I have been interested in Alternative energy since I was around 11 years old. I remember spending hours reading about electronics and solar panels system in the library at school.I guess that would be 1969 and I have no doubt that watching Felicity Kendel in "The Good Life" (BBC 1975-1978) at the age of 17 had a lasting effect on me tooThis picture shows my first REAL venture in to a PV system. It was 24V panels that direct charged a 24V battery of about 300AH and ran the lights in my house. This was about 1999. Just before the world didn't end in 2000 BTW the next end of the world is now 21st December 2012. Guess that never happened either...

Even this early system tracked the sun it was a simple Polar mount system with limited travel but in the garden it was in it was good enough. I used a PIC16F84 and an MSF rx to get the time and mearly moved the PV 15 degrees an hour. over the movable range at the correct time.

Over a while I expanded the system a little... Below is the Original 300AH battery system.

However what I discovered whilste playing was that a PV maximum power transfer point for a 24V system was not at voltage of 24V. Each nominal 12V Panel actually gave an OC volage of about 18V and peakpower transfer was around 16V. This varies a little with each panel and the light falling on it. Clearly to get ones monies worth you needed a Peak Power Point tracking Power conditioner. So I set about designing one...

The basic concept was to string all the PV's together and produce about 160V OC and use a switch mode technique to load the PV's whilste monitoring the PV voltage and current (V x I = Power) and adjusting this load for peak power from the PVs. The o/p Voltage of the system could be whatever you wanted I stuck with 24V as I had an Inverter of this voltage. This picture is the First switch mode PSU I designed to run the system I needed +-8V for the opamps +5V for the logic and processor and a floating 12V to drive the FET switches

The first PSU design had a few problems with the operating voltage range it work fine down to 3.5 Volt but was not happy at 29V (useless) This is the redesign. It worked between 12V and 60V much more usefull. While I was at it I added a few features I'd missed on the first prototype. A fuse and some connection points and some LED indicators.

This is the brains of the system, the processor board. It contains the processor a PIC18F252 all the signal conditioning for it's 5 ADC channels the 485 muti drop coms chip and the line drivers for the FET drive to the switching board.

This the switching board brand new and un-modified. It will not look quite this good ever again.

PV's have an o/p characteristic that is vaguely constant current like. i.e as you increase the load the current increases to a point and then the voltage starts to drop rather than current increase much more. So, for Initial testing I used a bench power supply with the voltage set to 60V and a constant current output set to 2 Amps. The device should, when working track the max power out as current limit and Voltage is adjusted.

Secondary testing took place connected to the PV and a 12V battery as the load. (Told you Any Voltage just a s/w change) You may notice the inductor has changed to one Significantly larger. The inductance of the original inductor was ok but it saturated. at about 2Amps and I needed it to still be good at 20A. Hence the change.

After a lot of work I had the system ready by the last half of 2003. We (susie and I) had been looking to move house to somewhere bigger. Hopefully with a workshop. We (Susie) found a place in the last half of 2003 so all wok on the PV system stopped and it was dismantled ready to move. As with most moves there were delays. We finaly moved Jan 2004.

As soon as we were settled in I installed the PV panels with the original Polar tracking system. Not the best tracking system for this location but all that was available.

After the system was up and running a while it very soon became apparent that that there was a problem with the Battery current monitoring section of the Processor board. The signals being very small and the amplification being on the processor board ment there was interference. This problem was rectified by placing the preamp directly across the Current shunt rather than on the Processor board.

The modification on the Processor board was simple I just connected the new Instrumentation Amp preamp to the chip socket of the removed Op amp I had used before.

After this modication the system worked well on test on the bench for several months.

The next modification needed was to shut the the PV power off to the system if the battery Over Current trip went. This was because if this happend, the supply voltage to the PV-Reg PSU would rise to 160V (the PV OC voltage) and the PSU would self distruct... The modification used Zenners to trigger Thyristors to Short the PVs themselves and Load the PV-Reg O/P with big Resistors, when the PV-Reg O/P Voltage went above 50Volts.

Next On the list of improvements was to use an Azimuth and Elevation tracking system. This would seriously improve the time that the PV's could be kept at the optimum angle to the sun and therefore attain maximum power from the sun for longer. Some experimentation had showed me that if the PV's are kept within +-20degrees of perpendicular to the sun you maintain >95% of attainable PV power o/p. The tracking system updates every 15mins (3.5 degrees Azimuth less on Elevation).

I did a rough estimate of what the increase in power would be and it should be a minimum of greater than 20% more Power collected compared with not tracking. This figure depends on the time of year. In the winter the sun tracks about 100degrees Az and summer more than 250 degrees. So in summer the collected power increase by tracking would be much greater than in winter. It turns out that in summer the gain is 30-35%

and I still have a thng about Felicity Kendle even if she is 60...