ATTENUATION IN PCB TRACES DUE TO PERIODIC DISCONTINUITESGustavo Blando, Jason R. Miller, Istvan Novak, Cheryl PrestonSun MicrosystemsJim DeLapAnsoft Corporation

2

Outline• Introduction• Background theory and simulation methodology• Test board definition and measurements• Measurement to simulation correlation• Periodic discontinuities characteristics• Parameterization studies• Summary and conclusions• Q & A

3

• Periodic discontinuities, what are they?• Refers to the loading of some element, for example, a

transmission line, at periodic intervals• Many kinds can be found in practice, ranging from the

periodic loading in multi-drop busses, to the periodic loading of plate capacitors

• Which ones are we going to study and why?• Traces routed over perforated planes.

We will present parametric studies on realistic cases and will show its effects

Introduction

Connpin-field

BGA pin-field

In most cases, a great percentage of the PCB trace length has to be routed over perforated areas

4

Lumped discontinuity

⎥⎦

⎤⎢⎣

⎡⎥⎦

⎤⎢⎣

⎡⎥⎦

⎤⎢⎣

⎡⎥⎦

⎤⎢⎣

⎡⎥⎦

⎤⎢⎣

⎡⎥⎦

⎤⎢⎣

⎡⎥⎦

⎤⎢⎣

⎡

4

4

11

11

3

3

2

2

11

11

1

1

IV

DCBA

IV

DCBA

IV

DCBA

IV

dd

dd

Background theory and simulation methodology

• Periodically loaded transmission lines can be analyzed by identifying a unit cell, which is repeated along the structure, and calculating the loaded propagation delay.

• Unit cell length -> half wave resonance• Number of unit cells -> size of the resonance• Discontinuity size/kind -> size of the resonance

dip

00

00

0

/}]1)){cosh(2/()[sinh(}]1)){cosh(2/()[sinh(

)sinh()2/()cosh(

ZdZZdCdZZdZBdZZdDA

d

d

d

−+=++=

+==

γγγγγγ

).sinh().2/().cosh().cosh( 0 dZZdd dp γγγ +=

Length of one unit cell

Unit Cell

• From a practical stand point, we could, for example, get the unit cell from a field-solver

• Or, do different type of analysis, for example

5

Design guidelines

Simulation plan

Real case simulations, definitions andparameterization

Resultsand conclusions

TransmissionLine generation

Simulation of:● Saturation● Frequency dependency

Gain modelingconfidence level

HFSS:● Solve a single unit cell

Same?

no

yes

Concatenate “N” unit cells

Create a test board:● Easily modifiable●Simple to measure●With “N” unit cells

HFSS:● Solve a single unit cell

Same? Concatenate

“N” unit cellsManyvariations

Periodical-disc effects study

6

TDR response (rho)

-0.04

-0.02

0

0.02

0.04

0.06

0.08

0.1

0.0E+0 2.0E-10 4.0E-10 6.0E-10 8.0E-10 1.0E-9 1.2E-9

Time [sec]

Half wave resonance1/ 2*500mils*150ps/in=6.6GHz

Test board 2.5 mm wide, 6in long microstrip on top of a 800 mil wide, 63 mil thick FR4 dielectric

500mils pitch

Fixed 125 mils from the edge of the trace to center of the hole

Hole diameter increased from: 125,164,194,250mils

Edge launch SMA

7

Test board measurementsS21 magnitude [dB]

-2.5E+01

-2.0E+01

-1.5E+01

-1.0E+01

-5.0E+00

0.0E+00

0.0E+0 5.0E+9 1.0E+10 1.5E+10 2.0E+10

Frequency [Hz]

164-mil194-mil250-mil

Solid plane125-mil

Half wave resonance

Full wave resonance

Baseline

Dip

8

Simulation setup• Ansoft HFSS, v10• Unit cell approach used• Material: FR4

• Dk = 4.5• Tand = 0.03

• The hole diameters were progressively increased to match the test board

• MATLAB was used to perform the concatenation

12 un

it cell

s

9

Measurement correlation (1)

