Steve,
Data collection can be a very valuable thing, provided it is specific enough
to guide you in the right direction for problem solving. After all that is
what true SPC is about.
Before one can properly set-up SPC, one must measure the processes
capability to bench mark the current processes being used. There are many
different tools available for this. The "Memory Jogger for SPC" breifly
lists a large majority of these tools. Once you have your processes bench
marked you will know exactly where your problem areas are and what kind of
data needs to be collected.
We usually start in RMA and Rework. The defect codes are very specific, and
when "other" is used too many times we review the defect codes again to see
what defect code needs to be added to the list. We have even gone as far as
bagging and tagging defective parts for further evaluation to determine what
type of defect code should be used for that particular failure. This is a
never ending process. The defect code list grows quickly at first, and then
levels off slowly.
Once we have this information we take the highest detractor and look at each
process to see what has caused these defects. Using the DOT STAR approach
we begin our problem solving process. D - Define the problem, O - Observe
the problem/process, T - Think of possible solutions, S - Select a solution,
T - Test the solution, A - Act on the solution (implement), R - Review the
solution. During the Act portion of the problem solving process we usually
leave an SPC chart behind. This allows us to go back and quickly review the
process to ensure it is still in control and the solution we implemented has
a positive affect.
To just throw SPC charts out on the manufacturing floor, will buy you
nothing. I have two years worth of worthless data to prove this. Sure the
charts will show you your sporadic spikes, but they will never address the
chronic failures. The sporadic spikes are easy to fix, and we have all had
years of pratice at fine tuning our fire fighting skills. It is the chronic
problems that are the hard ones to fix. The constant 6% failure rate that
we build our entire manufacturing processes around. These are the problems
you want to attack, and you will need much more than SPC charts to do it.
The Winsconsin University Continuing Education program offers the following
training programs (through Chad Cullen with the Internationl Quality
Institute at (916) 933-2318).
"Reducing Process Variations (Using the DOT STAR approach)
"Measuring Process Capabilities"
"SPC for Short Run Production"
These are great training classes, and will give you a head start on your
journey through the continuous improvement process.
Good Luck!
Sincerely,
Kathy Palumbo
Manufacturing Engineer
IPC Certified Instructor
Viking Components, Inc.
30200 Ave. De Las Banderas
Rancho Santa Margarita, Ca. 92688
BUS: (949) 643-7255 ext. 352
Fax: (949) 459-5360
CELL: (949) 422-4145
E-Mail: [log in to unmask]
Web: http//www.vikingcomponents.com/
-----Original Message-----
From: Stephen R. Gregory [mailto:[log in to unmask]]
Sent: Monday, October 25, 1999 1:55 PM
To: [log in to unmask]
Subject: [TN] SPC programs, how make the best use of the efforts..
Hi All,
I'd like to ask any input or opinions on how to implement a SPC program that
will really be meaningful. I don't want to waste time and effort on
something
if all it's going to produce is "window dressing"...
Our business is very high-mix/low volume stuff. From boards that are very
straight forward, standard 50-mil pitch stuff, to those that have every part
on the board (SOIC's, CQFP's) cut and formed, pre-tinned, and then placed
and
reflowed. A big run for us is 100-boards, most of the time it's 50 and
under.
I guess what I'm saying is that there are TONS of variables with a lot of
the
products we build here.
We do a lot of hi-rel stuff here, very custom, almost hand built boards. We
also hand place a lot of components here...for example the resistors on a
lot
of the boards come to us in little 2" X 2" matrix trays that can't be
machine
placed. Any design problems that cause defects on the boards most of the
time
won't be corrected because the document trail that has be completed is so
daunting that our customers tell us to not ask for changes because it won't
happen, and our time is best spent figuring out a way how to deal with it.
So what I'm wondering in a situation like this, what good is SPC going to
do?
I think the time spent trying to correctly identify all the variables and
gather meaningful data would be better spent in training and working with
our
operators to learn all the little nuances with the product we build so we
can
deal with them. What was proposed to me for example in screen printing, was
to create a simple attribute chart for example;
1. Is the paste coverage complete? Yes/No
2. Is it registered on the pads? Yes/No
3. Is the paste smeared? Yes/No
or such simple attributes such as that, and then plot them on a chart. To me
it'll be a waste of time. We don't have anything to quantify any
measurements...no paste height measuring equipment, or X/Y measuring
equipment, everything will be done by the ol' Mark-I / MOD-I eyeball with
all
it's subjectivity...but I am still being asked
to establish an SPC program in SMT.
Am I fighting this too much? Or is there some means that will be worthwhile
to make this happen? Oh yeah, it's gotta be cheap (next to nothing) too...
Thanks in advance!
-Steve Gregory-
##############################################################
TechNet Mail List provided as a free service by IPC using LISTSERV 1.8c
##############################################################
To subscribe/unsubscribe, send a message to [log in to unmask] with following
text in
the body:
To subscribe: SUBSCRIBE TECHNET <your full name>
To unsubscribe: SIGNOFF TECHNET
##############################################################
Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional
information.
If you need assistance - contact Gayatri Sardeshpande at [log in to unmask] or
847-509-9700 ext.5365
##############################################################
##############################################################
TechNet Mail List provided as a free service by IPC using LISTSERV 1.8c
##############################################################
To subscribe/unsubscribe, send a message to [log in to unmask] with following text in
the body:
To subscribe: SUBSCRIBE TECHNET <your full name>
To unsubscribe: SIGNOFF TECHNET
##############################################################
Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional
information.
If you need assistance - contact Gayatri Sardeshpande at [log in to unmask] or
847-509-9700 ext.5365
##############################################################
|
