Issue link: https://iconnect007.uberflip.com/i/340751
82 The PCB Magazine • July 2014 WHY REMOvING YOUR BOTTLENECK IS A BAD IDEA continues cycle time, yet they lose sight of the fact that the reduced amount of WIP does not provide enough buffer for the variability in their pro- cesses. The result is that the bottleneck starves frequently and total demand can't be met. So any one order can be completed very quickly yet there is not enough volume in output to sat- isfy all demand. The unintended consequence of this is that mangers start expediting most or- ders depending on the latest, hottest fire. This in turn: • Increases product mix, which • Accelerates the shifting of bottlenecks, which • Makes production management much more difficult and • Increases the variability in cycle time, which • Makes customer service worse, which • Leads to more expediting and the downward merciless cycle of ruination Figure 3 shows an unbalanced scenario from Factory Physics called Penny Fab II. Here we have one station taking two hours to punch, two sta- tions taking five hours per machine to stamp, six stations taking 10 hours per machine to rim, and two stations taking three hours to deburr. I am not going to provide a detailed explanation on the fact that the total process time is 20 hours, or that the bottleneck rate is at station two and is 0.4 parts/hr, and the critical WIP is only eight parts in this system. It should be obvious that this isn't a balanced production operation. Figure 4 shows the result of the throughput and cycle time as compared to the marginal case (the case where it would be single machines at each station, all operating at the same speed, with no passing allowed: the ideal case for level Figure 4: the superior performance gained by having too much capacity with a substantially improved throughput and cycle time over the marginal case. in the real world you have to decide if this extra performance is worth the investment in the additional capacity.