timothy r. minnick - amp technical seminar · pdf filetechnical seminar h412 nanosecond...

1

ProjectBackground

DiscontinuityDiscovery

PhysicalTesting

SimulationTesting

Solution

MaterialQualification

Conclusion

Technical Seminar H412

Nanosecond DiscontinuityNanosecond DiscontinuityImpact on Hot-SwapImpact on Hot-Swap

Hank Hank Herrmann Herrmann - AMP- AMPJack Kelly - MotorolaJack Kelly - Motorola

Timothy R. Minnick - AMPTimothy R. Minnick - AMP

2

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Nanosecond Discontinuities...

...are ...are NOTNOTpin bounce!!!pin bounce!!!

3

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Session Outline•• Project BackgroundProject Background•• Discovery of the PhenomenonDiscovery of the Phenomenon

–– Physical TestingPhysical Testing

–– Simulation TestingSimulation Testing

•• SolutionSolution•• Material QualificationMaterial Qualification•• ConclusionConclusion

4

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

The Basic System

•• Intense Error-Free Data Rates(telecom)Intense Error-Free Data Rates(telecom)•• Bus ArchitecturesBus Architectures•• Conductive Interfaces (i.e. connectors)Conductive Interfaces (i.e. connectors)•• Hot-Swap CapabilityHot-Swap Capability

Slot 2Slot 1 Slot X

ProjectBackground

5

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

The Hot-Swap System

•• Additional Energy TransferAdditional Energy Transfer•• Additional Signal Integrity RequirementsAdditional Signal Integrity Requirements

Slot 2Slot 1 Slot X

Hot-Swap CardProjectBackground

6

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Motorola CPX8000 System

•• Telecom Infrastructure ApplicationsTelecom Infrastructure Applications

•• CompactPCI ArchitectureCompactPCI Architecture

•• High AvailabilityHigh Availability

•• Data IntensiveData Intensive

•• Device RedundantDevice Redundant

ProjectBackground

7

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Detection of a Discontinuity


8

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Mechanical Connector Bounce

•• Analyzed at birth of CPCI specificationAnalyzed at birth of CPCI specification

•• Typical speeds in microsecond rangeTypical speeds in microsecond range

•• Pre-charge resistorsPre-charge resistors

Present signal integrity design can not Present signal integrity design can notrespond to nanosecond discontinuities!!respond to nanosecond discontinuities!!


9

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Magnified 2mm HM Contact


10

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Initial Discontinuity Test Set-Up

PhysicalTesting

Hot-Swap Card

5.0 V

GND

4.7k-Ohm

4.7k-Ohm

Backplane

Scope Probe(1st Channel)

Scope Probe(2nd Channel)

Zoomed-inView (~40 usec)

Engagement

11

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Electrical Testing

•• Simple voltage dividerSimple voltage divider•• Constant velocity fixtureConstant velocity fixture

Vmeas

PhysicalTesting Oscilloscope

1K

StandardReceptacle

Standard Pin

+5v

12

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Equipment References

•• Discontinuity DetectionDiscontinuity Detection–– Tektronix TDS 784A Digitizing OscilloscopeTektronix TDS 784A Digitizing Oscilloscope–– Motorola Motorola VxWorks VxWorks Operating SystemOperating System–– Motorola CPX8000 Series Chassis & CardsMotorola CPX8000 Series Chassis & Cards

•• Pin Testing and Resistance ProfilingPin Testing and Resistance Profiling–– Tektronix TDS 684A Digitizing OscilloscopeTektronix TDS 684A Digitizing Oscilloscope–– Mating Fixture with Pneumatic DriveMating Fixture with Pneumatic Drive

PhysicalTesting

13

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Test Set-Up Results

PhysicalTesting

-1.0

0.0

1.0

2.0

3.0

4.0

5.0

6.0

108.200 108.230 108.260 108.290 108.320 108.350

Time (usec)

Volta

ge (V

)

