Standards – A Double Edged Sword
Recent discussions on twitter led me to pen this blog post. It seems there is no deference for standards. It is better to do this alone than to rely upon standards. Standards are the bane of product development existence. I believe when it comes to product testing standards can be useful. I would not be willing to throw the baby out with the bath water making assumptions that the standards are bloated and untested. First of all, standards do not necessarily “solve” our problems any more than standards “create problems”. I would agree with some of the twitter arguments that our talented staff should think these things through rather than summarily dismiss or embrace standards.
Consider in the automotive world, specifically heavy trucks. A Class 8 truck is not like a car. These vehicles travel in one year the distance some cars travel in a lifetime. The vehicle is an instrument of profit and not an inconvenience when it is down. Additionally, the amount of variation in feature content and customer applications is considerable compared to the typical automotive applications. The Society of Automotive Engineers (SAE) is a body that provides standards for developing automotive products. Within that organization, as an example, there is a standard Recommended Environmental Practices for Electronic Equipment Design in Heavy-Duty Vehicle Applications also known as J1455. This standard identifies the electrical and physical stimuli to which an Electronic Control Unit (ECU) can expect to be subjected when mounted on a heavy vehicle. An electronic control unit is usually an embedded subsystem component such as Anti-locking Brake System (ABS) or an Instrument Cluster. For example, the vibration profile is illustrated as Power Spectral Density (PSD) expected considering the mounting location (dash, frame rails, front or rear etc.). The standard also identifies temperatures that can be expected by mounting location of the ECU on the vehicle. This standard, I know specifically, is the result of a collaboration sponsored by SAE with the Original Equipment Manufacturers (OEM). In other words, the OEM’s have studied some portions individually, taken some measurements, the results of which end up in this standard.
To decide that a standard is not applicable, we must know something about the standard, not summarily dismiss as bloated or trivial arbitrarily. In the case of our ECU environmental standard above, perhaps we need to know specifically what is not applicable. We should also consider what is required for us to understand the true implications (if we think the standard a bunch of puffery). For example, if we believe the standard does not reflect what happens it is incumbent of us to:
- Identify a variety of applications (in our vehicle example, heavy haul, construction, long haul etc.)
- instrument those vehicles
- ship them to a variety of climes (desert, arctic etc.) for testing
- take measurements
- distill the results into something manageable and tangible (actionable) on our part
We will need to do this over a period of time to ensure the validity of the range of stimuli to which the product will be subjected. Meaning this is not a one sample test. This is a nontrivial activity and will cost us in terms of time, talent and money, and perhaps it is a better solution than relying upon the standard.
Contrast this to doing what you think may be happening in the field. Guessing the environmental exposure to the product will require many assumptions to be valid.
Our making the measurements up or arbitrarily assigning values is of little help. Dismissing the standards as bloated without knowing is not productive. Standards are not evil, no more than ad hoc and on the fly are. In my view, standards can help us if taken for what they are, a homogeneous representation of a nonhomogeneous environment. It is up to our talent to make some sense of the standard, know the source of the standard where possible, and understand the implications of real life deviations from the standard. Testing does not end at testing to standards at any rate, as that will likely only tell you where the product passes. We are more interested in where, when and how the product fails. That requires more than a standard.
We will write more on standards in future articles.