The Value of Analysis (Part I)

I spent two years at Visteon Corporation (Visteon is to Ford Motor Company as Delphi is to General Motors) working on material cost reductions (and, as ancillary projects, labor & scrap reductions, as well as the occasional foray into floor-level problem solving).  One of the things we did was hold “Value Analysis” sessions in which we’d sit one of our climate control air handling systems on a table and have people wander in and out informally and scribble cost-reduction ideas on forms that our team would then take up as possible projects.

Note: I’m going from memory here as I – obviously – was not permitted to take any work-related materials during my one-day transition from employee to “alumni”.  (How’s that for spinning being laid off?)

Having learned about – and fallen in love with – Value Analysis/Value Engineering while studying for my Masters in Mechanical Engineer at Carnegie Mellon University, I understood that these sessions could be done much, much better.  In a bid to take over organizing and facilitating these sessions, I supported my attempt by actual demonstration of how I thought it should be done.

The first was to select a sample product; I picked the Ranger/Explorer platform.  I went through every single part, dividing them into purchased parts and in-house parts (all injection molded pieces but one; the exception, a heat exchanger was purchased from another Visteon plant).  Now that I had two groups, I did a Pareto chart of the costs; purchased parts were easy, the in-house parts required a little digging into material, labor, etc… but both were doable.

In the case of purchased parts, the results were predictable: two pieces of the 20-odd purchased parts represented over 80% of the purchased part costs.  In this case, the air blower motor and the “squirrel cage” fan that attached to the motor.  Logically, it made sense to focus on those, and part of my proposal was to hold these pieces out for special attention.  There was also another dimension: parts whose common functions across platforms opened up standardization possibilities.

Please indulge me as I go through a somewhat lengthy exercise.  This is not only to highlight my knowledge of the technique, but also to – I hope – pass on knowledge to help others.  Below each description will be a listing of things I proposed along with the benefits I identified… note, I was in the process of doing this when I was downsized in 2001, so I never did take over the sessions, nor do I know if any of my recommendations were taken.

Purchased parts

Electric motors:  Every product had its own unique motor, with its own unique flange.  Not only was this duplicating functionality into unnecessary variations, the unique mounting schemes caused enormous headaches.

1. Standardize to three motor sizes, small, medium, and large for coverage of all vehicle sizes.
   1. Economies of scale in cost due to fewer variants
   2. Fewer SKUs, lower inventories
   3. Less guesswork in the design world – three proven off-the-shelf possibilities to use

1. Standardize flange mounting features
   1. Common design removes one whole area of variation
   2. Cross-line commonality enables better employee mobility
   3. Common gasket across all lines; economies of scale, one SKU not many
   4. Change gasket design to allow larger tolerance on positioning (assembly time)
   5. Add positioning features in molded case to speed placement and alignment of motors
   6. Add snap-fits to reduce number of fasteners in motor installation (e.g., from 4 to 2)

 

Squirrel-cage fan

1. Standardize material to one material rather than the 3-4 used
2. Standardize to three sizes, small, medium, and large (economies of scale, SKU reduction)
3. Redesign interface to motor to allow for multiple correct alignments
   1. Current design only had one allowable position on motor shaft (flat on shaft to flat on molded female opening); adding second flat would reduce positioning labor time as workers had to spin the fan to get it to go onto the shaft

 

Fasteners

I can’t claim credit for this one: one of our intern projects was to do a board of all the screws and other fasteners we used.  It was… illuminating.  For the same geometry screw, for example, we had an assortment of – IIRC – 5-6 different coatings.  The same for washers and nuts.  Standardize on two: black and silver appearance.  Economies of scale, and fewer SKUs.

 

Cams

Many of the climate control doors directing air flow inside the case were controlled by cam mechanisms that connected the door (inside the case) to the actuator (outside the case).

1. Future designs: Try to reuse existing cams instead of designing new ones each time.
   1. Fewer SKUs
   2. Economies of scale
   3. Standardize to one material
   4. In the case of electric motor actuation, try to design the actuator (custom design) to interface directly with the door rather than requiring an intermediate cam component – space permitting.

 

Air gaskets

Many of these air gaskets were die-cut foam pieces.  Standardize materials and thicknesses; ideally, standardize sizes (requiring standardized openings) for economies of scale and reduced SKUs.

 

Firewall seals

Every climate control handling system has to have a large molded seal to allow the coolant tubes to pass into and out of the engine compartment.  Geometry is in a flux as the sheet metal separating the engine compartment from the passenger cabin – and the openings provided for the tubes – was not under our control and varied widely.  Still, there was opportunity for standardization and cost reduction on this material (which formed the core of a material substitution project I did saving over $250K a year).

 

Actuators

Climate control air handling systems had pneumatic actuators, vacuum actuators, pulley and cable systems, and electric motor actuators.  Phew.  Pick one.  Standardize not only on the type of actuator but select 2-3 and design around them to allow commonality across positions in the case and across vehicle platforms.

In-house parts

Injection molded pieces

Visteon had actually done a good job in standardizing the material; in fact, in a compromise between US and European operations, a 33% talc-filled polypropylene had been agreed upon to simplify things on a global scale.  There were a few exceptions, however, that IMHO some design work could have been done to move these few non-standard-material parts to the common material.

I had two observations on molded parts:

1. Tool change time.  I had, at the time, seen videos of a mold change where – from last part to first new part – the mold change took minutes.  Think SMED from Lean Manufacturing.  It was… awesome.  Both Ford/Visteon plants I’d worked in, it took hours, and usually an entire shift, to change a mold out.
2. Process optimization. I had seen, firsthand, how mold cycle times could be optimized.  One example: An engineer I know (who has forgotten more about injection molding than most people know) optimized a cycle time from 90-odd seconds to 50-odd seconds, still producing good parts.  This was an amazing cycle time improvement.  And yet, the next morning, he went to check the machine and it was back to the slower time.  Why?  Because the packer at the end of the line didn’t want to work that fast, so he had his mold tech buddy reset the cycle time.  Multiply this across the 100+ machines in the plant – this cost the company millions in lost productivity.  Not only that, but between SMED and optimized cycle times we could have pulled multiple molds back from suppliers, capturing that part of the dollar stream instead of paying someone else.

 

Evaporator core:

We purchased myriad variants of evaporator cores from a sister plant.  Almost every platform had its own tweaked core design.  Standardize, standardize, standardize.  Again, have 3-4 sizes as stock parts from which to choose.

Core lessons:

1. Standardize materials wherever possible
2. Standardize designs, especially at interfaces
3. Mass-customize designs by having an array of stock modules (e.g., fan, motor, actuator, and evaporator core sizes)
4. Drive designers to re-use existing modular components rather than customizing each time
5. Optimize processes (e.g., molding operations) for both productivity gains as well as pulling in parts from outside to capture that value
6. Formalize analysis sessions and have them facilitated by someone who knows the techniques rather than having be people wandering in and out.

In Part II I’ll discuss some other things I proposed doing in these sessions.

© 2013, David Hunt, PE