![[US Patent & Trademark Office, Patent Full Text and Image Database]](/netaicon/PTO/patfthdr.gif)
| United States Patent |
6,587,530
|
|
Conklin
, et al.
|
July 1, 2003
|
Method and apparatus for signal integrity verification
Abstract
The present invention provides a signal integrity measurement method and
apparatus which allows for signal characteristics to be measured by
obtaining samples taken at the midpoint of the data stream. The invention
provides a measurement device that is suitable for use in the field to
provide a measurement of signal characteristics within transmitted data
streams. The invention is particularly suitable for field measurement of
signal characteristics of data streams or continuous in-line monitoring of
signal characteristics within transmitted data streams. The signal
characteristics include, but are not limited to eye opening jitter, noise,
slope efficiency, average power and peak-to-peak amplitude.
| Inventors:
|
Conklin; Troy R. (Hillsboro, OR);
Hutchison, Jr.; Harold B. (Forest Grove, OR)
|
| Assignee:
|
International Business Machines Corporation (Armonk, NY)
|
| Appl. No.:
|
680819 |
| Filed:
|
October 5, 2000 |
| Current U.S. Class: |
375/372; 359/264 |
| Intern'l Class: |
H04L 025/00; G06F 001/07 |
| Field of Search: |
375/372,224,225,316,324,327,355,371
327/165,291,26,37
359/118,264
|
References Cited [Referenced By]
U.S. Patent Documents
Other References
Wavecrest Corporation, Technical Papers and Presentations,
www.wavecrest.com/technical/papers.html.
Wilstrup, Jan, A Method of Serial Data Jitter Analysis Using One-Shot Time
Interval Measurements, T11/98-130v0,
www.wavecrest.com/technical/papers.html.
Wilstrup, Jan, A Method of Serial Data Jitter Analysis Using One-Shot Time
Interval Measurements, Feb. 5, 1998,
www.wavecrest.com/technical/papers.html.
|
Primary Examiner: Tran; Khai
Attorney, Agent or Firm: Garnett; Pryor A, Paxson LLP; Dilworth
Claims
We claim:
1. A method for measuring the signal characteristics of a
telecommunications data stream, the method comprising the steps of:
obtaining at least one sample measurement at the midpoint of a waveform of
the data stream; and
determining at least one signal characteristic of the data stream based
upon the sample measurement,
wherein the signal characteristic is selected from the group consisting of:
eye opening; jitter; noise; average power; peak-to-peak amplitude; and
slope efficiency.
2. The method of claim 1, further comprising the step of:
producing a first indication if the signal characteristic is within a first
range.
3. The method of claim 2, wherein the step of producing the first
indication comprises presenting a visual or audible indication.
4. The method of claim 2, wherein the step of producing the first
indication comprises transmitting a data signal.
5. The method of claim 2, wherein the first range corresponds to a
preferred range.
6. The method of claim 2, further comprising the steps of:
producing a second indication if the signal characteristic is within a
second range; and
producing a third indication if the signal characteristic is within a third
range.
7. The method of claim 6, wherein
the first range corresponds to a preferred range;
the second range corresponds to a marginal range; and
the third range corresponds to a failure range.
8. The method of claim 1, wherein the signal characteristic is expressed as
a percentage.
9. The method of claim 1, wherein the signal characteristic is numerically
displayed.
10. An apparatus for a telecommunications data stream, the apparatus
comprising:
means for obtaining at least one sample measurement at the midpoint of a
waveform of the data stream; and
means for determining at least one signal characteristic of the data stream
based upon the sample measurement,
wherein the signal characteristic is selected from the group consisting of:
eye opening; jitter; noise; average power; peak-to-peak amplitude; and
slope efficiency.
11. The apparatus of claim 10, further comprising:
means for producing a first indication if the signal characteristic is
within a first range; and
means for producing a second indication if the signal characteristic is
within a second range.
12. The apparatus of claim 11, further comprising:
means for producing a third indication if the signal characteristic is
within a third range.
13. The apparatus of claim 12, wherein
the first range corresponds to a preferred range;
the second range corresponds to a marginal range; and
the third range corresponds to a failure range.
14. An apparatus for a telecommunications data stream, the apparatus
comprising:
an analog to digital converter;
a clock recovery unit;
a midpoint sampler responsive to the analog-to-digital converter and clock
recovery unit; and
a signal characteristic calculator responsive to the midpoint sampler.
