Ahhhh, a little computer communication problem. The dot in front of the
number of sigmas is not a decimal point, but it was part of a table list, a
fat round black circle. Next time I will have to change things like that to
another symbol.
The actual numbers were slightly rounded off.
Ahne.
-----Original Message-----
From: Louis Hart [mailto:[log in to unmask]]
Sent: Tuesday, August 25, 2015 6:00 AM
To: TechNet E-Mail Forum; Ahne Oosterhof
Subject: RE: [TN] PWB, PCB End Of Line Shorts/OPENS detection/Test
Ahne, Victor, et al.,
Check the numbers below. Are they supposed to be 0.2 sigma, 0.3, etch? I
think 3.0 sigma means 27 defects (or nonconformances, if you want to use a
non-pejorative word) per ten thousand.
My understanding from ~1989 was that the defective fraction at six sigma
was predicated upon a shift of the mean of a normal distribution of 1.5
sigma from the center of a tolerance range. The defect fraction was thus
greater than would have been if the distribution were centered within that
range.
The tale of the bags of good and defective parts, with what may have been
different numbers, I heard at that time, but thought it likely apocryphal.
What appears to be the colloquial meaning of six sigma now is social rather
than mathematical. Louis Hart
-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Ahne Oosterhof
Sent: Monday, August 24, 2015 1:08 PM
To: [log in to unmask]
Subject: Re: [TN] PWB, PCB End Of Line Shorts/OPENS detection/Test
.2 Sigma: 308K errors per million (69% accuracy). 30 errors per
100
.3 Sigma: 66.8K errors per million (93.3% accuracy). 6 errors per
100
.4 Sigma: 6.2K errors per million (99.4% accuracy). 1 error per
200
.5 Sigma: 233 errors per million (99.97% accuracy). 1 errors per
5000
.6 Sigma: 3.4 errors per million (99.999997% accuracy). 1 error per big
number
In the past a good number was '3-sigma', or 6 errors per 100. This company
ordered 100 parts from Japan and required meeting a '3-sigma' error rate. So
they got the batch back with 100 parts in the package, of which 6 rejects
were in a separate bag. (They were obviously telling: we can do better than
3 sigma.)
Today in many instances companies strive to meet '6-sigma', not always
successful yet. And even with excellent process control, including
excellent incoming inspection it remains very difficult.
Copied:
""Although one of the key concepts of Six Sigma is to strive for near
perfection, the practical goal of Six Sigma programs is to continually
improve the rate of accuracy as it approaches that nearly perfect goal. As
the quality control of an enterprise progresses, it traverses lower sigma
levels that have less accuracy. Six Sigma, however, is not just a measuring
stick for performance, nor is it a technique for improving performance: Six
Sigma as we know it addresses corporate culture and seeks to change it into
an environment that is at every point optimized for quality.
Six Sigma, therefore, is an attempt to unify all employees of a corporation
into a unified team that works together to produce high quality goods and
services.""
Yet it still is very difficult to get circuit boards at low sigmas,
especially if you are using fine pitch components.
Good yields come from everyone paying attention and doing the right thing
correctly.
Ahne.
PS: When Motorola started pushing for '6-sigma' somebody explained to me why
such an effort was important. His example: say a nurse handles babies all
day. How many times is she allowed to drop one? 3-sigma? 4 sigma? 5 sigma? 6
sigma?
-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Victor Hernandez
Sent: Monday, August 24, 2015 5:38 AM
To: [log in to unmask]
Subject: [TN] PWB, PCB End Of Line Shorts/OPENS detection/Test
Fellow TechNetters:
In today's world of High Volume Production. What is a good YIELD Rate.
Back some 30 years I remember 89 to 92% yield depending on board complexity.
Our AOIs weren't as good and accurate as todays. I need some kind of warm
and fuzzy number. We used to put external image numbers on all boards to
help sort repeat images printing anomalies. What happened to the PERFECT
COUPON that was to HELP solve these type of issues.
I also see frequent Front to Back cores misregistration!!!!! What's up
with that anomaly.
Victor,
______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please contact helpdesk at x2960 or [log in to unmask]
______________________________________________________________________
______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please contact helpdesk at x2960 or [log in to unmask]
______________________________________________________________________
______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please contact helpdesk at x2960 or [log in to unmask]
______________________________________________________________________
|