0 0 10 20-25

-20

-15

-10

-5

0

Frequency [GHz]

dB

0 10 20-14

-12

-10

-8

-6

-4

-2

0

Frequency [GHz]

dB

250mils holediameter

Simulated

194mils holediameter

Used a frequencyindependent dK

Goodcorrelation

Measured

10

Measurement correlation (2)

0 10 20-14

-12

-10

-8

-6

-4

-2

0

Frequency [GHz]

dB

Measured

Sim

164 mil holediameter

0 10 20-14

-12

-10

-8

-6

-4

-2

0

Frequency [GHz]

dB

Measured

Sim

125 mil holediameter

11

Periodically loaded line characteristics• Doing some mathematical post-processing, several

periodic discontinuity effects can be studied.• Study 1: Examine how the location of the

resonance dip changes as a function of unit cell length

• Study 2: Examine the first resonance amplitude as a function of the number of cascaded cells for a 500-mil long unit cell

12

Frequency vs. pitch dependency12

unit c

ells

0 5 10 15 20-45

-40

-35

-30

-25

-20

-15

-10

-5

0

frequency[GHz]

dB

0 5 10 15 20200

300

400

500

600

700

800

frequency[GHz]

leng

ht[m

ils]

1/(2*tpd)

13

Saturation effect

0 25 50 75 100-0.18

-0.16

-0.14

-0.12

-0.1

-0.08

-0.06

-0.04

-0.02

0

Number of Cells

dB

5.9 6.1 6.3 6.5 6.7

-0.18

-0.16

-0.14

-0.12

-0.1

-0.08

-0.06

-0.04

-0.02

0frequency[GHz]

dB

(Bas

eline

–Di

p)/ (

numb

ers o

f Cell

s)

0.5in

Incre

asing

numb

er of

cells

14

Parameterization• Examine the impact of additional losses introduced

by the periodic discontinuities in real-world designs• Large number of variables including:

• Number of periodic discontinuities• Distance between the discontinuities• Separation between the discontinuity and the

trace• Size of the discontinuity

15

Parameterization cases

• Case 1: Trace routed through a pin field, such as a connector, where the trace would periodically encounter a hole located on either side of a trace

ws

• Case 2: Trace routed near to a single cutout but due to misregistration and manufacturing tolerances, the trace gets routed over a portion of the plane cutout

-s

Example: 4 mil line width, BGA (1mm, 39.37 mil), 30 mil antipadRange Antipad edge to trace edge separation

Different core misregistration +/-5 mils+/-3 mils

-3.315 mils (-82%)Same core misregistration -1.315 mils (-33%)

16

Parameter ranges

• 50-ohm microstrip was simulated using a 4-mil wide trace and a 4-mil thick dielectric (2%, εr=4.5)

• Number of periodic discontinuities : 10• Distance between the discontinuities: 500 mils

(λ/2=6.6 GHz)• Trace to hole separation: -4 (-100%) mils to 6 mils

(+150%)• Size of the discontinuity (antipad diameter) : 50mils

to 150 mils

17

0 (0%)2 (50%)

4 (100%)6 (150%)50

100

1500 (0%)

0.5 (6%)

1 (11%)

1.5 (16%)

2 (22%)

2.5 (27%)

3 (32%)

edge-to-edge [mils(%line whole-diam [mils]

extra

-loss

[dB

(%ba

selin

e)]

Case 1, third resonanceParameterization

(20.23 GHz)

18

0 (0%)2 (50%)

4 (100%)6 (150%)50

100

1500 (0%)

0.5 (8%)

1 (16%)

1.5 (23%)

2 (31%)

edge-to-edge [mils(%line whole-diam [mils]

extra

-loss

[dB

(%ba

selin

e)]

Case 1, second resonanceParameterization

(13.44 GHz)

19

Case 1, first resonanceParameterization

0 (0%)

2 (50%)

4 (100%)

6 (150%)4060

80100

120140

0 (0%)