15. The apparatus of claim 14, wherein the signal characteristic is
selected from the group consisting of:
eye opening;
jitter;
noise;
average power;
peak-to-peak amplitude; and
slope efficiency.
16. The apparatus of claim 14, further comprising:
a first indicator responsive to the signal characteristic being within a
first range.
17. The apparatus of claim 16, wherein the first indication comprises
visual or audible indication.
18. The apparatus of claim 16, wherein the first indicator comprises a data
signal.
19. The apparatus of claim 16, further comprising:
a second indicator responsive to the signal characteristic being within a
second range; and
a third indicator responsive to the signal characteristic being within a
third range.
20. The apparatus of claim 19, wherein
the first range corresponds to a preferred range;
the second range corresponds to a marginal range; and
the third range corresponds to a failure range.
21. The apparatus of claim 14, wherein the apparatus is contained in a
receiver.
22. The apparatus of claim 14, wherein the apparatus is distributed between
a receiver and a host system.
23. The apparatus of claim 14, wherein the signal characteristic is
numerically displayed.
24. The apparatus of claim 14, wherein the clock recovery unit is selected
from a group comprising:
a phased locked loop circuit; and
a serializer/deserializer circuit.
Description
BACKGROUND OF THE INVENTION
1. Technical Field
The present invention relates to an apparatus and system that measures and
outputs a first order assessment of signal characteristics and a
determination of signal quality and integrity within telecommunications
networks such as gigabit and multi-gigabit networks.
2. Description of the Prior Art
Gigabit and multi-gigabit Local Area Network (LAN) and Storage Area Network
(SAN) markets are rapidly emerging industries, which are entirely
dependent on inexpensive network media components, including both copper
and fiber optics technologies. Since the late 1980's, the available
high-speed fiber optic components have been inexpensive screened parts
taken from compact disk (CD) production lines. CD laser technology is
based on edge emitting Fabry-Perot laser diodes, which offer data
communications bandwidth from hundreds of megabit/sec to gigabit/sec.
Other low-cost technologies, such as super-luminescent
light-emitting-diode (LED), have been incapable of achieving gigabit link
speeds. Since LAN and SAN markets demand low cost, the technically more
mature edge emitting Fabry-Perot and distributed Bragg reflector laser
diodes that are used in telecommunications applications were not adopted.
The CD laser technology was embraced by a number of suppliers, who applied
their respective level of sub-component supplier selection and parts
screening processes. Some CD laser vendors did an excellent job in
selecting sub-component suppliers, while others applied various levels of
parts screening. In the end, low cost CD lasers have a fundamental
characteristic: gradual reduction in "relaxation oscillation frequency."
Over time, these components experience un-damped ringing, resulting in
drastically increased, measurable jitter on the link (as measurable by a
digital-sampling oscilloscope). Some CD laser vendors have had to undergo
full-scale replacement of their product in the field, while others have
stated that it is just a matter of time before their product suffers from
the same degradation. This represents several years of production from
multiple suppliers, all of which requires support in the field.
Laser media suppliers are transitioning to Vertical Cavity Surface Emitting
Laser (VCSEL) diode based technology, which is specifically designed and
manufactured for the low cost, LAN and SAN data communications markets.
This will allow much better resolution of issues, since vendors are no
longer leveraging a fundamentally different, high volume, commodity
industry. VCSEL technology has the potential to offer data communications
bandwidth from 1 to 5+, gigabit/sec, and offers the initial promise to
attain very high reliability figures similar to the early days of the
telecommunications industry. On the cautionary side, VCSEL technology is
still relatively new, and primary failure modes are still being
characterized. Additionally, VCSEL manufacturers are transitioning from
ion-implant confined device structures to oxide confined device
structures. These device enhancements, coupled with overall
manufacturing/materials cost cutting efforts and higher data rate
applications operating at 2 gigabit/sec, represent areas of risk which may
impact life-time characteristics compared to current practice and
theoretical understanding.
However, beyond laser physics, there are a whole host of other materials,
design, and manufacturing issues which have the potential to adversely
impact signal integrity. Field service groups have already experienced
instances of cracked lenses within the laser assembly, glue that fails
over time, glue that has not been fully cured, as well as spattered glue
on the lens, all of which which darken the transmitted signal at some time
after installation.
