78 The FABRICATOR JANUARY 2016
both computer chassis and go-carts is the same. But for most processes, the
cycle time will differ depending on the product, which is why we need to work
from weighted averages.
To determine the operators required for a particular process, we add the total
labor input for the process—that is, the weighted averages of OCT and SUT—
and divide by our takt time of 115.2 seconds. Laser cutting, for instance, has a
weighted average OCT of 37. 5 seconds and SUT of 15 seconds (again, this is per
piece). So here, we get ( 37. 5 + 15)/115.2 = 0.4557, which we round up to 0.5.
You’ll notice in Figure 3 that for forming our machine cycle time is listed as
zero. This is because the manual machine doesn’t run independently of the operator cycle time, so we don’t need MCT in our analysis. As this shows, we need
0.8 (round up to 1) operator, so we obviously also need one press brake. The
same reasoning applies to hardware
When we add up the operator values
in all the process boxes in the value
stream (see Figure 4), we find that
on this particular day, with this particular product mix, we have a required
staffing level of 2. 6, which we’ll round
up to 3.
Staffing with three people probably
isn’t sufficient, though. Again, we run
a job shop, and as such we need to
deal with imbalances between work
content of each process. Figure 5
shows the labor content (weighted average OCT + SUT) of each process. For
every process we have a healthy buffer
except one: forming. The press brake
operation takes almost 100 seconds to
form the components for one unit ( 75
OCT + 22. 5 SUT = 97. 5). This again is
an average of a variety of actual cycle
times, so we’re really in danger of exceeding our 115.2-second takt time.
Staffing the value stream with just
three people doesn’t leave much room
for error. When we add up all the average labor times—that is, the average
OCTs and SUTs—from every process
box in the value steam, we find that
this particular mix of products has
an average of 302.6 seconds of labor,
or just over 5 minutes. Divide this by
three people, and we get a little more
than 100 seconds of work each to perform a 115.2-second takt time.
This may be doable in a highly pre-
Assigning Standard Work
dictable assembly line, but it’s a bit
more challenging in a job shop, where
the work content varies so much from
product to product. Moreover, people
don’t work every second of every day,
so we need to account for breaks and
other activities. For this reason, it’s a
good rule of thumb to load each per-
son to no more than 85 percent of his
or her capacity during a shift. We also want to drop that percentage a bit for each
additional work assignment. For this reason and others, we might be wise to
staff this value stream with up to four people.
This problem becomes abundantly clear when we assign standard work—that
is, work each person does each takt time or cycle. Figure 6 shows the standard work for a team of three, plus the total labor content cycle time (average
OCT and SUT from Figure 6) for each process step. If Stan’s job is to gather materials ( 8. 6 seconds), run the shear ( 40. 5 seconds) and the laser ( 52. 5 seconds),
that gives him a total cycle time of 101.5. In other words, he’s loaded to about 88
percent of his total work capacity, which is high.
Overcoming the Job Shop VSM Conundrum
All this sounds great, but again, no job shop processes the same product mix day
after day, month after month. Customer demand changes, as does the product.
Here is where this VSM approach really shines. If we have the formulas set in
Excel or similar software, we can change certain variables and then redistrib-
ute the standard work. Say our volume stays the same (250 units) but now you
have 200 computer chassis ( 80 percent of total) and 50 go-carts ( 20 percent).
Because we’re still producing 250 units a day, our takt time doesn’t change. But
the change in product mix does change our cycle time averages, increasing the
total work content from 302.6 seconds to 327.19 seconds. Being able to quickly
adjust the work content to balance the line may mean having a very flexible
workspace (see Figure 7).
Entering 1 beneath a person’s name assigns all of the work of a particular pro-
cess to one person; entering a 2 splits that time in half; entering 3 divides the
time into thirds, and so forth. If you enter a 2 under Dan’s name in this case, don’t
forget to enter a 2 under someone else’s work assignment—in this case, Jan.
Series 4® Press
With Automatic Feed
Economically installs PEM®
self-clinching hardware quickly,
safely and accurately
• Uses same tooling components as the
Series 2000® and Series 3000™ presses.
• Automatically installs nuts, studs, and
standoffs from M2.5 to M6 / #2-56 to 1/4”.
• Pneumatic and PLC logic control system.
• Touch-screen for operator friendly
• Vibratory feeder bowl for automatic
• Integral point-of-operation safety.
• Insertion force adjustable from 1.8 to
53. 4 kN / 400 to 12,000 lbs.
app now available
Standard Work Stan Dan Jan Leanne
Percent Loaded >>>> 80% 72% 72% 60%
Gather Materials 8. 6 1
Shear 36.0 1
Laser 48.0 1
Grain 54.0 2 2
Form 111.0 2 2
Hardware 69. 6 1
Total Operator Cycle Time 327.2
This shows standard work for a product mix of 200 units of computer components
and 50 units of go-cart components.
Standard Work Stan Dan Jan Leanne Barney
Percent Loaded >>>> 88% 85% 90% 0% 0%
Gather Materials 8. 6 1
Shear 40. 5 1
Laser 52. 5 1
Grain 60.0 1
Form 97. 5 1
Hardware 43. 5 1
Total Operator Cycle Time 302.6
This shows standard work for the team assigned to the mild steel value stream. Stan,
Dan, and Jan are fully loaded, so in this case it would be a good idea to bring in Leanne and Barney to spread the work load.
This value stream mapping method
has been a tipping point for many
make-to-order shops where a
standard product is not a reality.