Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar...
-
Upload
ethan-porter -
Category
Documents
-
view
220 -
download
0
Transcript of Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar...
![Page 1: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/1.jpg)
Polar InstrumentsWhy Effective Er is not the same as
Bulk ErKen Taylor
September 2002
© Polar Instruments 2002www.polarinstruments.com
![Page 2: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/2.jpg)
FR4 Dielectric
• Core and Prepreg have different r
• Both are Mixtures of Resin and Glass Fibers• Materials are non-homogenous
– r specified for laminate is the bulk value
– r for glass ~ 6.1 r for epoxy ~ 3.2
• So significant local variations occur for r
![Page 3: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/3.jpg)
Typical E-field distribution
Embedded Microstripr
![Page 4: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/4.jpg)
Microphotograph of FR4 structure
5 mil
Prepreg
P
P
C
Core Glass fibers Er = 6
Resin Er = 3.2
![Page 5: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/5.jpg)
FR4 structure
Er = 3.2
Er = 6.0
Bulk Er value in this direction is 4.2 approx
![Page 6: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/6.jpg)
Field in FR4 structure
3.1
3.1
4.2
![Page 7: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/7.jpg)
Field distribution in Differential Pair
Impedance value
Increases as
Er and C decrease
![Page 8: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/8.jpg)
FR4 structure
3.1
3.1
4.2Impedance Value Increases
![Page 9: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/9.jpg)
Resin Layer in Differential Microstrip
Glass/Epoxy mixture
Resin-only region
![Page 10: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/10.jpg)
Resin Layer in Differential Stripline
![Page 11: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/11.jpg)
Conclusions
• Over simplified modeling of differential structures leads to impedance discrepancies of several ohms
• Resin flow into region coplanar with tracks increases the impedance of typical embedded microstrip or stripline by 4%
![Page 12: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/12.jpg)
Conclusions
•Your predictions must consider–Structure–Track dimensions
•Thickness•Widths•Spacing•Etch taper•Symmetry
–Dielectric layer composition and Effective r values
![Page 13: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/13.jpg)
Conclusions
• Accurate implementation of predicted impedance requires– Accurate knowledge of dielectric composition
• Dielectric layers
• Resin region and local r value
– Accurate build of predicted dimensions• Track width• Track spacing• Track etch taper
– Increased etch taper leads to enlarged resin region
• Track symmetry– Loss of symmetry also affects differential performance
![Page 14: Polar Instruments Why Effective Er is not the same as Bulk Er Ken Taylor September 2002 © Polar Instruments 2002 .](https://reader036.fdocuments.us/reader036/viewer/2022062315/5697bf8f1a28abf838c8d419/html5/thumbnails/14.jpg)
Recommendation
• Build samples of small differential structures• Measure impedance using CITS500s• Microsection and micro-dimension the sample• “Back” calculate to obtain the Effective Er using Si6000• In future, use that value of Er for
– similar structures– same dimension range– same materials
• For small differential structures in FR4, expect Er(effective) ~ 3.7 (About 10% less than spec.)