The primary failure mode requiring new field instrumentation is degradation
of the transmitting device, including the laser diode and associated drive
electronics. This failure mode does not change average or peak signal
power levels, but rather results in increased "jitter" within the digital
signal. "Jitter" is an industry term which refers to the amount of
variance within the rising and falling edges of the digital signal. As
seen in FIG. 2, Jitter 23 is measured at the receiver, and appears on a
digital-sampling oscilloscope as though the rising and falling edges of
the signal are smeared across a broad area of the overall duty cycle. This
smearing results in a more closed eye opening 21 and may be quantified by
expressing the eye as the "percent eye closing" or "%eye." Noise 24 is the
signal unrelated to the signal of interest.
At the early stage of transmitter failure, the link may experience
occasional bit errors, which may only happen with specific data patterns
within an IO data-gram sent over the link, or with heavy IO load levels
transmitted over the link. At the later stages of transmitter failure, the
link saturates with errors, and becomes non-operational for transferring
IO. For both of these increased transmitted "jitter" cases, the optical
power level is within the original manufacturer specification, and field
service is not able to use an optical power meter to distinguish an error
free link from a link that is unusable. At the present time, there is no
direct measurement technique suitable for use in the field which provides
a measurement of signal integrity or transmitted "jitter" within gigabit
and multi-gigabit networks.
Without this capability, field service is placed in the difficult position
of trouble-shooting the onset of intermittent errors using only customer
data transmitted through the network and analysis of error counters
throughout the system. Trouble-shooting intermittent errors within an
on-line system is very difficult, due to the random nature of hitting the
right data pattern or load level which causes the error, as well as
collecting and analyzing the correct set of error counter data, which
identifies the most likely failing component or components. This
error-counter analysis procedure is extremely time consuming and results
in the identification of a small number of "most likely" failing
component(s), rather than accurately identifying "the" one or more failing
components. Additionally, this analysis depends heavily on the judgement
of the field service person, who must make the decision whether to
expedite a field service action based on a few increments in error
counters or wait to see what happens next.
Once field service has the opportunity to take a portion of the system down
for maintenance, they have a short period of time to complete any further
trouble shooting to narrow the list of most likely failing components,
replace the fewest number of components, and return the system to
operation. It is at this point in time, that field service requires a
low-cost, rapid, and deterministic pass-fail test to isolate the failing
component. Without this capability, field service has no way to verify
quality of the components being placed back into operation, and the system
is exposed to the risk of oscillating through many cycles of maintenance
actions. In the case of intermittent errors, this is especially
unacceptable since it may take a long period of time to confirm whether
the problem was fixed, based on the likelihood hitting the data pattern or
load level, which triggers the error.
Automated and continuous collection of on-line diagnostic data has, to some
extent, helped to reduce trouble-shooting time and improve accuracy.
However, even with improvements in on-line diagnostics data collection,
there is a need in the industry for a means to directly measure signal
integrity in the field. Whether in the form of a portable, hand held field
unit or incorporated into the communications equipment itself, the ability
to rapidly evaluate the optic signal at the customer site is a key quality
assurance factor.
All of these criteria point to the need for a means to directly measure
signal integrity both in-line and in the field. Additionally, many of the
risk factors apply equally to copper-based gigabit media.
SUMMARY OF THE INVENTION
The present invention provides a signal integrity measurement method and
apparatus, wherein the signal characteristics of a telecommunications data
stream are measured based upon obtaining at least one sample measurement
at the midpoint of the data stream. The signal characteristic determined
by this method is preferably one or more of: percentage of eye (%eye)
opening of the data stream, jitter, noise, slope efficiency, average power
and peak-to-peak amplitude. This information is preferably also stored,
and as necessary, displayed. In addition the invention provides a method
whereby a first indication is displayed if the signal characteristic is
within a first range; a second indication is displayed if the signal
characteristic of the data stream is within a second range; and a third
indication is displayed if the signal characteristic of the data stream is
within a third range.
Another aspect of the invention is an apparatus and a system related to the
aforementioned method.
Yet another aspect of the invention is an apparatus such as a hand held
device, a transceiver and host system, or a transceiver alone, comprising
an analog-to-digital converter, a clock recovery unit, a midpoint sampler,
and a signal characteristic converter.
