Susan, you are correct, up to the following point:
"If the coupon from one panel fails and all panels have been tested - those
that have satisfactory coupon may be shipped and the failed one scrapped."
In producing printed wiring boards, the manufacturer must correctly scale
their artwork so that the scaling of the laminated circuits (of differing
patterns, foil thickness, and even base laminates which all cause
stretching/shrinking of the individual laminates.) to match the drill
pattern. If we have a manufacturing panel with 12 individual pwbs and a
coupon strip on opposite sides, we may see that in the microsection review
one section from one end of the panel shows zero (undersize rim), while the
coupon from the opposite end in the same direction shows 5 mils of rim. The
pwbs near the "good" coupon, and part of the way down to the "bad" coupon
can reasonably be assumed to meet rim requirements. Knowing that we can't
sell pwbs on "assumptions", the manufacturer must prove that a portion of
the panel meets rim requirements. Since the manufacturer is normally
responsible for determining the manufacturing panel size and the number of
boards to build on each panel. they may elect to scrap one end of the panel
(bad end) and retest farther up the panel. They may use other coupons
farther up the panel or boards (scrapping all boards past the sample area).
All IPC documents for rigid boards that I can think of allow using boards
for conformance testing for rim as long as the requirements are met (number
of holes, directions, hole to pad relationships, etc.). So the manufacturer
uses coupons or boards that are further down the panel as test coupons to
verify that both ends of the "new size panel" are conforming. This is what
the manufacturer is trying to say that he did. What is the value in
throwing away good boards that can be verified good in this fashion? PWBs
don't always follow the " One bad apple spoils the barrel " philosophy.
This is a linear failure that can be verified and measured. The
nonconforming boards are scrapped, the customer's delivery schedule is
hopefully met, and the costs of rebuilding the good boards are eliminated
while the manufacturer re-scales their tooling for the next build. There
are no AABUS issues here, it's all allowed by IPC specs. For the US board
manufacturer to survive in today's market, they must build product in a
"quick turn" environment that does not allow for missed deliveries because
the manufacturer "just missed" the correct scale factor. They must get the
customer as many good boards as soon as possible. The scenario of scrapping
entire panels for U/S rim on one end of the panel will only lead to forcing
the manufacturer to build more panels than are needed to make sure that
they have enough "in case" they have a panel fall out of scale on one end.
This directly leads to increased costs for the manufacturer, so that they
are no longer competitive, and the business goes overseas where they can
afford to overbuild quantities. I hope I haven't rambled on too much, but
I've written this in several "windows of opportunity" and haven't checked
it over closely. Let's be careful here.
Tom Kemp
Rockwell Collins/ Collins Printed Circuits
Quality Manager
Susan Mansilla
<[log in to unmask] To: [log in to unmask]
M> cc:
Sent by: Subject: Re: [IPC-600-6012] Panel Acceptance
IPC-600-6012 Mail
Forum
<IPC-600-6012@ipc
.org>
01/30/2003 08:24
AM
Please respond to
"(Combined Forum
of D-33a and
7-31a
Subcommittees)"
Ron
The way the spec is written, if there are no deviations from the buyer.
All
panels have coupons on them.
Then there are sampling plans that are geared to the different classes and
are C=0 acceptance. In many of the small lots the
coupons from each panel must be tested. There are all sort of paragraphs
that say "except upon agreement
between buyer and seller", but that is how the base document is structured.
If the coupon from one panel fails and all panels have not been tested -
you
must test additional coupons or consider the entire lot to have failed.
If the coupon from one panel fails and all panels have been tested - those
that have satisfactory coupon may be shipped and the failed one scrapped.
As you indicate - the buyer can put whatever additional requirements and
testing protocols on the drawing and that would
take precedence.
Susan
-----Original Message-----
From: IPC-600-6012 Mail Forum [mailto:[log in to unmask]]On Behalf Of
McIlnay, Ronald
Sent: Wednesday, January 29, 2003 4:31 PM
To: [log in to unmask]
Subject: Re: [IPC-600-6012] Panel Acceptance
Hello Scott,
Thanks for responding. I follow you but now put on the "statistical and
reliability hats", since the original intent of having a coupon was to
perform DPA on coupons (verses arrays or PWB's) which should be
representative of the manufactured panel and the entire lot if the coupon
was only applied to selective panels. If the coupons were applied to all
panels, then you would have to test all coupons to determine what panels
are
acceptable. However, a second failure of a coupon on a different panel
would trigger significant cause to suspect the entire lot. A third failure
on a third different panel would dictate a lot failure condition.
