This story doesn’t sound like its got a lot to do with solar system reliability… but stick with it
Not many people know it, but when I graduated from my Science/Engineering degrees, my first job was as a hardware development engineer at a Telecommunications equipment startup. Apart from doing some hard core design work at Gbps speeds on optical fibre interfaces, I was also responsible for demonstrating that the equipment met its reliability promises – so the company could stand by its warranty without fear of in-field failure en-masse. Exactly the question that micro-inverter manufacturers must make when offering 25 years warranty before they have 25 years of operational experience.
Now what we were designing was revolutionary, the company was young, and we didn’t have 25 years of experience, so how could we promise our hardware would be reliable? Firstly by understanding the reliability of each individual component, right down to the 2mm resistors that were being soldered by robots onto complex Printed Circuit Boards. Like a chain with a weak link, the expected failure rate of our telecommunications device was governed by its most unreliable component. Thankfully integrated circuits and solid-state digital electronics aren’t very failure-prone.
Even so, my initial calculations showed that our theoretical worst case failure rate fell far short of our warranty promises, primarily because of a few high-risk components, which as a result we substituted for more reliable alternatives.
The reliability equation was based upon the reliability of individual components, and informed by experience in the field. Without in-field experience, the reliability calculation was assuming the worst, but we simply couldn’t wait for years of in-field experience before selling the devices. The solution was Highly Accelerated Life Testing, in which the components were subject to cycles of extreme temperature and humidity, while under operation. The reliability calculation could account for these tests, and if any failures were detected in the laboratory then our reliability would be very very poor, even worse than simply calculated from each individual part. But if there were no failures during three months of HALT then we could comfortably stand by our warranty. In the end that’s exactly what we were able to do, and I believe the in-field reliability turned out to be excellent.
Now, like many of you I have been cautious about micro-inverters, concerned about the reliability of electronics exposed to the high temperatures that regularly occur under solar panels. But while in the USA recently representing Australia in an International Energy Agency solar working group, something changed my mind, at least for one microinverter manufacturer. I was invited to visit a microinverter manufacturer’s laboratory in California, and the sight I saw was very familiar to me. Banks of hundreds of electronic devices packing chambers that were being subjected to length hours of extreme weather conditions, all while under operation. There was an entire warehouse dedicated to testing batches of pre-production and market-available devices, to the point that this company had a regular delivery of liquid nitrogen just to keep the place cool, such was its power draw. Add to that research laboratories in which root-cause analysis was being performed by very nerdy people in white laboratory coats. This wasn’t just impressive to me as a end-consumer, this was impressive to me as someone who has worked in a near-identical laboratory environment, except this was level-up impressive. I hope this isn’t the only inverter manufacturer that subjects its equipment to such exhaustive reliability checks, and the onus is upon micro-inverter manufacturers to demonstrate their product is reliable in Australian conditions.
Having walked into Enphase’s head office cautious about microinverters, I walked out of Enphase converted, having seen the extraordinary lengths they go to ensure product reliability. And I know I’m not the only convert.