I would like to thank everyone for their responses. I find it very
important to get an outside perspective on things when you have the "big
wigs" upstairs saying "We are supposed to be EXPERTS...we are supposed to
find EVERYTHING. Why can't we do that?" (Grin) I'm sure none of you have
ever heard something like that. Again, thanks for all your input!
Rich Lasko
Badger Electronics
-----Original Message-----
From: Smith, Russell (US LA) [mailto:[log in to unmask]]
Sent: Monday, March 03, 2003 2:15 PM
To: [log in to unmask]
Subject: Re: [TN] First Pass Yields
You are so right Steve, I knew someone would bring an example like that
, I was thinking sorting coins by denomination myself , the apples are a
great one too! .
Any time something is open to interpretation it is! Remember the infamous "
that depends on what your definition of IS is "...
I guess the point I am really trying to make is 100 percent inspection can
be made redundant,if the manufacturing processes are controlled. It costs a
lot of money to sort and scrap , or sort and repair. Money that should be
spent in prevention, prevent a defect and you not only save that part, but
the time would would have used for remake, can now be making you a salable
product.
It has always struck me odd in our industry , that a company will drop
several million on AOI equipment , and never hire an engineer to use the
data that can be generated by the AOI to solve the issues. I once worked for
an organization that the VP said: "the only reason he had a quality
department was his customers expected it". , and he didn't believe in
process controls. (it goes without saying that company is no longer in
business).
Are we having fun yet ! , I once heard someone at Nepcon many years ago
state that there were over a million variables and interactions in making a
PC board. I think I would have been much happier growing Green Apples, but
then............
best regards always , to the industry , that makes the world function!
My hats off to all of you that make it work no matter what it takes!!!
Russell Smith
-----Original Message-----
From: Shoda, Steve [mailto:[log in to unmask]]
Sent: Monday, March 03, 2003 11:27 AM
To: 'TechNet E-Mail Forum.'; Smith, Russell (US LA)
Subject: RE: [TN] First Pass Yields
I concur with Russell, except 80% inspection effectiveness is not the best
attainable. 100% effectiveness can be achieved. Visual inspection is
dependent on the skill of the inspector and the complexity of the inspection
performed. For example, someone who can see in color can sort yellow and
red apples with 100% effectiveness, but sorting ripe red apples is not as
discernable and will not be as effective.
For electronics solder inspection, inspecting for bridges demonstrates very
high effectiveness, but inspecting solder wetting is very complex and
requires more appraiser training to get the effectiveness needed (and even
there is an indefinite debate among the experts).
The 80% effectiveness phenom is taught as a rule of thumb, and is a good
approximation to use when developing and working to error proof processes.
However, when applied to a particular situation, actual measurement system
analysis is required before a conclusion regarding the effectiveness can be
stated.
Steve
-----Original Message-----
From: Smith, Russell (US LA) [mailto:[log in to unmask]]
Sent: Monday, March 03, 2003 1:56 PM
To: [log in to unmask]
Subject: Re: [TN] First Pass Yields
Dear Rich:
You are right, a 100% visual inspection is at best 80% effective,
every time it is done, The big issue with visual inspection is the
effectiveness is extremely variable depending on the class of defects that
you are looking for , and other items such as the visual acuity, training
level,fatigue and moral level of your inspectors.
What you might be looking for is From Juran-Melsheimer:
percent defects correctly identified = d-k / d-k+b
where d= number of defects reported by inspector
k= number of good units rejected by inspector
d-k = true defects found by inspector
b= defects missed by inspector
d-k+b =true defects originally in product.
A simple test to show the effectiveness is the following :
assume that the letter "G" is the defective product in the following story.
Allow someone to read the story and count all the "G"' s that are used. give
them 3 minutes. There are eighty -five g , determine the percentage they
catch and viola you have your effectiveness and accuracy. It should give you
the ammunition you need
WHILE STROLLING THROUGH A GLEN, A GIDDY ENGLISH
GIRL TRIPPED ON A RATHER LARGE , ALMOST GIGANTIC
FROG. THE GIRL STAGGERED BUT REGAINED HER
FOOTING AND WAS ABOUT TO GO ON WHEN THE FROG
BEGAN TO SPEAK AND GESTICULATE TO GAIN THE
GIRL'S ATTENTION. "I HAVE NOT ALWAYS BEEN A
FROG" HE CROAKED. THE FROG'S GREEN COLORING
SEEMED TO GLOW BRIGHTLY AS HE CONTINUED, "I
WAS ONCE A GRACIOUS KNIGHT, A GENTLEMAN
CALLED GALLANT GEORGE GRENVILLE,BUT WAS
CHANGED INTO THIS GHASTLY GREEN FROG YOU NOW SEE BY
AN UNGODLY, MAGICAL GENIE. THE SPELL CAN ONLY
BE BROKEN IF I GAIN A GIRL'S GOOD GRACES AND
SPEND A NIGHT IN HER GARDEN." THE AGOG GIRL WAS
SKEPTICAL, OF COURSE. SHE GAZED AT THE FROG'S
PLEADING EYES AND SOON HER GIDDY NATURE GAVE
WAY TO HER DOUBTS. GIGGLING,SHE DECIDED TO
GRANT THE FROG'S WISH AND TOOK HIM HOME
STRAIGHTWAY, PUTTING HIM BY HER GARDEN GATE.
