We’re available to help you plan, setup, and run Highly Accelerated Life Testing (HALT) on your products. We get involved early in the concept and design process to ensure adequate testability is designed into the product, prior to HALT being conducted.
We’ll provide a comprehensive test report, including recommendations based on our experience, and will be available to present this report to Management if required.
We operate a Qualmark Typhoon 2.5 located in Christchurch, New Zealand, and provide testing services for some of New Zealand and Australia’s most well-known brands. Contact us for pricing and availability.
Let’s make it robust
Highly Accelerated Life Testing (HALT) is specifically designed to improve product reliability, it is one of the most effective and efficient ways of characterising product response to stress. In HALT we are attempting to stimulate failures through the application of stress, the points of failure indicate an opportunity to improve product reliability.
The first challenge is identifying the failure mode, next comes repeatability and the last is debugging and corrective action – then repeat.
What, how, and when?
For a better understanding of what the process of Highly Accelerated Life Testing entails take a look at our library here. We conduct HALT to identify weaknesses in our product to determine the most likely failures in the field. By doing so we get to understand what fails and how, we also get to reliably replicate the failure mode which gives us the ability to fix it.
When to conduct HALT often comes down to when samples of sufficient quality are available with high enough test coverage to justify running stress tests. Typically we recommend looking at a two-stage process:
- Run HALT at a sub-assembly level – improve these first.
- Run HALT at a system level
Each step will pick up different failure modes, with the system level HALT being the more critical of the two.
Let’s take a look at a couple of relevant distribution illustrations, first off if we assume that a population of identical products (in process, BOM, and design) are manufactured and tested for a particular failure mode exposed when the product is stressed at high temperature we would find the following (ignoring for now the shape of the distribution):
If we then analysed our field environment for each of the population of units we’d equally find that the level of stress encountered by each product will vary, most likely fitting to a distribution of it’s own:
So what happens when field stress is greater than product strength? This happens – stress overcomes product strength, i.e. the weak products will fail.
Take that same scenario but fast forward in time, what has happened to product strength? It’s degraded over time, meaning more products are now under threat of early failure:
How does HALT help?
When we conduct HALT we are attempting to increase the operating margins of the product. Shown in graphical form below the operating margins (or limits) are the limits at which the product fails. In a product that is not robust it is reasonably foreseeable that the following situation may occur:
In other words – our margins are not sufficient enough to protect the weak from our predicted field stress levels.
By conducting HALT and finding these weaknesses (dominant failure modes) we are able to drive them to root cause, correct them and verify we have increased our margins. After HALT we’d hope to see something like this:
How far do we need to push? That’s a good question, and one that needs some thought.
What else can we use HALT for?
HALT as described here is practiced using a HALT chamber, these chambers provide wide thermal and vibration capabilities giving is the ability to stress to failure. A HALT chamber is a fantastic tool, it can be used for:
- HALT & HASS Testing
- Field failure diagnosis (NFF – see footnote)
- Regression testing
- Quality control
- Alternate part, or design qualification
The possibilities are limited only by your thought boundaries. Any functional test you currently perform on your products can be performed whilst the product is under duress – more to the point, it should be.