Browse Prior Art Database

High productivity no-load inline dimensional inspection device

IP.com Disclosure Number: IPCOM000004698D
Publication Date: 2001-Apr-12
Document File: 8 page(s) / 114K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a design for a high productivity no-load inline dimensional inspection device. Benefits include a method for accurately measuring the distance between two planes on the same side of the component.

This text was extracted from a Microsoft Word 97 document.
This is the abbreviated version, containing approximately 74% of the total text.

High productivity no-load inline dimensional inspection device

Disclosed is a design for a high productivity no-load inline dimensional inspection device. Benefits include a method for accurately measuring the distance between two planes on the same side of the component.

The disclosed design is an effective, accurate measurement device developed for in-line measurement of critical parameters during high-volume manufacturing (HVM) production. This device meets the requirement to test 100% of the thermal transfer plate (TTP) component to ensure the quality of the package assembled state, which directly effects processor performance.

Measurement device accuracy of 1 ┬Ám, repeatability of 15% (with a tolerance range of 52 gm), and reproducibility of 30% is required. Conventional industry methods and devices do not meet all of these requirements.

The metrology design objective is to measure the distance between two planes on the same side of the component. The force applied to the component during measurement must be equivalent to that exerted during package assembly operation resulting in a functional measurement.

The device consists of two sets of probes that are calibrated to a 0-reference plane. The first set creates a best-fit plane using three or more fixed probes contacting the component reference surface (see Figure 1). The second plane is created using three or more floating probes contacting the part-measurement surface. Software takes data from these probes to create a best-fit plane and then determines the distance between both planes and/or the individual contact points. The developed measurement device has cycle time of less than 1.33 seconds, an accuracy of 0.17 gm, a repeatability of 3.3%, and reproducibility of 6.3% (see Figures 2-19).

Fig. 1

Moments

Mean

2.60780

Std Dev

0.00017

Std Error Mean

0.00004

Upper 95% Mean

2.60789

Lower 95% Mean

2.60771

N

16.00000

Sum Weights

16.0000

Fig. 2

Quantiles

Maximum

100.0%

2.6197

99.5%

2.6197

97.5%

2.6197

90.0%

2.6196

Quartile

75.0%

2.6194

Median

50.0%

2.6193

Quartile

25.0%

2.6192

10.0%

2.6188

2.5%

2.6184

0.5%

2.6184

Minimum

0.0%

2.6184

Moments

Mean

2.61925

Std Dev

0.00029

Std Error Mean

0.00005

Upper 95% Mean

2.61936

Lower 95% Mean

2.61914

N

30.00000

Sum Weights

30.0000

Fig. 3

Fig. 4

Oneway Anova

Summary of Fit

RSquare

0.446148

RSquare Adj

0.435497

Root Mean Square Error

0.002724

Mean of Response

2.476421

Observations (or Sum Wgts)

160

Analysis of Variance

Source

DF

Sum of Squares

Mean Square

F Ratio

Model

3

0.00093249

0.000311

41.8880

Error

156

0.00115760

0.000007

Prob F

C Total

159

0.00209008

0.000013

.0001

Means for Oneway Anova

Level

Number

Mean

Std Error

1

40

2.47357

0.00043

2

40

2.48019

0.00043

3

40

2.47545

0.00043

4

40

2.47647

0.00043

Std Error used a pooled estimate of error variance

Fig. 5

Fig. 6

Fig. 7

Analysis of Variance

Source

DF

SS

Mean Square

F Ratio

Prob F

Part

3

0.000932

0.00031

24.058

0.00036

Operator

3

0.000026

0.00001

0.997

0.45052

Part*Operator

9

0.000084

0.00001

3.746

0.03110

Day

1

0.000866

0.00087

154.732

0.00457

Part*Day

3

0.000018

0.00001

2.444

0.13092

Operator*Day

3

0.000006

0.00000

0.805

0.52184

Part*Operator*Day

9

0.000022

0.00000

2.373

0.01...