🚀 go-pugleaf

In a message dated 9/3/97 10:39:06, you wrote:

<<Hi,

I have a situation that I need some help with.

I have a specification that is 2.870-2.872 on an inlay that mates with the
handle.
Parts normally run 2.870-2.874. Supplier is not capable of improving part.

Mating parts (handle) specification is 2.875 +-.002

Problem:
Supplier cannot improve process for Inlay.
Supplier on Handle cannot improve process.
Eng is unwilling to change prints in case parts start coming in to a
different part of the specification range.
WE are writing up each shipment as a deviation to the inlay supplier.

Both sides of each spec need to be maintained to prevent spec crossover or
gaps between the two parts.

Any suggestions?

Jeff Hubbard
Quality Engineer
Mailto:Jeffh@buckknives.com
WEb: www.Buckknives.com
>>

My understanding of the problem is this:

Two mating parts with tolerances; 2.870-2.872 & 2.875 +-.002
Min design clearance is .001
Actual data indicates a range of 2.870-2.874 for the first part.
No available data on second part in the posting.

Problem:  There is an interference in the parts do to lack of process
capability.

Organizational issues:
Supplier cannot improve process for Inlay.
Supplier on Handle cannot improve process.
Eng is unwilling to change prints in case parts start coming in to a
different part of the specification range.
WE are writing up each shipment as a deviation to the inlay supplier.


Suggestions:  ( Opinions based on my experience and the limited information
given, my biases may be many.)

First:
Either measure yourself, or ask the vendors for statistical data on the
characteristic(s) in question.  Specifically, I'd ask for capability data,
short term and long term. (short term: within a production run if quantities
are large, over 1000/Long term: 1-3 months worth of production).

I'd be interested in determining if they have a short term or long term
capability issue.  I'm assuming that there is an issue with both suppliers
(implied not directly commented on).  A short term problem maybe harder to
fix (probably machine related).  Long term problems may be set-up or material
related.  Angles to pursue for improvement, set-up variation reduction (gage
R&R, set-up verification, work instructions), material variation ( tighter
specs, sample runs with process adjustments to account for material
variation).

Don't ask for just a raw Cpk number, but for the indices and the raw data
time and date stamped.  What I'd look for is good short term capability but
lot to lot variation.

Gauge R&R has been noted as being the root of many product quality issues.
 Some 6 sigma companies have found this to be true with their supplier base.
 If the supplier's measurement system is questionable, they may not know they
really have a problem, and it will likely show up as lot to lot variation.

Personal experience has shown me to look at the measurement system (gauge
R&R, and customer-supplier measurement systems) first.  One product line I'm
responsible for had a long history of problems.  Reviewing the  measurement sy
stem showed R&R problems.  Revising that then highlighted the process
problems that were the root of the historical issues.  The new measurement
system showed us how to change processing methods (more material handling
 and sample processing) than machine issues.  Cpk's went from .9-1.1 to
1.9-3.0.

Second:
Investigate their process and their process controls.  Is there significant
room for improvement that would bare on this issue.  Are there control plans,
SPC used, etc.

Third:

Assuming that the vendors are giving you the best they can, what do you do?
 The next best solution is do you really need all that tolerance.  Your
planned for tolerance range is .001 to .007.  Would it hurt to move the
better controlled process to a new nominal that provides the proper clearance
based on process capability?

The design engineers need to answer that.  However, you can help by providing
the process capabilities of the two parts, current scrap/rework/yield costs,
yours and the suppliers).  It is not uncommon for designers to over tolerance
parts to "Ensure Quality".  They will need good data on the process and costs
for justifing a print change and determine a valid tolerance band.

Fourth:
Are there any other vendors with better capabilities?  Pursue this if you
need the current tolerances, and the current vendors can not provide
acceptable product.  If there are no better vendors, go back to engineering
with purchasing and layout the problem at their door with a report stating
that the design is not manufacturable.  The part/system may have to be
redesigned.

Fifth:

100% sort to yield acceptable product.  This provides more data to drive you
back to 4.

These are my thoughts.  I'm looking forward to any others.

Robert Drensek, CQE, CQA
Quality Engineer
RNAV@AOL.COM