The method and apparatus of the present invention provide a measurement
device that is suitable for use in the field to provide a measurement of
signal characteristics within transmitted data streams. The method, and
apparatus system are also suitable for field measurement of signal
characteristics of data streams or continuous in-line monitoring of signal
characteristics within transmitted data streams. The signal characteristic
preferably includes, but is not limited to eye opening, jitter, noise,
slope efficiency, average power and peak-to-peak amplitude.
Other features and advantages of this invention will become apparent from
the following detailed description of the presently preferred embodiment
of the invention, taken in conjunction with accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a flowchart of the invention according to the preferred
embodiment of this invention, and is suggested for printing on the first
page of the issued patent.
FIG. 2 is a diagram of an unfiltered eye.
FIG. 3 is a picture of an analog eye opening.
FIG. 4 is a graph showing the basis for a geometerical estimate of the eye
opening of FIG. 3.
FIG. 5 is a block diagram to show a generalization of the circuitry that
could be used to implement the present invention.
FIGS. 6-8 are waveforms taken by an oscilloscope used to illustrate the
present invention.
FIG. 9 is a block diagram of the hand-held implementation of the present
invention.
FIGS. 10-11 are graphs of slope efficiency.
FIG. 12 is a graph showing the basis for a calculus estimate of the eye
opening of FIG. 3.
DESCRIPTION OF THE PREFERRED EMBODIMENT
Technical Background
The present invention relates to an apparatus and system that measures and
outputs a first order assessment of signal integrity within
telecommunications networks such as, but not limited to, Fibre Channel and
Gigabit Ethernet gigabit and multigigabit networks. Fibre Channel is the
general name of an integrated set of standards developed by the American
National Standards Institute (ANSI). A novel aspect of the present
invention is the ability to troubleshoot significant degradation of signal
quality by using a handheld device or an in-line continuous monitor of
signal quality.
FIG. 1 is a flowchart of the presently preferred embodiment of the
invention. The contiguous samples are stored, step 1, into registers. The
next step is to initialize the counter, step 2, which indicates the number
of samples that are to be taken. The more samples that are taken, the
longer it takes between readouts but the accuracy of the obtained readings
is increased. A sample of the waveform is taken at the midpoint, step 3.
The sample (S) is compared to the stored Y4, Y3, Y2 and Y1 values, steps
4-8. If S is greater than the Y or M values, the value of Y is replaced
with the S value, steps 9-12. If S is less than all of the Y values, step
8, S replaces Y1, step 17. If S is greater than Y1 however, the sample is
discarded. This process continues until a predetermined number of samples
are acquired, at which time the counter is equal to 0, step 15. If the
counter is not equal to 0 the counter is decremented, step 14 and another
sample is obtained, step 3 repeated. The signal characteristics, such as
%eye, noise and jitter, are calculated, step 16, from the stored values
after the predetermined number of samples are acquired.
These values can be displayed in a number of different manners, such as in
the form of actual numeric values, or as a histogram, or it can be
determined if the values are within either of the acceptable, cautionary
or critical ranges.
Overview
Fiber optic link health assurance primarily relies on the measurement of
the optic signal strength. This is critical for the obvious reason that
sufficient power must be present to allow the receiver to properly
interpret the signal. However, the presence of a strong signal does not
guarantee that the signal has good integrity or quality. One universally
applied tool for the measurement of signal quality has been the eye
diagram. For example, the Fibre Channel Physical Layer specification calls
for a minimum of 57% eye opening for 1063 Mb/s data rates using short wave
laser transmitters. Unfortunately, eye diagrams require that relatively
expensive high-speed oscilloscopes be available for field applications,
which is impractical.
There is a strong relationship between the signal noise level and the
percentage eye opening. This relationship can be utilized to give a quick
measure of the eye quality, and hence the health of the optic link. FIG. 2
shows a stylization of an unfiltered eye as might be shown on a high-speed
oscilloscope, as well as noise 24, bit period 22 and jitter 23. The image
is acquired by overlaying successive traces one upon another, while the
scope is in a "persistence" mode. The trace is all positive; that is, the
entire signal is displayed above the ground reference (GND) 25, since
there is no "negative light." Also, the oscillation in the logic one level
would normally not be visible when viewed with the bandwidth filters on
due to the filter cutoff frequency. The transmitted square wave normally
appears as a sine wave.
