(+)+(+)(+)+(+)(+)+(+)(+)+(+)(+)+(+)(+)+(+)(+)+(+)+(+)+(+)+(+)+(+)
(+)+(+)(+)+(+)(+)+(+)(+)+(+)(+)+(+)(+)+(+)(+)+(+)+(+)+(+)+(+)+(+)
(+)+(+)
(+)+(+) S.M.I.L.E(sm)
(+)+(+)
(+)+(+) Setup Minimization Improves Line Efficiency
(+)+(+)
(+)+(+) The Lean Changeover e-letter
(+)+(+)
(+)+(+) Published monthly by Changeover.com
(+)+(+)
(+)+(+) Written by John R Henry, CPP
(+)+(+)
(+)+(+)
(+)+(+)(+)+(+)(+)+(+)(+)+(+)(+)+(+)(+)+(+)(+)+(+)+(+)+(+)+(+)+(+)
(+)+(+)(+)+(+)(+)+(+)(+)+(+)(+)+(+)(+)+(+)(+)+(+)+(+)+(+)+(+)+(+)

###############################
###############################
###
### OCTOBER 2005
###
###############################
###############################
============================================================

This newsletter is a free service of Changeover.com. Visit our website at
http://www.changeover.com or contact us at
john@...

Please feel free to forward this newsletter to anyone you think might be
interested

To subscribe send a blank e-mail to: quickchangeover-subscribe@yahoogroups.com

To unsubscribe send a blank e-mail to:
quickchangeover-unsubscribe@yahoogroups.com

Past issues are indexed and archived at www.changeover.com/newsletter.htm

=============================================================

Be sure to visit the packaging blog at www.packnation.com Lots of good
stuff there.

For those of you lucky enough to live in Puerto Rico, great news. I will be
teaching the course in Packaging Technology (MMP-6236) starting November
14. This is part of a program leading to a Master of Science in
Manufacturing Engineering degree. The syllabus is not online yet but
general program information is available at
http://www.pupr.edu/gs/gprogram.asp?ID=60

I hope I will see some of you there.



SOME THOUGHTS ON...
STARTUP Part II

In last month's newsletter I defined "start-up" as being the time spent
after line re-start getting the line settled down and running at normal
speed and efficiency. It is characterized by frequent stoppages for
adjustment, high reject levels and a variety of minor and sometimes major
problems in operation. As I said, start-up is caused by variability in
materials or variability in clean-up and/or set-up. For the purposes of the
following discussion, let us assume that all materials, components and
product is perfectly centerlined in it's specification. (That may not
always be a valid assumption)

Most of the following discussion will focus on set-up but clean-up can be a
source of variability as well. Chuck style cappers can sometimes build up
mold release from the caps of the previous run. If not properly cleaned,
this will cause slippage with resulting variation in on-torque in the
succeeding run. In general, the following comments, if not the necessarily
the examples, will apply to both clean-up and set-up variation.

In my experience, the absence of standard operating procedures (SOP's) for
changeover are the biggest cause of in my experience. Even where SOP's
exist, they may not be well written or may not be followed. Typically,
people are taught to do changeover by example. They then each evolve their
own techniques over time. Absent good implemented documentation, it is rare
to find two people doing the same changeover the same way. Often, even the
same person will not do changeovers the same way from one time to the next.
There is a best way. IT must be found and documented.

The first key to driving variation out of the set-up is good documentation,
properly implemented and used. An SOP must be more than a checklist. It
must include detailed instructions on everything that is to be done.

An SOP describing how machine settings are to be made is useless unless it
is possible to accurately measure them. Thus, measurability becomes the
second key to driving variation out of set-up. (This assumes set-up is
achieved via adjustment rather than changeparts. I will discuss this below)
If a capper set-up SOP calls for a "slight clearance" between the chuck and
the cap, this can be interpreted in a variety of different ways. Some might
think it means almost touching. Others may think it means 1/4". Even where
there is agreement on the distance, if it is done by eyeball, a short
mechanic may set it differently than a taller mechanic simply because of
difference in viewpoint.

All adjustments must be made to a fixed measurement. Digital indicators
work well for this. Scales and gauges are also useful. When using scales, a
critical element is the pointer. It must come to a fine point and must be
as close as possible to the number to be read. Otherwise, different people
may read it differently depending on the relationship of their eye to the
scale. If gauges are to be used, they should be permanently mounted to the
machine. There are several ways to do this including the revolver gauge I
mentioned in last month's newsletter.

Physical position is not the only thing that needs to be measured. Machine
speeds are all too often ignored and set by eye and feel. Conveyor speeds
are critical to a smooth running line yet I seldom see tachometers on
conveyors. All equipment or components capable of speed adjustment must
have tachometers. Temperature and pressure must also have adequate
indicators and controls. A temperature control consisting of a single turn
knob, numbered 1-10 with no indication of actual temperature is not
adequate. I strongly recommend digital rather than analog indicators as
they are less subject to interpretation error.

Machine construction and state of repair must be considered as well.
Machines with excessive wear cannot be set precisely no matter how good the
measuring device. A worn gate on a case packer, for example, may never stop
in exactly the same place twice. Some machines are built of light duty
construction or without safety clutches. When the inevitable crash occurs,
the machine will bend and all settings are now out the window, requiring a
complete recalibration and revision of the SOP to reflect the new setpoints.

There are two schools of thought on whether machines should be built with
changeparts or adjustments to accommodate different product sizes. The
argument in favor of changeparts is that, by eliminating adjustment, they
eliminate the possibility of mis-adjustment. While this is true, it also
eliminates the ability to compensate for variability in components or
product. (I'm still ambivalent about whether this is a good or a bad thing)
The use of changeparts also reduces the ability to rapidly produce new
products. Instead of simply making the adjustments and running, custom
changeparts must be purchased which can take anywhere from weeks to months.

Measuring startup time is difficult. It is easy to see when to start the
clock. It is harder to see when the line is running at normal speed and
efficiency. The best way to do this is to have a SCADA system that provides
a graphical display of line conditions. When the line showing speed
adjusted for efficiency flattens out, startup is over. Paul Zepf of Zarpac,
in his excellent book suggests that startup be considered over when there
is a 15 minute period with no stoppages. Other plants may have their own
methodologies. One idea I recently ran across intrigued me so I made it
this months "Tip of the Month". The key is that a metric be developed and
used. While the metric must be as accurate as possible, it is more
important that it remain constant. Only constancy will allow for comparison
and observation of trends.

Variability is the greatest enemy of quality. It is the great enemy of
changeover time as well.


TIP OF THE MONTH...
A QUICK AND DIRTY WAY TO MONITOR STARTUP TIMES

In a workshop recently at Kellogg's cereal plant we were discussing this.
David Bryan (Randy, I am not positive it was David. Do you have any way to
verify?) one of the mechanics, suggested a solution that is pretty simple.
His suggestion was to measure the amount of time it took to produce the
first 5,000 units of good product. This does not actually measure startup
time. It does give an indication of how well the changeover was done. By
comparing these times across batches, it is fairly easy to see whether
startup times are increasing, decreasing or holding constant. The reasoning
behind this is that the longer the startup time, the longer it will take to
reach the 5,000 products.

The 5,000 is not a magic number. The cut-off number should be picked based
on the normal amount of production before the line settles down. It may be
more or less in specific cases.

Additionally, if different products run at different speeds, a multiplier
can be used to normalize times. The time it takes to produce 5,000 units of
a product running at 100PPM would be halved to make it comparable to the
time needed to produce 5,000 units of a product running at 200PPM