February 2008


Solar thermal is an attractive renewable energy source. Like other solar energy sources, it works, but is often not economically viable compared with established sources, such as coal. With rising energy costs and environmental concerns, there is increased interest in solar today.

Solar Stirlling Engine

The design above is a modern, reflective parabolic dish from Stirling Energy Systems. The dish produces about 25 kW, enough for about 10 homes in the US. This video on YouTube has more details. This video shows a small scale Stirling engine operating on solar power, similar to the full scale dish. Based on data from a large installation, it looks like 4 of these large dishes pictured above can fit in one acre, so that’s 100 kW/acre. The United States consumes about 440 GigaWatts. Roughly 18 million of these units spread over 6875 square miles would satisfy the entire US electric power requirements. That’s about 6% of the land area of Arizona, one of the two states with high solar insolation. It could be built using only 5.7% of the Sonoran desert.

The US congress is currently debating an economic stimulus package of $150 billion. If all of this money was used to build a solar array and each collector could be installed for under $50000, the power output would be equivalent to 44 Hoover dams, or about 20% of US electric power consumption. At current market prices, such an array could pay for itself in under 4 years. After that, the array could yield as much as $38 billion per year in gross revenue.

Of course, transmission of this power from the deserts of the Southwest to where it’s needed is a challenge, but it’s something to think about.

I really like Atmel AVR microcontrollers. They are simple, powerful, and inexpensive. There are lots of development tools out there, but I like AVR Studio. AVR Studio is made for Windows, but it works great inside Parallels desktop on my iMac. Parallels sees the AVRISP mkII USB programmer also. I had no problem upgrading firmware on the programmer. Now that I have a nice development environment set up, I might be able to finish some of the numerous hardware projects that have been languishing for years.

One minor problem I ran into was the clock rate on the ATtiny24. By default, the clock is divided by 8, so it runs at about 1 MHz using the internal RC clock. This was too slow for the default program clock, which caused programming errors. This is to be expected, but it ends up crashing Windows! After testing on a native Windows machine, I located the problem, reduced the programming clock and disabled the clock divider fuse. After that, everything was smooth. I even got AVR-GCC working using WinAVR.
I’ll post some of the circuits I’ve been working on in the coming months.