44ns

-1.0

0.0

1.0

2.0

3.0

4.0

5.0

6.0

0 30 60 90 120 150

Time (usec)

Volta

ge (V

)

Zoom-in Region

14

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Simulation - CPCI Network

Slot 2Slot 1

1.0 V

Hot-Swap Card

Slot 8

1.0 V

1.0 V

10k-Ohm 10k-Ohm

10k-Ohm

1.5"

1.5"

1.5"

10-Ohm10-Ohm

10-Ohm

0.8"0.8"0.8"

DriverReceiver

SimulationTesting

15

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Simulation of the Discontinuity

-2.0

-1.0

0.0

1.0

2.0

3.0

4.0

5.0

6.0

0 5 10 15 20 25 30 35 40Time (ns)

Volta

ge (V

)

Driver CardHot-Swap CardReceiver Card

bus signal above receiver threshold

disengagement pointre-engagment point

SimulationTesting

16

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Simulation Summary

•• Initial point of contact - low normal forceInitial point of contact - low normal force•• Highly conductive interfaceHighly conductive interface•• Microscopic irregularitiesMicroscopic irregularities

A B C

Receptacle Pin

SimulationTesting

17

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Simulation Conclusions

=> Nanosecond discontinuities=> Nanosecond discontinuities=> Energy transfer onto the bus=> Energy transfer onto the bus

=> Signal Impact=> Signal Impact

Any conductive hot-swap interface!!Any conductive hot-swap interface!!

SimulationTesting

18

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Solutions

•• Software control during live insertionSoftware control during live insertion

•• Decrease low normal force zoneDecrease low normal force zone

•• Limit rate of energy taken from the busLimit rate of energy taken from the bus–– lower daughtercard capacitancelower daughtercard capacitance

•• physical restructuringphysical restructuring•• impact on signal integrity and timingimpact on signal integrity and timing

–– add resistance during engagementadd resistance during engagement

Solution

19

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Resistive Pin DevelopmentA B E

Receptacle Pin

C D

Resistive Coating

Unmated Receptacle and Pin

A B E

Receptacle Pin

C DSufficient Normal Force Engagement PositionSolution

20

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Isolation-to-Resistive Transition

-2.00

-1.00

0.00

1.00

2.00

3.00

4.00

5.00

6.00

0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0Time (ns)

Volta

ge (V

)


disengagement pointre-engagement point

gradual release onto backplane

acceptable level

Solution

21

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Electrical VerificationVDC

VmeasR1

Rtip


measdc

meastip VV

RVR-

1∗=

Oscilloscope

1K

StandardReceptacle

ResistiveTip Pin

+5v

22

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Resistance Profile


-0.5

00.5

11.5

2

2.53

3.5

44.5

55.5

6

0 0.5 1 1.5 2Time (msec)

Res

ista

nce

(kO

hms)

Cycle #1

Cycle #100

200-Ohm min

500-Ohm max

23

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Mechanical Testing

Different Lighting on Magnified Pin - 250 Engagement CyclesDifferent Lighting on Magnified Pin - 250 Engagement Cycles

•• High-cycle durabilityHigh-cycle durability•• Exceptional adhesion to goldExceptional adhesion to gold•• Absence of debrisAbsence of debris


24

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Qualification Testing

•• 2mm HM product spec (108-1622)2mm HM product spec (108-1622)–– 250 cycle durability250 cycle durability–– pin stagingpin staging

•• BellcoreBellcore GR-1217-CORE ( GR-1217-CORE (TelcordiaTelcordia))

•• Resistance requirementsResistance requirementsMaterial

Qualification

25

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Conclusion - The Problem

-2.0

-1.0

0.0

1.0

2.0

3.0

4.0

5.0

6.0

0 5 10 15 20 25 30 35 40Time (ns)

Volta

ge (V

)




Conclusion

26

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Conclusion - The Solution

Conclusion

-2.00

-1.00

0.00

1.00

2.00

3.00

4.00

5.00

6.00

0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0Time (ns)