THAT NIGHT THE GIRL SLEPT GRANDLY AND SURE
ENOUGH , WHEN SHE AWOKE THE FOLLOWING MORN-
ING, THERE ALONGSIDE HER GARDEN GATE WAS THE
GRACIOUS KNIGHT, GEORGE GRENVILLE. WELL,
STRANGELY ENOUGH, FOR A LONG LONG TIME THE
GIRL'S MOTHER DID NOT BELIEVE THAT STORY.
Inspection accuracy is rarely better than 85%. If you add complexity, like
other type fonts, different format , multiple colors , or distractions such
as noise etc. the accuracy can be very low
With the complexity of the product you are dealing with and the number of
defective opportunities per assembly, visual inspection is not the answer.
It should be used as a method to collect data to be used back in the
manufacturing process.It sounds like at this point in time the only rapid
band aid for the problem is to perform some type of testing. I fought many a
customer about paying for inspection , and invariably they reply with " I am
buying good product, how you assure I get it is up to you" . That should be
accounted for in the original pricing .
If you follow modern quality theory then your guiding philosophy should
be " If you control your processes, and all your processes are in control
then you don't need to inspect the product." of course this assumes that
you are controlling the items that effect the outgoing quality of the
product!
While all this seems an insurmountable challenge, the rewards are vast, and
will carry directly to the bottom line of the ledger sheet assuring that
the company achieves it ultimate goal of amassing great piles of coin of the
realm!!!!!
PS: I never was one for the "World Class" moniker, I always felt that we
should strive for "Galactic Class" ! Yes in my youth I was a Trekky!
Another overused phrase while I am on the subject is "State of the Art" ,
Don't you go to a museum to see art? Wouldn't "State of the Science " be
more appropriate?
Thanks all for allowing my ramblings on this dreary Monday Morning in LA .
Russell Smith
Original :
Good morning everyone,
I am having a debate with regards to first pass yields and how testing
strategy (ie.. ICT or Functional Test) plays a role.
I am looking for any "factual data" or "industry studies" that have been
done to show the difference in yields between a batch of assemblies that are
subjected to only visual inspection and to those batches that receive visual
inspection and some type of ICT or functional test. I have my own internal
data that shows the differences but want to compare that against the
industry.
Summary - I have a customer that is expecting 99% First pass yield
(including board functionality) to their facility but does not utilize or
want ICT or functional testing done to their products. Now the old rule of
thumb that I've gone by for many years is that visual inspection is approx.
80% accurate and there are bound to be defects that slide through this
visual inspection process. To consistently achieve yields in the high 90%'s
additional test strategies should be looked into. Note: The boards in
question range anywhere from 200 components per board to approx. 2000
components per board.
Any feedback or published industry articles documenting the yield
differences between visual inspections only and using ICT/Functional testing
would be very helpful!
P.S. I have a meeting at 1:30pm today (Central Time) so any feedback before
then would definitely be appreciated.
Thanks,
Rich Lasko
Badger Electronics
---------------------------------------------------
Technet Mail List provided as a free service by IPC using LISTSERV 1.8e
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt or (re-start) delivery of Technet send e-mail to
[log in to unmask]: SET Technet NOMAIL or (MAIL)
To receive ONE mailing per day of all the posts: send e-mail to
[log in to unmask]: SET Technet Digest
Search the archives of previous posts at: http://listserv.ipc.org/archives
Please visit IPC web site http://www.ipc.org/html/forum.htm for additional
information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700
ext.5315
-----------------------------------------------------
---------------------------------------------------
Technet Mail List provided as a free service by IPC using LISTSERV 1.8e
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt or (re-start) delivery of Technet send e-mail to
[log in to unmask]: SET Technet NOMAIL or (MAIL)
To receive ONE mailing per day of all the posts: send e-mail to
[log in to unmask]: SET Technet Digest
Search the archives of previous posts at: http://listserv.ipc.org/archives
Please visit IPC web site http://www.ipc.org/html/forum.htm for additional
information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700
ext.5315
-----------------------------------------------------
---------------------------------------------------
Technet Mail List provided as a free service by IPC using LISTSERV 1.8e
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt or (re-start) delivery of Technet send e-mail to [log in to unmask]: SET Technet NOMAIL or (MAIL)
To receive ONE mailing per day of all the posts: send e-mail to [log in to unmask]: SET Technet Digest
Search the archives of previous posts at: http://listserv.ipc.org/archives
Please visit IPC web site http://www.ipc.org/html/forum.htm for additional
information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315
-----------------------------------------------------