Electronics Forum / Electronics / July 2005

Here are the figures I'd researched a month or so ago for another
discussion. Solar cell makers buy surplus semiconductor grade
silicon for manufacture of their cells, so they should count the
energy cost to refine silicon to this stage. There are processes
to make "Solar grade" silicon out there but none are past the lab
stage of testing: these take 1/3 of the energy to make the silicon.
At any rate here is the figures I've come up with.

The calcs assume:

1. You are living on the sunniest place on the planet with 3000Hrs
of sun per year, somewhere on the equator on a cloud free plateau,
Sydney/Melb figures will be greatly reduced as you are closer to
the south pole at sea level in air pollution.

2. The solar flux is a constant unvarying 1.4Kw per square metre
from the sunup to sundown.

3. There is no age related loss of power at all(more like average 10%
and peaking at 20%).

4. You have invented some new way to slice silicon that has no losses.
The figure is closer to 50% at the thickness I have quoted.

5. The only energy costs are that to make the silicon, nothing else.
Please not that this does not count the lack of efficiency of energy
storage in batteries, the cost of making the batteries, the cost of
making the inverter, the cost of converting the silicon into a cell
the cost of refining the metals and glass used in the solar system or
the cost of shipping and installation.

6.It also does not count the multiple sets of batteries over the life
of the system.

7. It does not count the CO2 made during the chemical combustion part
of refining the silicon with carbon from quartz.

Here's what I'd researched earlier, Trevor on the 19/05/05:

"I've read it's a six step refining process that requires 2130 KW/h per
kilo of silicon. The density is 2330 kg/cubic metre, so assume you can
slice the silicon to 0.5mm thickness and have no sawing / polishing
waste and no energy expended to saw / dope / connect the cells,you'd get
2000 square metres of silicon for just under 5 GW/hours of energy. Solar
radiation is ideally 1.395 KW/metre square by 2000 square metres cell
area @ 3000 hours per year of sun under ideal conditions at the best
spot on the planet.

This gives 8.370 GW/hrs of energy per year by 15% cell efficiency to
give 1.255 GW/hrs per year of electricity.

so 5 GW/hrs divided by 1.255 GW/hrs per year gives 3.953 years to
cover the energy costs of making the silicon?."

payback = Four years???

So in a nutshell you get 3.953 years energy payback time assuming no
losses in the best part of the planet, assuming no energy losses, that
your silicon is at least 15% efficient?? and assuming no energy went
into making the solar cell from the raw silicon.

I would put it to you Trevor, that making the raw silicon into a panel
will double that figure, shifting the panel from the world's sunniest
site on a cloud free high altitude plateau on the equator to a region
far from the equator at sea level, like Sydney, would double that figure
again as you'd no longer have the CONSTANT 1.4KW per sq metre solar flux
I used in my calcs.

    Finally a Telstra tech once told me that as a rule of thumb you
get half the energy out of a battery as you put into it, so your figure
would double again, so 4 years x 2 x 2 x 2 = 32 years assuming your
batteries never die, your inverter is 100% efficient and you solar cells
never age. If you could use/sell your solar power at the 24 Volts it
makes without needing to store or convert it you would still have a
payback of about 16 years.

The only commonly available cell on the market with any real
usable performance WILL have this problem....regardless of
what marketing tells you Trevor. There are experimental cells
of different design but they are not available retail yet.

It sounds like you've been listening to your solar cell
retailer Trevor, over it's lifetime, an alternator is
environmentally less polluting to manufacture than a solar
installation but would pollute more if run constantly all
day instead of when needed, which I thought your use would
be?.

Regards
Mark Harriss

If you want a quiet RF environment, some "pure" sinewave inverters put
out a bit of RF. Mine (SEA Voyager) puts an S5 signal on some of the
amateur bands. Can be suppressed of course - but that's more work.
It is very efficient though - even at low levels  (<50W).
Alan