Volta

ge (V

)




acceptable level

27

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

Conclusion - The Connector

Conclusion

28

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

The Quiet MateTM Contact

Conclusion

29

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

3.3V CPCI System Discontinuity

-1.50

-1.00

-0.50

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0

Time (ns)

Volta

ge (V

)Receiver CardHot-Swap CardDriver Card

disengagement point

re-engagement


30

ProjectBackground


PhysicalTesting

SimulationTesting

Solution


Conclusion

3.3V Quiet MateTM CPCI System

-1.50

-1.00

-0.50

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0

Time (ns)

Volta

ge (V

)Driver CardReceiver CardHot-Swap Card



acceptable level

Nanosecond Discontinuities & Quiet MateTM Contacts

1 04/20/00

1 February 10, 2000

The Quiet MateTM Contact

2 February 10, 2000

Nanosecond Discontinuties??

•• Initial Contact Point - Low Normal ForceInitial Contact Point - Low Normal Force

•• Microscopic IrregularitiesMicroscopic Irregularities

•• Highly-Conductive InterfaceHighly-Conductive Interface

•• Different From Pin “Bounce”Different From Pin “Bounce”

A B C

Receptacle Pin

Telecommunications equipment continues todrive faster and faster data rates andbandwidths (levels approaching 1Tbps).These high-speed systems must beextremely reliable and demand that the datathey process meets this level of reliability(five 9’s). Bus-type architectures, includingCompactPCI, provide the backbone for suchsystems. Therefore, bused signal systemsoperating within these requirements mustalso assure that the data they carry is errorfree, even during live insertion operations.During the development and testing of thesesystems, nanosecond discontinuities werediscovered with repeatability. Nanoseconddiscontinuities are electricalconnects/disconnects occurring in the arenaof several nanoseconds to tens and hundredsof microseconds. The electrical disruptionresulting from these intermittencies canadversely affect certain systems.Nanosecond discontinuities occur at theinitial point of closure between two highlyconductive separable interfaces. It isprimarily a result of the microscopic

irregularities and the relative motion of thestructures at the near-zero normal force areaof engagement.Nanosecond discontinuities are differentfrom “pin bounce”, in that they are not amechanical deflection and return of thecontact beam, but rather a result of therelative motion and initial contact of theconductive interface.

3 February 10, 2000

Measured Discontinuity

-1.0

0.0

1.0

2.0

3.0

4.0

5.0

6.0

108.200 108.230 108.260 108.290 108.320 108.350

Time (usec)

Volta

ge (V

)

44ns

-1.0

0.0

1.0

2.0

3.0

4.0

5.0

6.0

0 30 60 90 120 150

Time (usec)

Volta

ge (V

)

Zoom-in Region

Nanosecond discontinuities are extremelydifficult to capture with consistency. Anextensive understanding of the system andhow it responds to a live insertion event iscritical in detecting the phenomenon.Advanced (and expensive) test probes,measurement scopes, and specialized testfixtures aid in the identification of theintermittencies.Data samples from one such test set-up areclear enough to discern the fasterdisruptions, due to the optimally reducedtime constant of the test-bed. The measuredwaveforms were plotted on a voltage scaleto 5V against a physical engagement time-scale. The recorded data is then focused onthe specific areas around the transition edgesof engagement and disengagement.The measured waveform displaysintermittencies detected during a cardengagement event. When the irregularitiesare magnified [seen as the waveform on theright], a discontinuity of approximately 44nanoseconds is measured.The intermittencies are extremely erratic andvary anywhere from 100’s of microsecondsto single digit nanoseconds. The


2 04/20/00

phenomenon has been detected during bothengagement and disengagement operations.

