Taking the HASSle out of reliability

In a recent blog I explained how Highly Accelerated Life Testing can improve a product’s design and make it inherently more reliable. This time I’d like to look at it’s production-line partner, Highly Accelerated Stress Screening or ‘HASS’ (also invented by Gregg Hobbs) which improves reliability by eliminating manufacturing defects.

As with HALT, HASS works on the basis that high levels of stress applied to a product for a short time will result in exactly the same failures as low levels of stress applied for a long period of time. In this case it’s a production test not a design test, it is there to turn latent (potential, hidden) manufacturing defects into patent (realised, visible) faults that can be repaired prior to shipment. This is, of course, slightly different to free stuff and methods you can research on the web, but they all have the same aim and outcome – to maintain product life and ensure it does not die in its infancy.

By precipitating or highly accelerating these faults the manufacturer can find and fix them rather than the customer experience them after several weeks or months with the accompanying cost, loss of reputation and commercial damage.

If we think of the typical bathtub reliability curve – see the diagram – you are trying to eliminate the infant mortality part of it, i.e. the early life fails (shown in red).

You can’t do HASS testing without having done HALT first as it relies on specific stress figures that are determined during the HALT process. These stress figures have different scaling factors applied to them and are used to set the initial HASS parameters.

As with HALT a conventional environmental chamber is unsuitable, a special HALT/HASS test chamber must be used. Again as with HALT, the unit under test needs to be set up in a continuously monitored test mode so than any failures, however momentary, are immediately visible.

Because HASS is deliberately trying to ‘use up’ a small proportion of the product’s lifetime, i.e. removing the infant mortality part of the bathtub curve, we need to make sure we aren’t over-doing it and using up most of its lifetime. Before HASS is introduced into production a Proof of Screen test is done, which involves running a unit through the HASS test twenty times in succession. If it hasn’t failed after twenty HASS tests you know that one test doesn’t used up more than 5% of its total product lifetime.

Unlike HALT the test is quick; typically between 30 and 90 minutes of combined rapid thermal and physical shock cycles. This is a lot shorter than most soak or burn-in tests, yet is much more effective at finding faults. The duration, and sometimes the stress intensity, can be adjusted over the course of several months depending on when failures occur within the test period.

HASS can be expensive. It uses up a costly HALT/HASS chamber and skilled staff for a perhaps an hour or so. Because of this it’s not for everyone; far more people use HALT than use HASS, and it tends to be restricted to particularly high reliability or mission-critical products rather than every product in your catalogue.

To help reduce cost, a variant of HASS known as Highly Accelerated Stress Audit (‘HASA’) has been developed. This works by only testing a small proportion of the manufactured units on a sampling basis, but if a specific number of faults per batch is exceeded the whole batch is then 100% tested. (I’ll talk about sample testing and Acceptable Quality Levels in another blog).

However, if your products are of high value and/or complexity, and if you are looking to provide maximum reliability to your customers, then HASS (or HASA) is well worth investigating.

But please be aware, just like HALT, HASS needs to be done ‘consistently, completely and correctly’ or it can cause more problems than it solves; it’s more complex than I have space for in a blog like this, so do read up about it and make sure you get some good advice.