The bit period is the time between two adjacent pulses and its base unit is
the second. The bit period depends on the data rate and can be extracted
from a Fibre Channel data stream via a Phase Locked Loop (PLL). Phase
Locked Loop is a circuit in which a phase comparator will compare the
phase of an input signal and the voltage controlled oscillator (VCO) in
the loop, and is readily available at gigabit speeds. The output of the
phase comparator is filtered and the resulting voltage controls the VCOs
frequency. This causes the loop to stabilize in such a state that the
phase of the input signal and the VCO are equal.
The noise measurement requires a combination of a sample-and-hold circuit,
and a flash analog-to-digital converter (ADC) capable of sampling at the
data rate of the input signal. This is a common feature of today's
high-speed digital scopes.
An important aspect of the present invention is that the critical
information that was formerly obtainable only by use of a high-speed
oscilloscope, can now be obtained by acquiring samples at the waveform's
midpoint. As a result, circuit complexity is considerably reduced which
facilitates the invention's use as a hand held device or an in-line
monitor for signal quality/integrity.
An aspect of the present invention is the relationship noise has to the
total area enclosed in an eye envelope. FIG. 4 shows an approximation of
the analog eye of FIG. 3, from which the relationship can be derived. The
initial assumption is that the filtered waveform is essentially
symmetrical, and that it is fundamentally a sine wave. Note that as eye
quality degrades, it becomes less symmetrical, which affects the
derivations, but the invention is still effective for first order
approximations.
As an assumption for eyes of good quality, the rising edge slope is equal
to that of the falling edge, and therefore calculations of the percentage
eye opening and noise measurements can be made by obtaining the abscissa
of the polygons (or the upper and lower extremes) over time. Eyes that
have poor quality will generally have fall times that are greater than
rise times, but this method is still adequate for a first order
approximation. There are several different techniques available to make
approximations or first order measurements for signal quality. The
measurements of the percentage of eye opening (%eye), jitter and noise can
be determined geometrically using algebra, or by the use of calculus.
Eye Opening Approximation
With reference to FIG. 4, the percent eye, noise and jitter are calculated
as follows using one of two derivations: a Geometric derivation or a
Calculus derivation.
Method 1 Geometric Derivation:
x=(a to c)=(c to e)=1/2 bit period;
.DELTA.x=(a to b)=jitter;
.DELTA.x/2=(c to d);
x'=(d to e)=(c to e)-(c to d)=x-.DELTA.x/2;
y=Y4-M;
.DELTA.y=Y4-Y3=Y2-Y1=noise;
M=(Y4+Y1)/2=(Y3+Y2)/2=midpoint of the envelope=average power;
Amp=Y4-Y1=peak-to-peak amplitude.
In light of the above, the percentage of eye opening 30 (%eye) can be
expressed as follows:
%eye=(Eye opening/Eye envelope)*100=A'/A*100.
But: A'=4[(x-.DELTA.x/2)(y-.DELTA.y/2)/2]=2(x-.DELTA.x/2)(y-.DELTA.y/2)
And: A=4[x*y/2]=2xy
So:
##EQU1##
Therefore,
%eye=[1-1/2(.DELTA.x/x)-1/2(.DELTA.y/y)+1/4(.DELTA.x/x)(.DELTA.y/y)]*100.
(Formula 1)
Note that Ay is equal to the noise present in the signal at the sample
point, and can be calculated from the logic 1 level by Y4-Y3, or from the
logic 0 level by Y2-Y1. Also note that x is known; specifically for Fibre
Channel 1 Gbaud traffic, x is equal to 470.5 ps; for 2 Gbaud FC traffic, x
is equal to 235.25 ps, etc. The remaining independent variable, .DELTA.x
(jitter), can be determined by the principle of similar triangles,
assuming that the slope of the rising edge is constant.
Slope=(.DELTA.y/2)/(.DELTA.x/2)=y/x=(Y4-M)/470.5 ps
.DELTA.y/.DELTA.x=(Y4-M)/470.5 ps,
So: .DELTA.x=(470.5 ps)(.DELTA.y)/(Y4-M)
=(470.5 ps)(.DELTA.y)/(Y4-((Y4+Y1)/2)) # substituted M
=(470.5 ps)(Y4-Y3)/(Y4/2-Y1/2) # substituted .DELTA.y; simplified
denominator.
=(470.5 ps)(Y2-Y1)/(Y4/2-Y1/2) # equivalent statement, using logic 0 noise.
Jitter can now be approximated in terms of independent variables, which are
determinable at the midpoint of the acquired waveform.