0.1 (3%)

0.2 (6%)

0.3 (9%)

0.4 (11%)

0.5 (14%)

0.6 (17%)

edge-to-edge [mils(%line width)]hole-diam [mils]

extra

-loss

[dB

(%ba

selin

e)]

(6.89 GHz)

20

Case 1, etch to antipad separationParameterization

0 (0%)2 (50%)

4 (100%)6 (150%)50

100

1500 (0%)

0.5 (6%)

1 (11%)

1.5 (16%)

2 (22%)

2.5 (27%)

3 (32%)

edge-to-edge [mils(%line whole-diam [mils]

extra

-loss

[dB

(%ba

selin

e)]

0% 50% 100% 150%-0.5

0

0.5

1

1.5

2

2.5

3

edge-to-edge [%line-width]

extra

-loss

[dB

]

0 1 2 3 4 5 6

-0.5

0

0.5

1

1.5

2

2.5

3

edge-to-edge [mils]

f3

f2

f1

100 mil antipad diameter

21

Case 1, antipad diameter changeParameterization

60 70 80 90 100 110 120 130 140

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

2.2

hole diameter [mils]

extra

-loss

[dB

]f3

f2

f1

0 mil separation0 (0%)

2 (50%)4 (100%)

6 (150%)50

100

1500 (0%)

0.5 (6%)

1 (11%)

1.5 (16%)

2 (22%)

2.5 (27%)

3 (32%)

edge-to-edge [mils(%line whole-diam [mils]

extra

-loss

[dB

(%ba

selin

e)]

22

-4 (-100%)-2 (-50%)

0 (0%)2 (50%)40

6080

100120

0 (0%)

2 (22%)

4 (43%)

6 (64%)

8 (85%)

10 (106%)

edge-to-edge [mils(%line widhole-diam [mils]

extra

-loss

[dB

(%ba

selin

e)]

Case 2, third resonanceParameterization

(20.23 GHz)

23

-4 (-100%)-2 (-50%)

0 (0%)2 (50%)40

6080

100120

0 (0%)

1 (16%)

2 (31%)

3 (46%)

4 (61%)

5 (77%)

6 (92%)

edge-to-edge [mils(%line widthole-diam [mils]

extra

-loss

[dB

(%ba

selin

e)]

Case 2, second resonanceParameterization

(13.44 GHz)

24

-4 (-100%)-2 (-50%)

0 (0%)2 (50%)40

6080

100120

0 (0%)

0.5 (14%)

1 (28%)

1.5 (41%)

2 (55%)

edge-to-edge [mils(%line widhole-diam [mils]

extra

-loss

[dB

(%ba

selin

e)]

Case 2, first resonanceParameterization

(6.89 GHz)

25

Case 2, etch to antipad separationParameterization

-100% -80% -60% -40% -20% 0% 20% 40%0

1

2

3

4

5

6

7

edge-to-edge [%line-width]

extra

-loss

[dB

]

-4 -3 -2 -1 0 1

0

1

2

3

4

5

6

7

edge-to-edge [mils]

f3

f2

f1

58 mil antipad diameter-4 (-100%)

-2 (-50%)0 (0%)

2 (50%)4060

80100

1200 (0%)

2 (22%)

4 (43%)

6 (64%)

8 (85%)

10 (106%)

edge-to-edge [mils(%line widhole-diam [mils]

extra

-loss

[dB

(%ba

selin

e)]

26

Case 2, antipad diameter changeParameterization

50 60 70 80 90 100 110 1201

2

3

4

5

6

7

8

9

10

hole diameter [mils]

extra

-loss

[dB

]

f3

f2

f1

-4 mil separation-4 (-100%)

-2 (-50%)0 (0%)

2 (50%)4060

80100

1200 (0%)

2 (22%)

4 (43%)

6 (64%)

8 (85%)

10 (106%)

edge-to-edge [mils(%line widhole-diam [mils]

extra

-loss

[dB

(%ba

selin

e)]

