Reprinted from Quality in Manufacturing
GAGING GEARS FOR
HARLEY-DAVIDSON
By Donald B. Dobbins
Senior Technical
Editor
Recently I had an opportunity to sit down with
Donald L. Baird, vice president, engineering, and Mark Krzewinski,
applications engineer, at Ellison Machinery Company, Pewaukee, WI,
to discuss two manufacturing cells that Ellison Machinery installed
at Harley-Davidson, Milwaukee, WI. These lines are used to produce
six different transmission gears. The first cell began operation
about two years ago. The second cell went on-line in February.
“Perhaps as long as 15 to 20 years, Harley-Davidson bought two gear
manufacturing lines from a company by the name of Suga,” Baird says.
“There were four machines in each line plus a series of elevators
and gravity chutes that delivered components to spindles for
machining. When these lines were installed, they were considered
state-of-the-art, and they ran successfully for years and years.”
These machines were described as “semi-CNC machines” by Baird, who
says they often were called peg-board machines. When they were
introduced, these machines were considered pretty far advanced
because a “programmer” could make them go through various iterations
and hit hydraulic stops based upon the peg-board assembly, which
determined how a machine was going to process workpieces.
“Elevator mechanisms were used to transfer parts onto gravity chutes
to send them from spindle to spindle. That was rarely done outside
of what I’d call automotive installations,” Baird explains.
“Automakers had been using that type of equipment for 50 to 60
years, maybe longer, but for a motorcycle builder with a
comparatively small volume, this represented a big leap forward.”
Of course, time marches on and a lack of flexibility and the need
for additional processes to make up for the lack of defined accuracy
those machines were capable of holding led to Harley’s decision to
replace them with modern, high-tolerance manufacturing cells, he
says. “If we had done a statistical analysis of that former
process,” Baird says, “it probably was capable of holding maybe
0.010_. With that type of in-process capability on their turning
equipment, they had to employ a lot of other machines that normally
wouldn’t be involved, or perhaps have engineers in the line to
remove extremely large amounts of stock to make up for the lack of
accuracy in upstream operations.”
A better way
The Okuma machines that Ellison installed can hold tolerance of at
least 0.001_. “Statistically, they’re good to half-a-thousandth,”
Baird says. “We’ve done demonstrations on that type of equipment
with in-process gaging where, if we’re running in a metric mode, we
can hold ±4 microns successfully.”
Downtime in the Suga machines and unavailability of repair parts
because the builder had gone out of business prompted Harley
engineers to look for a better production technique.
Before Harley-Davidson selects new production equipment, they form a
committee of maybe 10 or 12 people from various segments of their
engineering and manufacturing areas to look at different ways to
produce the parts they will be running on the new lines. “They have
their predictive change of engineering methods going on at the same
time so that group could swell to twice that size during the time
when they are trying to assess model changes or modifications for
the next five or six years,” Baird says.
“That’s a speculative group and their engineers probably would be
somewhat involved in that,” he adds. “However, their design people
would be the ones feeding information into them. That way, if they
buy a machine tool system, it has the flexibility to incorporate the
changes they would be considering. They do not want to be left in a
position where the equipment they bring in is only a five-year
solution. They want a 10- to 15-year solution for their production
problems rather than just a short-term solution for an immediate
problem.”
New lines
When Harley decided to upgrade its first transmission gear line, it
chose an Okuma twin-spindle CNC lathe with dual gantry loaders/unloaders.
“They are using the twin-spindle lathe to machine both sides of a
gear,” Krzewinski says. “They can do that in one operation, hands
free, using the gantry loaders.
“The right gantry arm will load the first spindle with a green
forging and the lathe will finish the front face, rough the bore,
and finish the bore,” Krzewinski says. “The gantry then will pull
this part out, load a fresh part, and take the half-finished part to
what we call a turn-around station. Next it will turn the part to
the back side, which hasn’t been machined.
“Now, the left gantry will take the half-finished part and place it
in the left side of the lathe. The lathe employs OD chucking on the
right side and ID chucking on the left station,” Krzewinski points
out. “Because the bore is finished, concentricity shouldn’t be a
problem. The first face is used for a datum plane. Next the lathe
will finish the last side and the diameter.”
After the lathe has finished the part, the loader picks the part off
the spindle and loads another part onto the left station. It then
takes the finished part over to the Edmunds Air Gage located at the
side of the lathe. Here it will orient the part into a horizontal
position, and lower the part onto the air gage. The air gage will
measure the part and send an RS-232 message back to the Okuma
control. This signal can be used to execute an offset change in the
tooling, if required.
Although every part is gaged, the system is programmed so that it
will accumulate measurements of three parts and then wait for four
or five parts before it sends back an offset. This avoids the
possible problem of hunting or cycling back and forth that could
occur if every part were used to correct the next one.
