Ehem...(clearing my voice as I climb on the soap
box)
With
tracking efficiency, you are essentially trying to get the most amount of
product out of a set of machines that you possibly can. To figure that,
you want to know what is theoretically possible and how close you come to
theoretical. Since many brands of pick and place machines keep a record of
what state they are in, you can query that data and get some eye opening
results.
I have
found that cycle times can be figured three different ways, giving very
different results. First, looking at a 24 hour period, how many boards
came out of that machine. This is real world cycle time planning should
use, and is the ultimate bottom line. This time includes set up time and
any decreases in change over will positively affect you cycle
time.
Second, looking at the time that the line is actually
running divided by the number of board produced. This is the cycle time
that determines your constraint and paces the line.
Third
is the time it actually takes for the machine to assemble the board. This
is the theoretical cycle time.
I'm
guessing here, but you probably do 2-3 change overs per 24 hour period.
Not using docking carts, change overs could be as high as two hours.
Assuming you get it down to 10 minutes, that means you gain 5 1/2 hours of
number two time. That would be awesome, and would improve efficiency by
25%. This is probably an extreme case, so adjust the numbers
accordingly.
However, if you look at the numbers between number two
time and number three time (theoretical time), you will typically find a
difference by a factor of three or more. This is the "mother load".
Your efficiencies are lost by not focusing all attention on your constraint
and keeping that machine fed, up and running. In other words, any
time the constraint "waits" for product to arrive so it can start working, you
have lost time. Any time the machine stops for any kind of error, you have
lost time. Any time the machine stops for a feeder to be reloaded, you
have lost time. These time losses are only a dozen seconds or a minute or
two, but they add up over a 24 hour period. As a constraint, there is no
reason that machine shouldn't be able to run continuously at near its
theoretical limit. I'm not talking about improving the placement time,
just keep that machine busy. By improving the "working" time on your
constraint, there is an easy 100% improvement in line efficiency, and it doesn't
cost anything. Just an attitude change from management down to the
production line to identify the constraint and focus on keeping it
working.
I have
worked in factories where this concept has improved output by a factor of three,
therefore I'm extremely sold on it. As far as what is possible, I have
seen lines where the constraint was "working" 80% of the time the machine
was plugged into the wall. What is realistic... I have heard reports from
machine manufactures that the CM industry is at about 20-40% machine
working time. Of course, those at 80-90% working time are not
high mix, low volume.
Not
that improving change over time is bad, I just think that attention should be
focused on the place with the greatest potential for
improvement.
Ryan
Grant
Dear TechNet subscribers,
I
am curious about the level of productivity of your SMT lines, particularly for
those in a high mix environment. We are getting pressure to reduce
changeover and cycle times, as I am sure most of you are too, and I am curious
as to what level of effectivity (hours available in the day do you actually
build product) that you are at. Would 80% be reasonable? Also, how
efficient are you at building your product to standards, for those who track
efficiency?
As far as changing over
from one job to the next, what are your average times? It was suggested to
me by management that the goal for changeover should be 1 minute or less.
I am a reasonable person, and I believe that single digit (SMED)
changeovers, like 8 or 9 minutes might be reasonable, but 1 minute is a stretch
that I can't see happening, even when using docking carts for the feeders.
Thanks in advance for your input.
Howard Watson
SMT Manufacturing
Engineer
--------------------------------------------------- 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
-----------------------------------------------------
AMETEK/Dixson