.DELTA.x/x=((470.5 ps)(Y4-Y3)/(Y4/2-Y1/2))/470.5 ps
=(Y4-Y3)/(Y4/2-Y1/2)
=2(Y4-Y3)/(Y4-Y1)
=2(Y2-Y1)/(Y4-Y1)#[since (Y4-Y3)=(Y2-Y1)]
.DELTA.y/y=(Y4-Y3)/(Y4/2-Y1/2)
=2(Y4-Y3)/(Y4-Y1)
=2(Y2-Y1)/(Y4-Y1)#[since (Y4-Y3)=(Y2-Y1)]
This shows that the ratio of noise-to-signal is equal to that of
jitter-to-bit period. Further, since from Formula 1:
%eye=[1-1/2(.DELTA.x/x)-1/2(.DELTA.y/y)+1/4(.DELTA.x/x)(.DELTA.y/y)]*100
which can be expressed as a function of "logic one" noise,
%eye=[1-2(Y4-Y3)/(Y4-Y1)+1/4(.DELTA.x/x)(.DELTA.y/y)]*100
%eye=[1-2(Y4-Y3)/(Y4-Y1)+(Y4-Y3).sup.2 /(Y4-Y1).sup.2 ]*100
or as a function of "logic zero" noise,
%eye=[1-2(Y2-Y1)/(Y4-Y1)+(Y2-Y1).sup.2 /(Y4-Y1).sup.2 ]*100
In effect, this is simply: %eye=[1-2*(noise/Amp)+(noise/Amp).sup.2 ]*100
Samples obtained at the midpoint of the bit period yielding Y4, Y3, Y2 and
Y1 can be used to calculate:
Noise: the greater of (Y4-Y3) and (Y2-Y1)
Jitter: (470.5 ps)(Y4-Y3) (Y4/2-Y1/2) or equivalently
(470.5 ps)(Y2-Y1)/(Y4/2-Y1/2)
Method 2 Calculus Derivation:
Referring now to FIG. 12 to estimate the eye opening using calculus, one
must find the area of the eye 30' as a percentage of the area between a
sine and negative sine curve for one-half cycle. .theta.j 85 is the jitter
phase. The derivation is as follows:
The first step is to determine .theta.j 85 as a function of Y4 and Y3.
The difference between the sine waves at the clock midpoint is given by:
##EQU2##
Dividing through by Y4 and evaluating the first sin( ) yields:
##EQU3##
The second step is to determine %eye as a function of .theta.j.
First derive the maximal eye area:
##EQU4##
Second, shift the limits of integration by .theta.j 85, and derive the
jittered eye area:
##EQU5##
The third, and final step, is to find %eye as a function of Y4 and Y3.
##EQU6##
Substituting the results from the first derivation:
##EQU7##
Applying half-angle formulas for cos( ) and sin( )
##EQU8##
Simplifying the expression under the square root:
##EQU9##
The values for jitter, noise and %eye can be obtained by the geometric
derivation, the calculus derivation or by any other suitable means for
determing such signal characteristics from the sample measurements at the
midpoint of the data stream.
Hardware Implementation
FIG. 5 is a block diagram that is intended to show a generalization of the
existing information flow for an optical to electrical conversion stream,
and an alternative method that could incorporate the present invention for
capturing optics quality data. The new host system 57 is intended to
represent a host-based implementation, useful for continuous in-line
signal monitoring. The cross connections represented by the dotted lines
are included to indicate that the old GBIC 51 would be accepted by the new
host system 57, and that the new GBIC 73 would be accepted by the old host
system 56.
Currently the optic stream, in the form of inbound light, is channeled to a
photodetector 52 which does the actual conversion from light to current.
The resultant electrical representation has the very high frequency
content filtered by filter 53 before being converted to a digitized
differential signal by differential converter 54 and digitizer 55. This
differential signal (+/-Rx) is presented to the output pins of the GBIC
(or MIA, 1.times.9 etc.) which are the input pins to the Host 56. The
inbound Rx signal pair is presented to the serializer/deserializer
(SERDES) 58 which produces 10 or 20 bits of paralleled data. This wide
data is fed to the protocol controller 59, which converts the 10b data to
8b data, and presents it to the system for routing.