I'm not indicating I agree with the logic, but to interpret a requirement,
I
feel you have to understand what the original intent was that caused the
initial requirement. This is the logic that the original coupon usage was
intended to expose. It just seems that the procuring agency would be the
one to "dictate" what was a acceptable condition for acceptance and what is
not. Cost would then be based on the fabricators ability to predict what
their yield would be.
Ron McIlnay
-----Original Message-----
From: Scott Bowles [mailto:[log in to unmask]]
Sent: Wednesday, January 29, 2003 10:29 AM
To: [log in to unmask]
Subject: Re: [IPC-600-6012] Panel Acceptance
Ron,
I believe you are confusing arrays with panels. We are using the term
panel
as the manufacturing panel the fabricator put the individual boards or
arrays on. What Susan is trying to get correct language on is that if you
have coupon failures on a "manufacturing panel" then all of the individual
boards or arrays are considered nonconforming. You are correct that many
companies accept X-outs on arrays but they are x'd-out for reasons other
than for coupon failures.
Scott
----- Original Message -----
From: "McIlnay, Ronald" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Wednesday, January 29, 2003 12:55 PM
Subject: Re: [IPC-600-6012] Panel Acceptance
John/Susan/& Committee
This issue seems to be a procurement issue and not a specification issue.
Companies that are outsourced may not care as they are only interested in
individual boards if they don't lay out the array (leaving array design to
CM & fabricator). Other companies want to control the array so they submit
that as part of the drawing package. Other companies have sufficient data
to and yield information to know that a panel with defects sufficient to be
scrapped would indicate all boards in that array or possibly even the
entire
lot are suspect so they may want to reject the entire lot. For companies
using CM's, the procuring agency may be the final customer, a central
procurement group at a different facility (but still the final customer) or
the CM who may be able to handle x-outs/punch-outs or not.
Every company should know their own requirements for acceptance and
therefore should control yield issues from a procurement standpoint. I
would hate to see a specification issue drive yield requirements without it
being a requirement to be negotiated between fabricator and procuring
agency, whether they are the final customer or not. Otherwise, it would
seem you would have to identify individual requirements for the different
classes of product.
I think John is on the most correct track by identifying the
procurement/acceptance requirements (ref: MIL-PRF-55110F) as it seems this
issue should be on a purchase order or general procurement agreement.
Does this make sense, or am I missing something here (please respond back
to
me ply if I am). Thanks.
Regards,
Ron McIlnay
Sr. Procurement Engineer
Medtronic Physio-Control
Phone: 425 867 4500
e-mail: [log in to unmask]
-----Original Message-----
From: Scott Bowles [mailto:[log in to unmask]]
Sent: Tuesday, January 28, 2003 4:19 PM
To: [log in to unmask]
Subject: Re: [IPC-600-6012] Panel Acceptance
John/Susan,
The following does not explicitly state that all boards on a panel fail if
a
coupon fails but it states the purpose and relationship of the coupons.
This is from IPC-2221 Section 12.1, second paragraph, last sentence:
"Test specimen are used in these types of tests as representatives of the
printed boards fabricated on the same panel." But, the next paragraph then
goes on to say that "the design and location of the test specimen are
critical to ensure that the specimen are truly representative of the
printed
boards".
Be careful in your wording about "any failure...shall constitute a failure
of the entire inspection lot...". That is why we are using coupons, to
represent that specific panel and the boards on that panel, not the entire
lot.
You may be able to extract some language from MIL-PRF-55110F, A.4.6.2.2
Sampling Procedures. Maybe use something like "all rejected printed wiring
boards or panels of printed wiring boards based on coupon analysis shall
not
be supplied as compliant without a waiver from the customer".
