TDR system and the linearity of the
timebase. However, if the TDR system exhibits drift in the
timebase, averaging too many waveforms may result in a
reduction of t
sys
and a commensurate reduction in the
temporal/spatial resolution of the TDR.
The number of samples (data points) in the waveform epoch
will affect the accuracy and uncertainty of the computed value
of t
d
because this value is typically computed by interpolating
between adjacent datum values. Therefore, the more samples
in the waveform epoch, the smaller will be the error and stan-
dard deviation of the computed t
d
value.
6.4.3 Selection of Waveform Period
Inconsistency in
defining the waveform epoch may cause a significant but
unknown error than can exceed two sample intervals. Speci-
fying the waveform period to be consistent for both long and
short line measurements improves t
d
repeatability, and this
can improve assessment of design and fabrication quality and
vendor capability. This waveform epoch should be long
enough to accurately determine if the waveform has settled on
both sides of the waveform transition but should be short
enough, given the number of samples in the waveform, to
accurately compute the transition instant. The waveform
epoch is defined in 5.1.3.
Number
2.5.5.11
Subject
Propagation Delay of Lines on Printed Boards by TDR
Date
04/2009
Revision
IPC-TM-650
Page
16
of
16