One aspect of the current invention seeks to alter this flow. The GBIC 73
presents to new host system 57 proportional (analog) differential signals
61 rather than digitized differential pairs. The SERDES 65 processes this
data as above, as long as the signal levels are scaled from 370 to 2000
mV. Additional circuitry 62, then converts the analog differential signal
61 to single-ended data, which can be read by a flash analog-to-digital
converter 63 (ADC) and latched by a register 64. If the full-speed clock
signal (CLK) 68 is not available external to the SERDES (the extracted
clock is usually presented a factor 10.times. or 20.times. lower than the
link rate), then an additional phased locked loop (PLL) block 67 clocks
data into the register 64. The SERDES 65 and PLL block 67 alternatively
perform the function of a clock-recovery-unit (CRU). The CRU function can
be performed by these and any other suitable means.
The register 64 sits on the host local bus, and can be probed and read by
the host firmware as required. The width of the register depends on the
width of the local bus and the resolution of the flash ADC (the number of
bits). High speed FIFOs are used initially to clock in the first six,
contiguous samples to ensure that both a high signal level and low signal
level are captured. After those samples are read, following samples can be
overwritten without real data loss, since the method seeks only to capture
the extremes of both the logic one and logic zero levels. Since the data
rate is over 1 GHz, an adequate sample size is obtained in only a few
seconds, even if just a small percentage of signals are being recorded by
the firmware.
FIG. 9 is a block diagram of a stand-alone, hand-held unit that
incorporates much of the same structure, but without the SERDES and
without the conversion of the single-ended output from the filter to
analog differential. Photo-detector 74 converts the incoming light to an
electrical current. The modulated signal is applied to the PLL 75, which
recovers the clock. This is used to gate the S&H 76/Flash circuitry 79,
which is registered and read by the midpoint sampling and signal
characteristic calculation logic 78. Since there is a direct relationship
between power and current, the average power received is also discernable
by integrating the signal over time. The integrator function 77 operates
similarly, with the period of integration a multiple of the clock period.
The oscillator 82 feeding the PLL 75 can be made switchable to provide the
ability to sync up to a number of different fiber optic protocols or
speeds. The filter 81 shown in FIG. 9 could be configurable for various
optic protocols (e.g., Fibre Channel 1 Gbit traffic requires a 797 MHz
Bessel Thompson filter).
Referring again to FIG. 1, there are a number of logic schemes that can be
used to determine the abscissa values, but provided below is a simple
condition tree that yields the Y4, Y3, Y2, and Y1 values. This condition
tree is usable in either the on-line monitor or as a hand-held unit. In
the case of Fibre Channel, the sample and hold function must acquire six
contiguous samples at full fiber speed to ensure that valid starting
values are acquired prior to entering the body of this routine. Other
protocols that are not based on 8b/10b code may require different sync up
periods. Similarly, the device logic 78 from FIG. 9 can be in the host
system or on the receiver, and can be in the form of software, hardware,
firmware, an application specific integrated circuit (ASIC), or a state
machine.
Step 1 Load initial Y4 and Y1 values,
Acquire 6 contiguous samples, S1, S2, S3, S4, S5, S6
Set Y4=S1
Set Y1=S1
For i in 2 3 4 5 6
do
If sample S(i)>Y4 then
Y4-S(i)
Else if sample S(i)<Y1 then
Y1=S(i)
fi
done
Set Y3=Y4
Set Y2=Y1
Step 2 Initialize count C.
Set C=(number of samples to take)
Step 15 If Count>0, then Decrement count Step 14.
Step 16 Else calculate signal characteristics.
Step 3 Acquire sample (S).
Step 4 If S>Y4, then Y4=S Step 9 and test count Step 15.
Step 5 Else if S>Y3, then Discard Step 10 and test count Step 15.
Step 6 Else if
##EQU10##
then Y3 S Step 11 and test count Step 15.
Step 7 Else if S>Y2, then Y2=S Step 12 and test count Step 15.
Step 8 Else if S>Y1, then Discard S Step 13 and test count Step 15.
Step 17 Else Y1=S; test count Step 15.
The instrument runs until a sufficient number of samples are acquired,
after which, the calculations for eye opening, noise, and jitter are
generated and displayed on a digital readout (such as, but not limited to,
a digital multimeter).
