How Reliable is Reliable? Our Premium Offerings Come With Hard Data

Trial by Fire: How we Test

In order to simulate damaging events for testing, we work with external reliability testing partners to subject our hardware to standardized conditions and to achieve reliability certifications. Not all tests apply to all modules, but they broadly fall into a few categories: vibration testing, mechanical shock testing, and thermal / thermal shock testing. These environmental factors – impacts, vibrations, and fluctuating temperatures – are chief among the stressors that electronic components face in the real world. 

As mentioned, there are standards that help a manufacturer be sure their tests represent those real world challenges. JEDEC is a global leader in providing standards for manufacturers in  microelectronics, and they are the authors and originators of the relevant standards here. We test to the following:

• Vibration, Variable Frequency JESD22-B103B: Testing to this standard involves subjecting module samples to destructive levels of continuous vibration. A sine wave-shaped vibration pattern subjects the samples to vibration that displaces them from 20 to 80 Hz across 1.52 mm, and accelerates from 80 to 2000 Hz with 20g of force. All of this occurs across four minutes, four times per 3D axis for a total of 48 vibration cycles. This simulates extreme vibrations with many different characteristics, looking for any scenario where a module could fail. 
• Mechanical Shock JESD22-B104C: Here, samples are subjected to a half-sine pattern of mechanical displacement, simulating impact followed by a hard stop. This is performed with an acceleration force of 1500g, or about 3.3 pounds, 30 times and in all six directions (from the bottom, top, left, right, front, and rear). 
• Temperature Cycling JESD22-A104E: This test simulates dramatic temperature shifts over very short intervals of time, which have the potential to deform electronics and their internal electrical connections. Over a window of 30 seconds, test hardware is subjected to fluctuations between -40˚C to +85˚C (-40˚F to +185˚F) and back again. This shock could damage many electronics over even a few repetitions, but this is repeated over and over for 350 cycles. 

These are the primary tests we conduct with our modules. There are others, such as drop testing, which involve literally dropping hardware and then inspecting for damage. But the above are commonly tested across our modules, and are representative of the likely wear conditions for microelectronics in challenging fields. 

Information on the reliability testing we conduct for each module can be found in the module’s datasheet, in the chapter titled “Reliability Tests.” The following is an example for our BL654 Bluetooth LE module: 

Mean Time Between Failure: Reliability Quantified

One additional way of assigning a number to reliability is something known as Mean Time Between Failure, or MTBF. This metric is a calculation of the number of operational hours of a device divided by the number of failures observed. This results in a value, in hours, of the average time that hardware can be expected to operate without failure, taken as an aggregate of multiple samples and tests. Importantly, this is a mean, so it represents an average. While there are often outliers, the MTBF provides a good estimation of overall life spans and success rates for a given test sample. 

What represents a “good” MTBF score varies across types of devices, where they’re used, and the competitive matrix of offerings across multiple manufacturers. For that reason, it can be difficult to assert what constitutes excellence, especially if MTBF data is not available for every competing product. But as an example, our BL654 with internal antenna and BL654 with external antenna were evaluated for reliability prediction at 85˚C. The external antenna variant received an MTBF score of 1,883,242 hours. The internal antenna variant received an MTBF score of 2,108,359 hours. 

So the average device will operate for over 200 years right? Obviously not, as this does not account for multiple factors including the degradation of materials due to physical processes and wear and tear. To convert this to something more useful, we must convert that MTBF score to a reliability rating. The standard way of expressing this reliability is to find the length of time at which 90% of units can be expected to still be operational. Using this metric, a group of devices with an MTBF score of 1,883,242 hours has an expected reliability of 90% after 22.638 years -- a stellar claim to reliability and a source of great confidence for our customers. 

The Bottom Line: We Do What Many Others Don’t

If you’re not designing your hardware for these kinds of scenarios, you’re designing for the best case – and leaving yourself vulnerable to the worst. Ezurio’s reliability testing means that we offer hardware that OEMs can trust to do exactly what it says on the tin: provide long term reliability in rugged and challenging environments. And even in more tame operational conditions, the value is in knowing that a fluke incident is unlikely to lead to device failure. Our testing ensures that if your product is accidentally forgotten in a freezer, dropped, or mounted on a machine that begins wobbling rapidly, our wireless offering is more than likely to survive. 

This is an extension of the many ways we provider our customers a product that they can trust, elements of their design that resist failure, perform to expectations, and improve brand reputation. Modules from Ezurio allow our customers to leverage our expertise, simplify integration efforts, and as demonstrated here provide superior reliability. 

For more information on our product portfolio, please visit our website: 

https://www.ezurio.com/products