This week we decided we should have a look at another case study. In this case a European Cell Phone Manufacturer/Assembler with manufacturing plants in China.
This Cell Phone manufacturer assembled cell phones for one of a leading global cell phone brands. In this case production was already very high however the client was struggling to meet demand and had undertaken a Lean Sigma program to improve performance.
The manufacturer had several production lines and one in particular had a lower rate than the others. The initial project objective was to increase throughput rates from 451 units per hour to 527 units per hour. There was plenty of production data available.
Measure the current Process
Initially we started by mapping out the production process and using the process map and ‘Gemba’ walks (‘Gemba’ is Japanese for place where the work happens) to identify the operational step which was the process bottleneck.
In this case we identified the final testing operation as the bottleneck. Inventory was building up in front of the testing station, providing clear indication this was the bottleneck. Once the bottleneck was identified we used multi level pareto charts to drill down and help us focus on the critical areas. Multilevel pareto charts are a key tool we use to help the project focus.
The high level pareto chart showed throughput was affected by four areas. The top three areas in order of importance were:
- Potential Tester Operator Improvements
- First pass quality rates
A further Pareto chart was developed for each of the three areas that resulted in 8 further areas for further investigation including:
- Phone Microphone Design and Fit.
- Poorly Performing Buzzer QA test -high delay times.
- Computer Handshaking between the Testing operation and the production line.
- Work stoppages by the Tester.
- Testing Machine maintenance.
- Accuracy of placement of the Testing Pallet.
- Missed Printing Codes.
A root cause analysis was carried out to identify potential root causes for the low throughput rates. Once the potential root causes were identified statistical hypothesis testing was carried out to determine if the potential root causes were in fact true root causes.
The following solutions were implemented:
- The Testing Operation was automated and the software was updated to the latest version which resulted in lower cycle times.
- Testing Operation handshaking method was implemented that resulted in lower cycle times.
- New Printed Circuit Board was used resulted in higher first pass yields.
- New Microphone type was implemented which resulted in higher first pass yields. The new microphones not only increased average performance but had reduced variation.
- Missed Printing codes were rectified.
Throughput increased from 451 units per hour to 551 units per hour resulting in $270K in benefits per annum for the one line.
The results were transferred across the other four production lines which resulted in $860K in total benefits.