4 February 10, 2000

Susceptible Systems

•• Bus ArchitecturesBus Architectures•• Hot-Swap CapabilityHot-Swap Capability•• High Fault-Tolerance RequirementsHigh Fault-Tolerance Requirements

Slot 2Slot 1 Slot X

Hot-Swap Card

Specific systems which are susceptible tonanosecond discontinuities are those whichhave a ‘bused’ or ‘multi-drop’ architecture,and possess the capability of hot-swappingcards. An operating system requires that atleast one card is present and ‘driving’, witha second (or additional) card(s) ‘receiving’,when a third card is inserted into the system.Typically, systems which require high fault-tolerance operation will be more affected bynanosecond discontinuities (as theintermittencies tend to adversely impactsystems on a relatively infrequent basis).

5 February 10, 2000

Hot-Swap System Diagram

BusDriver

Receiver Card

Hot-Swap Card

Pre-ChargeEstablished

1st Contact 1st Break 2nd Contact

NanosecondDiscontinuity Event

DataTransition

A simple timing diagram best illustrates theworst-case impact of a nanoseconddiscontinuity:- At first contact between a hot-swap cardand the backplane, all waveforms coincide(including the hot-plugged card), as allpieces are connected electrically andrunning.

- At the first break point (or beginning of thenanosecond discontinuity), the hot-swapcard electrically separates from the runningbus, and continues to hold the energy it hadwhen the separation occurred. This energywill leak off as a result of pre-charge orother termination devices, but it is typicallysubject to extremely slow time constants, asa result of the high resistance valueassociated with pre-charge elements.- During the nanosecond discontinuity event,the backplane bus may switch logic levels(in the example, the bus transitions fromhigh to low). High-speed drivers will allowthis transition to occur as quickly as 1-2nanoseconds, or possibly faster.- Shortly after the bus transition, thenanosecond discontinuity ends, and the hot-swap card comes back in electrical contactwith the backplane bus. Two separatevoltage potentials come together at onepoint.- As a result, energy will immediately betransferred to/from the daughtercard,depending on relative potentials. Thisenergy transfer has the capability ofimpacting a received/sampled signalwaveform significantly enough to result infalse data transfer.

6 February 10, 2000

Detection of the Discontinuity

-2.0

-1.0

0.0

1.0

2.0

3.0

4.0

5.0

6.0

0 5 10 15 20 25 30 35 40Time (ns)

Volta

ge (V

)




An actual system waveform plot displays theimpact of a nanosecond discontinuity on asampled waveform. The additionaltransferred energy boosts the level of thesignal above the threshold of the receiver(blue line), and could result in a falsesampling of data.


3 04/20/00

7 February 10, 2000

Nanosecond Discontinuity Impact

=> Nanosecond discontinuities=> Nanosecond discontinuities=> Energy transfer onto the bus=> Energy transfer onto the bus

=> Signal Impact=> Signal Impact

Any conductive Any conductivehot-swap interface!!hot-swap interface!!

Simulation and testing of hot-swap networkshas shown that nanosecond discontinuitiesresult in energy transferred onto or out of thebus. Additional energy transferred to anactive bus has the potential of altering signallevels. (Existing pull-up terminations areineffective during a nanoseconddiscontinuity, due to their relatively slowtime constants.) Nanosecond discontinuitiescan occur at any conductive hot-swapinterface at the near-zero normal force zonethat occurs at the very beginning ofengagement.

8 February 10, 2000

Solutions

•• Software control during live insertionSoftware control during live insertion

•• Decrease low normal force zoneDecrease low normal force zone

•• Limit rate of energy taken from the busLimit rate of energy taken from the bus–– lower daughtercard capacitancelower daughtercard capacitance

•• physical restructuringphysical restructuring•• impact on signal integrity and timingimpact on signal integrity and timing

