The SMART Group PPM Monitoring Project Workshop, which was held on the 22nd January at The National Physical Laboratory, was another great success with over 30 engineers attending to discuss manufacturing assembly performance. The workshop was FREE to members of the SMART Group and PPM project members.

The PPM Monitoring Project was set-up by the SMART Group with support from the DTI to provide a benchmark for small and medium volume companies to compare their manufacturing yields and has now been running since May 2002. The project monitors the average assembly yields for screen printing, component placement, reflow and wave soldering. From the start of the project over 20 companies were visited and 18 companies regularly contribute to the industry average yields featured on the project web site www.ppm-monitoring.com. These yields are also displayed to other linked sites and can be featured on your own site with a linked result service provided free.

During the day Bob Willis thanked all the companies who had contributed support for the project, he also presented two special awards to companies who had contributed to the project since its inception and had provided the results every month on time. A glass trophy and plaque were presented to Dolby Laboratories and Primagraphics.

The workshop's aim was to discuss the project, the benefits gained so far and to suggest future improvements to the defect data gathering. There was also discussion on the methods of future support and the ways in which the results cab be best disseminated in the future. After much discussion the project will now be including the collection of defect types and the components most affected at each of the manufacturing stages, which currently include screen print, placement, reflow and wave soldering. This will help engineers improve design rules for future products based on the published assembly yields. All companies represented agreed that the project was very worthwhile, a benefit to the industry and a credit to the SMART Group.

In addition to the PPM discussions Bob Willis and Alan Brewin from NPL discussed different ways in which solderability had a direct impact on PPM yields.

Solderability Testing and Soldering Defect Levels - Alan Brewin
The advantage of solderability testing as a QA tool was highlighted by the link to resulting soldering defect PPM levels on products. The techniques were described and examples illustrated on how discrimination between good and bad levels of solderability on incoming parts are achieved. Wetting force determination using a wetting balance provides a measurement of the solderability of component leads or board finish.

NPL is currently undertaking work to improve the test method by better modelling production practice into the measurements. New test fluxes and sample pre-heating are approaches that are under evaluation. There is opportunity for UK companies to benefit from testing on their parts and materials within the project to link the component types or finishes which are shown to give the highest defective PPM levels.

Production Solderability Testing Made Simple - Bob Willis
Although there is no substitute for solderability testing using defined measurement and documented standards production testing is often necessary. If parts fail using a simple production test, parts can be re-tested in line with current standards prior to rejection to suppliers. In the real world of production, do parts solder or do they not?

Bob Willis illustrated a simple test method that anyone could use and hopefully can be calibrated in the future with standardised techniques and help engineers understand the real cause of many soldering defects. Examples and case studies of Bob's work were shown along with some of his unique video clips, some of which are shown in the SMART Group's members' area.

To get a copy of Alan Brewin's paper click here. For Bob Willis paper click here.

If you want to keep up to date with the SMART Group PPM Monitoring Project monthly results or participate in the project visit www.ppm-monitoring.com or email info@ppm-monitoring.com

The PPM Monitoring Project is unique as there is no other collaboration of companies producing this type of data and providing it to the industry FREE.

Some of the issues discussed are summarised here and are produced from notes made by Peter Grundy, Siemens, who was volunteered by Bob to be the scribe.

Some of the changes will be incorporated over the coming months and fully documented on the PPM website.

CHANGES DISCUSSED
It was requested that changes to processes should be included in any monitoring so that, for example, if an aperture in a stencil is changed because of insufficient bridging, the subsequent performance of the joint is monitored for success or failure.

If month 1 is all level 3 but month 2 is all level 1, how does this affect the results? Can we track changes in levels to go with defect types?

It was requested that the mean value of faults per month be recorded rather than the average to avoid distortions.

It was suggested that any results for lead-free statistics be separated out into a new project. This may need funding from other projects.

It was suggested that defects be categorised against set-up criteria, e.g. defect vs stencil type/thickness or defect vs component size (0201 vs 1206)

If we expand the criteria for data collection, will it put off participants from contributing as regularly as they do now?

Chris Hunt offered a CD with a catalogue of NPL's defects as collected by Martin Wickham for free distribution to Group members. All agreed this is a good idea.

In order to expand the data set and get more useful results it was agreed that defects should also be collated against component types. This will automatically add number of terminations and pitch into the equation. If we use this terminology will it replace the levels already used? It was decided to run both in parallel and see if there is too much data to collect sensibly or whether we can abandon levels in future.

It was suggested that we could focus on one type for a month and analyse it in greater detail so that the most accurate pareto of its propensity for faults can be established. The Defect of the Month was found to be useful.

The current spreadsheet recording method was widely accepted and it was requested that all defects be logged against all component types.

If we offer to expand into Europe or beyond with the survey, we should add a country code to establish if some faults are regional.

Wave soldering unchanged. Proposal to add a level 4 which is the same as level 3 but including array packages and difficult technology items such as 0201

Those contributors who can evaluate all joints on array packages and give good defect feedback will use level 4. Those placing arrays or 0201s but cannot provide quantitative data should remain at level 3.

As expanded data becomes available, spreadsheets will be used on the website to analyse and display results for all defects and aspects of production.

Add the appropriate industry sector to each member in case some faults are indigenous to a particular sector.

Once component type vs defect data is gathered, prepare and issue top 5 hit list of problems and include industry sector.