27

Conclusions (1)• Due to periodic discontinuity the slope of transfer function

increases and sharp dips appear in the loss profile • Extra attenuation varies almost linearly with frequency

Increases linearly with frequencySimplified discontinuitymodel

Z = jωL(ω)

Decreases weakly with frequency• Due to this effect, at lower frequencies (less than ~3 GHz),

for the type of discontinuities studied here, the additional loss is not too pronounced

• The extra attenuation starts to sharply increase as the separation between antipad and trace approaches zero

28

Conclusions (2)• Due to misregistration the loss can increase dramatically• Loss has a close to linear relationship to antipad diameter• Loss scales with the number of discontinuities until it

saturates• Distance between the discontinuities determines the lowest

resonance frequency

29

Next steps• Understand different types of discontinuities

> Including a complete via structure (some studies have already been done)

> Understand the crosstalk effect between same layers and adjacent layers due to perforated planes

> Understand the same effect on stripline structures, (all the measurements and simulations have been done on microstrips)

> Create behavioral models to account for these effects

Gustavo BlandoGustavo.blando@sun.com

THANKS!!!!

ATTENUATION IN PCB TRACES DUE TO PERIODIC DISCONTINUITES

31

Eye mask degradation due to periodic discontinuities (1)• Four cases have been created:

> Six inch trace without perforation: F-6in, (Baseline)> Six inch trace formed by twelve, 500mils perforated unit

cells: D-6in (Dip)> Eighteen inch trace with no perforation: F-18in> Twenty seven inch trace without perforations: F-27in

• Different bit times have been used• Internal differential eye contour, area and UI have

been computed for every simulated bit time

32

Eye mask degradation due to periodic discontinuities (1)• Simulations have been run using a 10ps rise-time.• The discontinuity has been generated fitting the half

wave resonance of the 250mils hole diameter using a RLC T-network

33

Frequency domain channel characteristics

0 2 4 6 8 10 12 14 16 18

x 109

-50

-40

-30

-20

-10

0

Freq

FULL-6in[1,2]D-194mils,6",500mils-sep[1,2]FULL-27in[1,2]FULL-18in[1,2]

• Notice the cases have been created to study the difference between:> Baseline> Discontinuity> Longer uniform line,

presenting the same attenuation than the periodic discontinuity case at the resonance frequency

> Medium line length, falling in between the baseline and the long case

D-250mils

34

Back-Up-Slides

0.2 0.4 0.6 0.80

0.05

0.1

0.15

0.211.1111GB/s

0.2 0.4 0.6 0.80

0.05

0.1

0.15

0.2

10GB/s

0.2 0.4 0.6 0.80

0.05

0.1

0.15

0.2

9.0909GB/s

0.2 0.4 0.6 0.80

0.05

0.1

0.15

0.2

8.3333GB/s

0.2 0.4 0.6 0.80

0.05

0.1

0.15

0.2

7.1429GB/s

0.2 0.4 0.6 0.80

0.05

0.1

0.15

0.2

5GB/s

0.2 0.4 0.6 0.80

0.05

0.1

0.15

0.2

4GB/s

0.2 0.4 0.6 0.80

0.05

0.1

0.15

0.2

3.125GB/s

0.2 0.4 0.6 0.80

0.1

0.2

2.5GB/s

UI

V

F-6inD-6inF-27inF-18in

35

Horizontal eye opening• Note how the

periodical discontinuity case is closely following the 18” case. Maybe a shorter line, 12” would be closer.

• The 27” has by far the biggest horizontal eye closure as seen on the eyes in the previous page

2 4 6 8 10 120.4

0.5

0.6

0.7

0.8

0.9

1

UI

f[GHz]

F-6inD-6inF-27inF-18in

36

Eye area• The area

computation, follows the same trend as the horizontal eye opening computation

• Clearly in all these cases, the line with periodic discontinuities is the one that have more fluctuations

2 4 6 8 10 120

0.05

0.1

0.15

0.2

0.25

area

,UI*V

f[GHz]

F-6inD-6inF-27inF-18in