The Donner Company

The Donner Company is experiencing production and quality issues, and as a result they are missing deadlines and losing customers. Their process, from order generation to shipping (see below), has some "low hanging fruit" that could help their productivity immediately. The following assumes 8 hour work days and that September is a typical month.

First, Bruce Altmeyer, the design engineer, spends about 10 hours a week investigating unusual production problems, and about 50 hours/wk programming CNC machines. This leaves him with 100 hours to inspect, estimate, and bid customer orders, and write material specs and factory orders once the bid has been accepted. For every order, writing material specs and factory orders takes about a day and a half (4 days from customer acceptance to an order reaching the production supervisor estimated as 1.5 days with Altmeyer, 1.5 days for the purchaser, 1 day for entering into log, factory order, and blueprint to supervisor). If Donner hired a fulltime programmer (more justification later), Altmeyer would gain 50 hours a week to devote to dealing with customer orders, possibly reducing cycle time on writing factory orders and material specs by a half day (100 free hours ? 150, 1.5 days ? 1 day).

The purchaser takes one to two days to locate and buy low priced materials once he or she receives material specifications from Altmeyer. Four materials, A, D, G, and M, make up 80% of the orders (Table 1). Since Donner uses 30 base materials, this is 13.3% of the possible materials that account for a larger percentage of the orders. If Donner were to stock these four materials (15 sq ft worth of space), they could gain a day and a half, and possible secure better prices by buying in bulk. Storing the top 6 materials used (20% of base materials, consisting of A, C, D, E, G, M), Donner would have almost 93% of orders covered. The production supervisor, David Flaherty, waits until the material arrives to start planning production, so having material on hand will save several days here too.

Material Boards % of Total

A 1509.85 26.21%

B 69.85 1.21%

C 357.6 6.21%

D 1250.75 21.71%

E 417.75 7.25%

F 160.1 2.78%

G 814 14.13%

H 9.1 0.16%

I 0.00%

J 108 1.87%

K 64 1.11%

L 0.00%

M 1000 17.36%

Total 5761

Table 1. Number of boards produced for each material type in September

Currently Donner uses the new CNC machine to drill orders over 100, which is 20% of the orders in September. Looking at the breakeven point for manual drilling vs. CNC, Donner should be using the new machine for orders over 5 (see graph 1). Orders over 5 account for 98.9% of boards and 58% of orders, so there will be some time savings, as well as quality improvements (many of the quality issues were missed holes, CNC is less likely to miss a hole than workers are). For the 1.1% of boards (42% of orders) that go will be manually drilled, there is a cost savings from only using one drill instead of 7, and 5 boards do not take up more than one typical panel, multiple drills could not be used anyway.

The CNC router should be used for orders over 200 (see graph 2). This will account for 75.5% of boards, but only 11.7% of orders. The rest of the orders will continue to be cut on the punch press.

The changes in CNC machine utilization will require 157.5 programming hours (240min/order * 35orders ÷ 60 mins/hour for the drill; 150min/order * 7orders ÷ 60min/hour), which will keep a full time programmer busy all month.

Simply changes along the production lines, such as moving the desk closer to the plating tanks could result in some process improvements. It is hard to quantify these types of changes without seeing a layout and getting more cycle time information. More large scale layout changes might also help the process, but, again, without seeing the current plant layout it is hard to make suggestions.

Adding new equipment would also help the process. Table 2 shows how many panels are possible per day through each process. Adding new equipment for the electroplate, soldermask, and solder dip area, adding another punch press, and adding another person to inspect, test, and ship, will help make the cycle time/ panel more even, which would allow Donner to implement a pull system. These are the apparent bottlenecks, though if one person was operating the DFPR area and moved one panel from step to step instead of having one panel in each step simultaneously, this would also be a bottleneck (10.13 panels/day) since it would be the second slowest possible step, but the slowest required step.

Setup Run Panels/day

Prep

Artwork Generation 29 0 16.5517241

Inspect and Shear 20 0.5 23.4146341

Punch Tooling Holes 10 0.5 45.7142857

Image Transfer

Drill

-Manual drill 15 320 1.43283582

-CNC drill 16 30

Metalization 10 0.75 44.6511628