–– add resistance during engagementadd resistance during engagement

Solutions to nanosecond discontinuities cantake several forms. The situation can becontrolled through software or hardware.Software solutions are possible, but slow theoverall response of the system, thus ahardware solution that does not slow thesystem is preferred.A decrease in the time of low normal forcemating would help to minimizediscontinuities. Stiffening backplanes andpre-loading spring members help to decrease

the amount of time that the phenomenon canexist and would reduce the probabilities of asystem detecting discontinuities. However,since there is always an initial point ofcontact, and since the inconsistencies occuron the molecular level, probabilities aredecreased but not eliminated.The final option is to limit the amount ofenergy the daughtercard injects/absorbsto/from the bus when the two metal surfacesmake electrical contact. One way ofapproaching this is to lower the capacitanceon the daughtercard as much as possible.Minimization of contact pads and othermetallic structures will reduce thiscapacitance and effectively make thedaughtercard more inductive. However, thisreduction in capacitance can have an ill-effect on the loading and timing of the bus,especially since the reduction must occur onall cards. In addition to signal integrityconcerns, the reduction in capacitance stillmay not guarantee that the energy flow willbe reduced to an acceptable level. Thesecond option to reducing energy flowduring the mating sequence is to create aresistive “shock absorber” between thereceptacle contact and the header pin, untilan acceptable normal force can beestablished. This resistance would then beinvisible to an operating system (followingthe live insertion event).

9 February 10, 2000

Resistive Application

Receptacle Pin

Resistive Coating

Unmated Receptacle and Pin

Receptacle Pin

Final Engagement Position

The requirement of a guaranteed error-freelive insertion points to the application of aresistive material on the pin as the bestsolution. The material must exist on aspecific region of the pin. The material


4 04/20/00

must exist on the tip of the pin in any areawhere the pin and receptacle could possiblymake first contact. The material must alsoextend back far enough on the pin to assurethat sufficient normal force is obtained priorto the receptacle sliding to and makingdefinitive contact with the gold surface ofthe pin. Sufficient normal force guaranteesthat the receptacle contact will not loseelectrical contact with the resistive materialor gold finish. The final resting position ofthe pin is noted by the vertical dashed line.

10 February 10, 2000

System Diagram Improved

BusDriver

Receiver Card

Hot-Swap Card

Hot-Swap Card &Quiet Mate Contacts

Pre-ChargeEstablished

1st Contact 1st Break 2nd Contact

NanosecondDiscontinuity Event

DataTransition

& Quiet Mate Contacts

An adjusted timing diagram displays thenew behavior of a hot-swapped card, and theresulting energy response seen at the inputof a sampling receiver device, when theresistive pin tip is used. The total amount ofenergy transferred is the same as before,however the rate at which it is transferred ismuch different. The reduced transfer rateprovides an impact at the receiver that nolonger can result in a false sampling of thewaveform data.


The Solution

-2.00

-1.00

0.00

1.00

2.00

3.00

4.00

5.00

6.00

0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0Time (ns)

Volta

ge (V

)




acceptable level

Electrical simulations provide waveformsshowing an actual bus signal’s response to alive insertion event, with the resistivematerial present on the tip of the pin. Theresulting rate of current into the backplanenetwork is not enough to force thecontinuously running data bus to cross thethresholds of the detecting receiver.


Qualification Testing

•• 2mm HM product spec (108-1622)2mm HM product spec (108-1622)–– 250 cycle durability250 cycle durability–– pin stagingpin staging

•• BellcoreBellcore GR-1217-CORE ( GR-1217-CORE (TelcordiaTelcordia))–– Mixed Flowing Gas (MFG)Mixed Flowing Gas (MFG)–– Temperature/HumidityTemperature/Humidity

•• Resistance requirementsResistance requirements

The new resistive material has been tested tothe 2mm HM product specification (108-1622) and qualified to the Bellcore GR-1217-CORE (Telcordia) specification. Thepin continues to meet all specifications. Thedurability of the new material exceeds thealready high (250 cycle) productrequirement. Pin staging requirementscontinue to provide sequential mating, dueto the location of the resistive material. Thematerial is also being qualified, throughresistance profiles, to the required resistanceranges specified by the simulations andverification testing.

timothy r. minnick - amp technical seminar · pdf filetechnical seminar h412 nanosecond...

Documents