GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Audio Play Icon Headphones Icon Plus Icon Minus Icon Check Icon Print Icon Picture icon Single Arrow Icon Double Arrow Icon Hamburger Icon TV Icon Close Icon Sorted Hamburger/Search Icon
Musings of an Energy Nerd

Testing a Forty-Year-Old Photovoltaic Module

A decade after my last test, I take my oldest solar panel off the roof again to see how it's performing

This photovoltaic module was manufactured by Arco Solar in 1979. Installed on my Vermont roof in 1980, it has been exposed to the weather for 40 years. Photo courtesy of Karyn Patno.

I purchased and installed my first solar electric panel—technically known as a photovoltaic module—in 1980. The PV module, an Arco model 16-2000, was rated at 33 watts. Installed on the roof of my house in Vermont, the panel has been exposed to the weather for 40 years.

A decade ago, I temporarily removed the panel from my roof and performed a series of tests to determine the panel’s electrical output. I reported the results in a May 2010 article called “Testing a Thirty-Year-Old Photovoltaic Module.”

Those tests revealed that after 30 years, the PV module was exceeding its original factory specifications. Now that the solar panel is 40 years old—significantly older than the presumed lifespan of a PV module—it is time to once again bring the panel down to the ground for another round of testing.

Here’s the short version of the test results: The solar panel’s electrical output has measurably declined. My imperfect test method shows that, compared to ten years ago, the solar panel has experienced a decrease in electrical output of between 3% and 8%.

If we assume the worst, the panel’s output has declined 8%. In other words, after 40 years of service, the PV module has an output equal to 92% of its output when new. Few homeowners would place such a PV module in a dumpster; after 40 years, the panel still belongs on my roof.

In my first test, the panel was connected to a 35-watt incandescent light bulb. (This incandescent bulb is rated for 12 volts DC.) The PV panel was connected directly to the lamp, with no intervening battery. The test occurred at about 11:00 a.m. on a sunny June day; the air temperature was in the mid 60s.

Under these conditions, the light bulb was drawing 1.948 amps—about 3% less than the same light bulb…

GBA Prime

This article is only available to GBA Prime Members

Sign up for a free trial and get instant access to this article as well as GBA’s complete library of premium articles and construction details.

Start Free Trial

17 Comments

  1. User avater
    Ryan Lewis - Zone 4A | | #1

    This is impressive, but, wouldn’t a modest investment in one new panel yield you 10x more power ?

  2. User avater GBA Editor
    Martin Holladay | | #2

    Ryan,
    The topic of this article is PV module performance after 40 years -- not the drop in PV prices over the last 40 years.

    That said, you're right, of course. If you want to buy a PV module with ten times the output of my old 33-watt module, you could buy a 330-watt Panasonic PV module today for between $300 and $340. In short, PV modules last a long time, and they are cheaper than ever.

    1. Ben Elliston | | #12

      Martin, do you remember what you paid for those 33W modules?

      1. User avater GBA Editor
        Martin Holladay | | #13

        Ben,
        As I noted in my previous article on this topic ("Testing a Thirty-Year-Old Photovoltaic Module"), I paid $275 for the 33-watt Arco module. That's $8.33 per watt.

  3. Derek Roff | | #3

    Martin, your data is very valuable and the longevity of this panel is very impressive. I'd like to read more about another element in your system– the batteries. Have you done any recent tests on how your current batteries are performing? Can you post the link to your newest article on your battery life and how much cost per kWh the batteries add to your operation?

  4. User avater GBA Editor
    Martin Holladay | | #4

    Derek,
    Here is a link to the relevant article: "Batteries for Off-Grid Homes."

    Anyone contemplating the installation of a PV system should strive for an on-grid PV system without batteries. If you have an off-grid house, the batteries (and generator maintenance) will be by far the biggest expense for your system.

    For an off-grid house, expect your batteries to last for about 8 years -- 10 years if you are very lucky. Our batteries cost $4,600 and usefully store about 19 kWh of electricity.

    Calculating the cost of PV modules, PV racks, batteries, a gasoline-powered generator, fuel for the generator, and generator maintenance, I've heard that the cost to the off-grid homeowner for home-made electricity is between 50 cents and $1 per kWh.

    So -- connect to the grid.

