HALT Versus ALT
Chris and Adam discussing HALT and ALT. What are these? HALT stands for Highly Accelerated Life Testing. ALT stands for Accelerated Life Testing. They sound very similar. But they are not. HALT is a destructive test regime. In fact, a good HALT plan will involve that product failing many times. This is done by subjecting the product to stresses (vibration, thermal cycling et cetera) well beyond actual operating stresses. Some of the failures this creates will not be relevant. That is, they will simply never occur when the product is used ‘normally.’ But many failures are relevant. And by undertaking HALT, we now have a good idea of which failure mechanisms and modes are likely to occur when it is used normally. And this information is incredibly valuable to a design team. ALT on the other hand starts with a failure mechanism you know about. And in a short period of time, you can predict how long that failure mechanism will cause your product to fail when used normally.
Still confused? Well listen to this podcast.
Join Chris and Adam as they discuss HALT and ALT. HALT stands for Highly Accelerated Life Testing. ALT stands for Accelerated Life Testing. So no big difference, right? Wrong.
HALT involves subjecting a product to stresses that are well beyond those likely to be experienced during ‘normal’ operation. This will sometimes cause the product to fail in ways that simply aren’t possible when used correctly. But many of the failures that are observed when conducting HALT will eventually occur in ‘normal’ operations. So when a product ‘fails’ during HALT, we (where possible) simply patch it up, and keep testing it. Because the aim is to create a list of the likely or dominant failure mechanisms and modes that can then be forwarded to the design team. And as a designer, how useful would it be for someone to give you a list of the (likely) weakest points in your design?
ALT is entirely different. Say that you know that you know that fatigue that causes one of the actuators in a relay to fail open is going to be a problem in your product (… perhaps using HALT to work this out). This means you can use a physics of failure (PoF) approach to model the corrosion, understand what factors are in play (likely temperature and humidity), and work out some relationship between these factors and time to failure. For chemical reactions, we often use the Arrhenius model. This means you can work out how much faster your component degrades when exposed to a higher temperature and humidity when compared to ‘normal’ operating conditions. This is called an acceleration factor (AF). An AF of 10 means that your ALT will make your product fail 10 times faster. If your product fails after 1 hours of ALT, it will likely fail after 10 hours of ‘normal’ use. If we increase our AF to a much higher value, we might be able to replicate an entire lifetimes use in a matter of weeks.
HALT gives you a list of likely dominant failure mechanisms. ALT helps you predict when one of these failure mechanisms will cause failure.
As with anything to do with reliability, things aren’t as straight forward as they seem. So Chris and Adam talk about:
Apollo program … rocket motor … exhaust cone … black powder … see how it broke … dampened the vibration.
- How some organizations don’t get the value of either activity, saying things like ‘of course it failed, you exposed to ridiculously high stresses.’
- How HALT can help you protect your product from ‘abnormal’ usage such as the user dropping your product
- That no matter what you do, either activity is useless if it doesn’t inform a design decision to make that product more reliable.
- How HALT has been successfully used on so many products from all eras. For example, HALT was used in the Apollo program to design a better rocket motor exhaust cone. Engineers actually destroyed the cone with black powder to see how it broke, producing a much more robust product as result.
Enjoy an episode of Speaking of Reliability. Where you can join friends as they discuss reliability topics. Join us as we discuss topics ranging from design for reliability techniques, to field data analysis approaches.