Scott
----- Original Message -----
From: "John Perry" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Tuesday, January 28, 2003 2:56 PM
Subject: Re: [IPC-600-6012] Panel Acceptance
Hi Susan,
I agree that there is a lack of explicit verbiage. I suppose one could
start out with section 4.2, Acceptance Tests, of 6012a wherein it states
"Acceptance testing shall be performed as specified in Table 4-3 [Sampling
Plan] to the requirements of this specification and IPC-6011 using either
test coupons and/or production boards." One could then go from there to
4.3.2.3 in 6011 where it states "When sampling plans are used, any failure
on any of the samples [test coupons] identified per the Performance
Specification requirements shall constitute a failure of the entire
inspection lot...."
I realize this is a stretch and is really an inference. Maybe we can beef
up these two sections in each of the documents to better address this. Or
take your previous recommendation to allow the customer to waive the
non-compliant condition.
John
-----Original Message-----
From: Susan Mansilla [mailto:[log in to unmask]]
Sent: Tuesday, January 28, 2003 1:44 PM
To: [log in to unmask]
Subject: Re: [IPC-600-6012] Panel Acceptance
Thanks, John
I, too, found that - but it really doesn't address why we test coupons OR
explicitly say "if a coupon fails to meet specification requirements, all
boards from that panel are considered to fail to meet specification
requirements."
Susan
-----Original Message-----
From: IPC-600-6012 Mail Forum [mailto:[log in to unmask]]On Behalf Of
John Perry
Sent: Tuesday, January 28, 2003 12:28 PM
To: [log in to unmask]
Subject: Re: [IPC-600-6012] Panel Acceptance
Hello Susan,
The only other verbiage that I have been able to find that may be
applicable
in this case is in the IPC-6011 Generic Specification:
4.3.2.3 Acceptance - Acceptance of product for delivery
shall be based on product that has passed the testing
requirements shown in the applicable Performance Specifi-
cation. When sampling plans are used, any failure on any
of the samples identified per the Performance Specification
requirements shall constitute a failure of the entire inspection
lot, and shall be processed in accordance with the
requirements of paragraph 4.3.2.4.
4.3.2.4 Rejected Lots - If an inspection lot is rejected,
the supplier shall screen out the defective units, i.e. 100%
lot inspection, or other documented supplier quality system.
The defective units shall be dispositioned per the supplier's
quality system.
John
-----Original Message-----
From: Mike Hill [mailto:[log in to unmask]]
Sent: Monday, January 27, 2003 10:33 AM
To: [log in to unmask]
Subject: Re: [IPC-600-6012] Panel Acceptance
Susan,
1. I would think this situation might fall under para 4.2.2 "Referee Tests"
in 6012.
...back to "microsection defects that are isolated or random in nature"
2. ERGO, It appears this person is saying the registration defect is in an
"isolated" area of the board.
3. If this para applies, he would need two additional sets of sections
without defects.
Mike
-----Original Message-----
From: IPC-600-6012 Mail Forum [mailto:[log in to unmask]]On Behalf Of
Susan Mansilla
Sent: Friday, January 24, 2003 11:53 AM
To: [log in to unmask]
Subject: [IPC-600-6012] Panel Acceptance
Even tho we are not finished with the first topic of Zone B I would like to
start another conversation.
I have a note from a board fabricator that was sent to his customer that
says
"Group A microsection submissions for this part number do nt meet your
internal annular ring requirements. This notification relates to parts
from
panels 1 and 2. In an effort to minimize cost and meet customer delivery
requirements, it is our policy to alsvage partial panels on lots with
internal annular ring failures. In this case we used additional
microsections or rgistration coupons to show that the internal annular ring
on some of the parts on the panels mentioned ar compliant. The
non-compliant parts next to the failed microsection coupons have been
scrapped. We hope that yu see the benefit of this practice and that it has
not created you any inconvenience."
This note was submitted because some of the coupons I received for Group A
testing for this lot, as an independent check, were rejected since they did
not meet specification or drawing requirements for annular ring.
The board fabricator asked "well where does it say that the whole panel
fails, if the coupons fail?"
Of course at this moment I have not been able to find those exact words.
SO - do we need to add those explicit words with the instructions that ONLY
THE BUYER OF THE BOARDS CAN WAIVE NON-COMPLIANT CONDITIONS ON A PANEL OF
BOARDS.
Thanks again for all the help
Susan Mansilla
Technical Director
Robisan Laboratory, Inc
6502 E 21st Street
Indianapolis, IN 46219
317-353-6249 phone
317-917-2379 fax
www.robisan.com
|