FIG. 10 is a light output power vs. applied current (diode current) graph
for a laser diode. The bias current is applied in response to the level of
average optical output power received in a feedback loop. The modulation
amplitude is kept constant. The LI curve 84 reflects the quantum well
efficiency of the diode. FIG. 11 depicts the characteristics of a diode
that has degraded. The efficiency has reduced, which results in an
increase in the bias current (due to the feedback loop). Modulation
current is kept constant, but due to the lower slope of the LI curve 84'
the optical output peak-to-peak swing has decreased. This lowered slope of
LI curve 84' reflects the degraded performance of the diode. The present
invention can be used to track the slope efficiency of a transmitter
device such as a diode by using the Y4 and Y1 values. The difference
(Y4-Y1) is the optical peak-to-peak value, which once collected can be
monitored over time for signs of degradation.
Other measurements that are obtained by using the present invention include
average power, peak power and peak-to-peak amplitude.
EXAMPLES
The following examples are included to further illustrate the present
invention. In these examples, the geometric derivation is used.
FIG. 6 is a trace of a marginal telecommunications device with 152 ps of
jitter (as measured with an oscilloscope). Using the geometric derivation:
Jitter (calculated at high point)=470.5 ps*(6.9/17.05)=190.4 ps
Jitter (calculated at low point)=470.5 ps*(7.2/17.05)=198.7 ps
%eye(calculated at high point)=[1-2*(6.9/34.1)+(6.9) 2/(34.1) 2]*100=63.6%
%eye(calculated at low point)=[1-2*(7.2/34.1)+(7.2) 2/(34.1) 2]*100=62.2%
In this example, the jitter value (worse case) given with the present
invention is 198.7 ps. The %eye (worse case) calculates to be 62.2%, which
is close to the Fibre Channel minimum of 57%. This telecommunications
device would rate within the marginal range.
FIG. 7 is a trace of a known failing telecommunications device that
measured 340 ps of jitter. Using the geometric derivation:
Jitter (calculated at high point)=470.5 ps*(8.4/19.2)=205.8 ps
Jitter (calculated at low point)=470.5 ps*(13.8/19.2)=338.2 ps
%eye(calculated at high point)=[1-2*(8.4/38.4)+(8.4) 2/(38.4) 2]*100=61.0%
%eye(calculated at low point)=[1-2*(13.8/38.4)+(13.8) 2/(38.4) 2]*100=41.0%
In this example, the calculated eye opening (worse case) is 41.0%, which
places the device in the failing range of eye opening.
FIG. 8 is a trace of a known good telecommunications device that measured
76 ps of jitter. Using the geometric derivation:
Jitter (calculated at high point)=470.5 ps*(8.4/25.3)=156.2 ps
Jitter (calculated at low point)=470.5 ps*(8.6/25.3)=159.9 ps
%eye (calculated at high point)=[1-2*(8.4/50.6)+(8.4) 2/(50.6) 2]*100=69.6%
%eye (calculated at low point)=[1-2*(8.6/50.6)+(8.6) 2/(50.6) 2]*100=68.9%
The high jitter value given by the present invention (159.9 ps) is due to
the high noise level (8.6 mV). The %eye (worse case) calculates to be
68.9%, which is within the acceptable range of eye opening.
It will be appreciated that, although specific embodiments of the invention
have been described herein for purposes of illustration, various
modifications may be made without departing from the spirit and scope of
the invention. In particular, the invention can be used as either a
hand-held field device or as an in-line monitor. The display of signal
characteristics can be either discrete or continuous, including numerical
display and can be either or both visual and audible, or by a data signal
transmitted to another electronic device. The signal characteristics can
further be expressed in the form of ranges that indicate that the signal
characteristic is within either a preferred range, a marginal range or a
failure range. Further, the in-line monitor can be a combination of the
host and a receiver or in the form of a transceiver alone. Accordingly,
the scope of protection of this invention is limited only by the following
claims and their equivalents.
* * * * *
![[Image]](/netaicon/PTO/image.gif)
![[Home]](/netaicon/PTO/home.gif)
![[Boolean Search]](/netaicon/PTO/boolean.gif)
![[Manual Search]](/netaicon/PTO/manual.gif)
![[Number Search]](/netaicon/PTO/number.gif)
![[Help]](/netaicon/PTO/help.gif)
![[Bottom]](/netaicon/PTO/bottom.gif)
![[View Shopping Cart]](/netaicon/PTO/cart.gif)
![[Add to Shopping Cart]](/netaicon/PTO/order.gif)
![[Top]](/netaicon/PTO/top.gif)