  5. Derek Roff | | #5

    Thank you, Martin. As always, this information is very valuable to me.

  6. John Downie | | #6

    Martin,

    You wrote:

    “ In other words, after 40 years of service, the PV module has an output equal to 92% of its output when new. ”

    This seems a stronger conclusion than you can make here.

    Unless you tested the panel when new, you can now compare it to its rated output when new or its actual output 10 years ago but not its actual output when new. We don’t know how much power over its rated output the panel produced when new.

    1. User avater GBA Editor
      Martin Holladay | | #7

      John,
      You're right. But following the principle of Occam's Razor, it's unlikely that the electrical output of this PV module, which already exceeded factory specs when it was 30 years old, exceeded factory specs even more when it was new.

      1. John Downie | | #8

        Martin,

        I don’t see how Occam is helpful here.

        There are a number of important assumptions in the analysis, including the specific performance of this particular panel when new, forty years ago.

        We don’t know what the underlying panel-to-panel performance variance was in 1979, where Arco specced the mean panel so that they met the label rating for most or all of the panels, or where this particular panel sat among the range.

        You have a panel which represents a draw from the population of remaining 1979 Arco 16-2000 panels. In addition to the performance variance in 1979, there may be a survivor bias at work, as well.

        The curve you are inferring is flat for 30 years, followed by a slow decline.

        To me, this is no simpler or compelling than a curve that started somewhat higher and declined at a slightly rate over time.

        The starting performance level is critical here. If it was 8% above spec, for example, the overall decline would be approximately double what you have concluded.

        Given that the t=30 values were higher than the assumed t=0 values, and the 20-ish year design life of the panel, this seems in a reasonable range of possibilities.

        None of this changes your overall point that PV panels work well for a long time, I just think your conclusion is more precise, and possibly rosier, than merited.

        John

  7. Charlie Sullivan | | #9

    Next time you want to do this measurement (after another 10 years), let me know, and I might be able to meet you somewhere halfway with equipment to plot out the full current-voltage curve and measure the incident insolation. As John points out, that's less useful that if we had an initial version of that curve but it might be interesting.

    1. User avater GBA Editor
      Martin Holladay | | #10

      Charlie,
      I'll be happy to invite you to help me with the 50-year test. Thanks!

      1. Charlie Sullivan | | #11

        See you in 10 years then.

  8. Nick Defabrizio | | #14

    This is good news for those of us who recently installed pv panels based on part on return calculations that assumed at least a 20 year life for the panels with only modest reduction in performance. The salesman who sold me the system told me that reasonable indications are that most panels meet or exceed their rated life and performance metrics. Of course this is just one test but it seems like a nice outcome.

    I wonder about whether ( in general) modern panels-mostly made in China and with much higher performance -should be expected to be more reliable or less so than your 1979 versions.

    1. User avater GBA Editor
      Martin Holladay | | #16

      Nick,
      Q. "I wonder about whether ( in general) modern panels--mostly made in China and with much higher performance--should be expected to be more reliable or less so than your 1979 version."

      A. That's a good question. We're going to have to wait a few years to find out the answer.

  9. Mark Klein | | #15

    For a little perspective that $275 panel would be around $1000.00 today if we adjust for inflation. So those early panels we were buying would be around $30.00 a watt in 2020 dollars. We thought they were quite a deal...

  10. User avater
    Chris Byers | | #17

    We work with planning board across the northeast, and primarily in Maine to permit large scale solar projects. We often get asked how long these systems last, and even though I've worked in the solar industry for 11 years, it's hard sometimes to convince someone that just because there is a warranty that it means the product will last as long as the warranty. Especially when most modules are made in China--the optics of where these products are made aren't helpful to those that may be skeptical. It's one thing to list the performance guarantee of modern solar panels (eg. 30 year @ 85% at year 30), but it's great to have this data as a side bar piece of evidence. If modules that were made 40 years ago and are performing at 92% of their power output, then seemingly you could assume that the more superior technology of today's solar modules would be equal to or greater than the one you bought. Of course, as you noted, n=1, so I'll take it for what it is. It's a test nonetheless. Appreciate you sharing the data.

Log in or become a member to post a comment.

Related

Community

Recent Questions and Replies

  • |
  • |
  • |
  • |