DConf 2013 Day 3 Talk 1: Metaprogramming in the Real World by Don Clugston

[Don]

On Tuesday, 11 June 2013 at 20:02:29 UTC, Walter Bright wrote:
> On 6/11/2013 12:21 PM, John Colvin wrote:
>> On Tuesday, 11 June 2013 at 18:47:35 UTC, Walter Bright wrote:
>>> On 6/11/2013 8:28 AM, Adam D. Ruppe wrote:
>>>> It is great stuff, solar power is almost free money if you 
>>>> can wait 20 years for
>>>> it.
>>>
>>> Yeah, but you'll have to replace it before 20 years!
>>
>> Source? There's not much that wears out in a photovoltaic 
>> AFAIK. The associated
>> electrical components may break however, especially on some of 
>> the more complex
>> setups.
>
> Don't have a source, I read it long ago. Note that none of the 
> advertisements, brochures, etc., mention expected life of the 
> PVs.

That's not correct. Almost all manufacturers provide a 20 or 30 
year warranty.
Warranty periods have been slowing increasing as the industry has 
gained field experience.


> I do know that the life of any semiconductor is measured as the 
> integral of the heat it experiences. Heat causes the doping to 
> migrate, and when it migrates far enough the part fails.

That's true for certain kinds of dopants (it's particularly true 
if you have copper involved), but dopant migration is not an 
issue for any commercial solar modules that I know of. (The 
situation may be different for exotic technologies). This is 
because solar cells are very simple devices, they're just 
enormous diodes.

Virtually all solar module failures in the field are caused by 
mechanical issues (bad solder joints, cracks, delamination), not 
by semiconductor degradation.


> PV panels can get pretty hot in direct sunlight.

They do. Still not as hot as a CPU though!

> Heating/cooling cycling will also cause cracking.

Most of these problems were solved in the 80's.

We were continuously doing accelerated lifetime testing of our 
own modules and ones from various manufacturers. Temperature 
cycling, humidity freeze, hail impact testing (that's fun), wind 
load testing (that's really fun), etc.
For some silicon modules you can get oxygen-boron complexes 
induced by UV, which causes a slow reduction in power, but our 
modules survived 200 years equivalent UVB exposure with no 
degradation whatsoever.


> Circuit boards, inverters, etc., also fail, and you'd need some 
> assurance you can get replacement parts for 20 years.

That one is definitely true. Even worse is batteries for off-grid 
systems, batteries have a very short lifetime.