Solar energy -- the math

Here's a link that explains the math, and keeps me from just plagiarizing what is already out there.

greenecon.net

OK, so as my last post revealed, I'm going to attempt to create solar panels from scratch. This is kind of a neat engineering problem, where I have to address the following questions:

1. How do I protect the solar cells while still exposing them to the sky? What about hail? How much reinforcement do I need if 2 feet of snow falls on them?

2. How do I attach them to my roof? Or would ground-level poles be better?

3. How big should the final system be?

4. What is the best wattage/voltage for the panels to run at?

5. how big should the panels be?

6. How much heat are they going to generate, and if I build the panels from wood, will I end up starting a fire on my roof?

7. Is it better to put more panels up, mounted flat to the roof, or to use less panels, and mount them at a sharper angle? What kind of extra reinforcement will I need to deal with wind, if I mount the panels at an angle.

So, to help answer these questions, I have looked at a lot of websites, watched a bunch of youtube videos, and really haven't come up with a lot of answers. The main thing is to avoid over-engineering this project, and paying way more than I should. So what I've done is ordered a enough supplies on Ebay to build a single panel, at 129 watts, and 18 volts, along with a simple grid-tie inverter that plugs directly into a wall socket, and a 'kill a watt' power meter to monitor what amount of electricity is generated. The shopping list so far?

$99.99 for a 300 watt grid tie inverter, rated for an input of 14-28 volts
$138 for 72 1.8watt, 0.5 volt solar cells plus miscellaneous parts to attach them to each other
$21.99 for a sheet of plywood to mount the panels to.

So, the math on the solar panels is like this:
I'm building 2 mini-panels of 36 cells connected in parallel, which gives us 18 volts (0.5*36), and 64.8 watts (1.8*36) I can then connect those 2 panels in series, to keep the volts steady, but push the amps up, essentially to stay under the 28 volt limit on the inverter.

I'll add some pics once they arrive, and once the panel is built, so that its construction will become clearer. The construction of the actual panels will probably require about 50 hours of soldering/building time, so I'm imagining getting the panel into operation about mid-April.