“The benefit of gaging every part,” Krzewinski says, “is that one
operator is running this line (it’s about 80 feet long). A single
operator is responsible for processing parts from green gear
forgings through the Okuma, on which he finishes the bore and rough
turns the diameter. Depending on which gear it is, it will be sent
through a broach, a profilating machine, a hobbing machine, and then
it’s ready for heat treating. Plus, there’s some other equipment as
well.
“The advantage of the gaging is the operator can walk away from the
Okuma area where the boring and turning takes place, for a long
period of time. There also is redundant tooling on that machine.
Like any other company, Harley-Davidson is trying to increase
production — I think they’re trying to almost double it. They’re
also trying to satisfy several other things, such as maintaining or
exceeding quality, lowering labor costs, and increasing production.
With the Edmunds gaging they can do it all.”
“Another requirement,” Krzewinski says, “is for fast changeover
between parts. Each one of those gear lines will do six different
gears. The requirement is a 10-minute changeover for the Okuma
equipment, which includes workholding, changing all the air plugs
for the Edmunds, and changing programs.”
“Sometimes there’s a minor tooling change necessary,” Baird says.
“Typically the changeovers are just programs. The new program tells
a tool to do something different. Even if there’s a workholding
constraint, they can change a set of jaws, call up a new program and
they’re producing a new part. Their old system really was
inflexible. They could have been looking at an hour or two
changeover on their previous line and still not have the capability
the new one affords them.”
Cell development
Ellison Machinery worked closely with Harley-Davidson in selecting
equipment and marrying the various cell components. “We integrated
the Edmunds gaging into the cell,” Krzewinski tells QM. “We
specified on the order to Okuma that we were going to use gaging.
There is an option called RS-232 post-process gaging, which simply
is just software and a port. Setting that up is fairly easy. I
worked with Edmunds in Connecticut. I like dealing with them
directly.”
“We commonly provide a machining system with integrated gaging,”
Baird says. “These gages are expensive but they work well and they
have a very good durability record. I’d say we have four or six
projects per year where we’ll furnish Edmunds as the gage of choice.
We also work with other gage companies. We’re not exclusive with any
of them.
“Actually,” Baird continues, “if we were picking gaging equipment
ourselves and weren’t being driven by the customer, we would lean
toward Edmunds because we have in-house knowledge of how to deal
with problems that may arise. Usually there will be a fair amount of
problems — any time you have two computers communicating with each
other and their operating systems are alien to each other, you have
to exercise your wits to get them communicating correctly.”
“All of the components of the system were shipped to us here and we
assembled the system for testing and runoff,” Krzewinski says.
“We’ll have a statistical runoff here, which is spec’d out in the
purchase order. Usually this is a four hour run, but sometimes the
purchaser will specify an eight hour run — with no stoppages.”
Once the runoff is completed, Ellison Machinery tears down the
machine, packages it, and brings it to the customer’s plant. There
it is reassembled and another runoff takes place.
Training
According to Krzewinski, training can be handled in a number of
ways. “With this system,” he says, “we contracted with Edmunds for
two days of training. Harley opted to wait until the machine was on
its floor and then they brought somebody directly in from Edmunds to
do the training.
“Being the integrator, and taking this system from cradle to grave,
I was pretty familiar with the gage and did some fairly extensive
training on it as well. The in-depth training where they really want
to get deeper into the gage, is where Edmunds can provide the needed
expertise.
“The lathe, the gantries, and the gage really are a whole system so
training has to deal with this equipment as a system,” Krzewinski
says. “Operators have to be familiar with the inbound operation
where the process starts, with programming the gantry, mechanical
training such as changing over workholding devices, changing over
programs, setting up the gage, and qualifying the gage.
“Maybe a month or so after runoff in the customer’s plant, we go in
for some follow-up training for a day or two,” Krzewinski says. “As
with any new system, we can’t possibly answer all questions until
it’s in full production. What I typically tell the operators to do
is to make a list. Being local as we are, I can easily come back in
and work through some of their problems.
“With the complexity of automation and machine tools, we need to do
that,” Krzewinski said. “We need to follow up with customers, not
drop ship a machine and hope that they understand it. Our customers
appreciate that.”
|
This twin-spindle Okuma lathe is
used to machine both sides of a gear blank at
Harley-Davidson. Gantries are used to load and
unload parts. An Edmunds Air Gage is used to measure
every part and send corrections back to the machine
control to compensate for any deviation from
tolerances.
This photo shows the second lathe operation. There is a raw
forging in the spindle and it is ready for machining.
After the part is machined in the second spindle the gantry moves
over to the left of the work cell and is shown in the process of
putting the finished part onto the Edmunds Air Gage for checking the
bore.
Transmission gears are shown in four stages of completion: green
forging (upper right); after the first side and bore have been
finished (lower right); after machining has been completed (lower
left); and the finished gear after hobbing, broaching, heat
treating, etc.
An overview of an Okuma lathe with dual gantry system and Edmunds
Air Gage producing transmission gears at Harley-Davidson.
|
