Solar PV

Original:   Feb ’99

July ’99  -  Step-up switcher module addition
Jan ’01  -  Battery store capacity increase

If you feel interested in this topic, and you haven't heard of the Centre for Alternative Technology, you really must try their web-site first. If that doesn't impress....well, lets just say that the following isn't likely to either. It is but a toy compared to the CAT version.

[Why not visit the CAT anyway, or go on a weekend course. And if you don't like the thought of PV panel cost, go on the Solar water heating course instead, and build a low cost panel - see pic]

Above: Six "13W" electronically ballasted flourescent lamps supplied directly from the Solar PV systems 150v DC battery store. Most currently available lamps of this kind will work satisfactorily from such a voltage (see table) 

These lamps were £3.99 each at Homebase, and although they are no longer available, Tesco's are currently (June '99) Philips lamps of different types for £3.99 and £4.99 - so always shop around, since they can be anything up to £12. Incidentally, the Chinese manufactured lamps are every bit as reliable as the recognised European brands in my experience.

The panels/battery combination also power the kitchen kettle and a small Moulinex cooker - without this load, the batteries remain near enough fully charged, with the majority of the power going into the (immersion heater)* dump load. With 150v DC one has to be careful with contacts - the kettle auto switch off device welded up the first time it tried to switch off the DC (believe it or not, an arc quenches very quickly once the supply is removed and 50Hz zero crossings every 10ms are slow enough to prevent the sort of self perpetuating arc one gets with these small gap contacts when operated at DC).

A kettle operating from 150v instead of 250v is a very pleasing thing. Boiling time is doubled, but that's not a problem for a couple of cups worth of water in a standard kettle. At switch off (if the thermostat hadn't welded), boiling is very gentle - reminiscent of older days,when a kettle would be left permanently boiling on the kitchen range. You can safely leave the thing boiling away (well, a slight exageration, but it does take along time to boil dry). A "2kW" kettle will take 5/6A at 150v (ie about 900W). I find that the Moulinex oven takes about 2.5/3A.


*Top fitting Immersion heater Dump load W A R N I N G:

As the following picture shows, problems can arise even with an apparently low duty cycle Immersion heater dump load.

This is a top fitting dual element Immersion heater - a 3kW thermostatically controlled element being run from the standard 240v AC mains and the smaller '2kW' element acting as the PVs dump load (actually dissipating 900W from the 150v DC). The reason the dump element is separated from the main fixing boss is that it UNSOLDERED ITSELF in-situ! - hence the warning. Looking at the element discolouration gives the secret away as to what happened. Water has boiled, and steam displaced the water, allowing the top half of the element to get much hotter than 100 deg C, eventually melting the solder. Fortunately, because the overflow was seen to be running, an investigation was made - just in time. There was water/steam escaping from the hot tank, which is in the roof space of the bugalow, but it hadn't started coming through the ceiling.

Conclusions: Use a tank with two, side fitting immersion bosses, and use separate immersion elements. This has now been done, ie

End of diversion, back to main story: 


*Seven of the eventual nine panels (BP275 75W) installed on the roof*

Each row is fastened to a pair of horizontal aluminium angles, which are themselves fastened through the roof with a couple of steel tie-bars. The nine panels are connected in series, and charge twelve series connected 12v lead acid gel batteries. OK, it should be 12 panels for 12 cells, but the batteries were available at modest cost - which is more than can be said for the panels, which were about £250 each.

Batteries, switching, metering and charging hardware:




Proposed and pretty much implemented interconnection diaram:


Twelve of these will sit on a motherboard, as below


At the moment, only one such board is fitted, with the 18k resistor taken from the +bat connection and fed via a 220k resistor to the overall 150v rail. This at least prevents too high an overall charge voltage, but does not protect individual batteries that may overcharge before the rest.


Conclusions so far:

i) It works!

ii) Contacts need to be checked to see if they will operate on DC without welding up.

iii) 150v DC is probably not very safe.

iv) 9 panels for 12 batteries, now it is summer (June), results in low charge current mid day (because when the panels are hot, the output voltage falls). I think this will need some rethinking, but it may be enough simply to modify the dump load switching (ie, to sense panel temp, and switch the dump load on when this is high)

Alternatively, a simple boost regulator in the output connection of the panels could be very efficient in this application (the output voltage never falls to much below the input voltage, even when the switching FET is not being driven. Loss is restricted to the generation of the extra boost voltage, which would never need to be greater than 5 to 10% of the input voltage in this case*). 

 *This would be to under use the switcher though. If the boost regulator is capable of producing full output at lower input voltages, then it could be used to 'match' or 'power-track' at all light levels, maximising the panel output.

iv) Combined with a solar water heater, the immersion heater dump load is a very good way of improving what will sometimes be low grade heat (ie, water at lower-than-needed temperature) to higher grade heat - there being no real equivalent with the pv/immersion element of loss of efficiency (and resulting limited max temp) that imperfect insulation of the solar water panel results in on a non too sunny day.

Incidentally, my nine panels usually put enough power into the immersion heater dump load to allow me a hot bath each evening (this is in addition to keeping the battery store fully charged0.


To be tried....

Turning the DC into AC is easy enough and very efficient (99% easy enough), as long as a voltage change is not also required. Most 12v to 240v converters these days use a bridge powerFET circuit to give inversion at 50Hz. Loss is very small at such a low frequency.

Alternatively, the DC could be retained, and ovens/kettles etc fitted with electronic on/off switches - after all, the current supply for even the kettle is only 5A, and a low cost TO-220 powerFET will easily switch this without needing to be heatsunk.

During the summer, when the panels are really hot, the output voltage falls and there is very little charging current (remember I'm using 9 panels to charge 12 batteries). Adding a boost switching converter after the panel should solve this - and there are two good things about this switching topology. Firstly, even without drive to the switching transistor, output is as before (less one diode volt drop), so it can be put into circuit even if not driven. Secondly, the switcher drive is referenced to -ve, and thus, safe to deal with.

Stop press (30th July '99): Tried this, using a pulse generator to drive an IRF740 device (and BYX71 series diode), and results were promising - firstly, a peak in battery current could be seen as the duty cycle was increased, followed by a fall off as the optimum loading point was passed. Secondly, the panel current was almost constant, even when the immersion shunt load came into circuit, exactly as would be expected.

I was driving the FET at 30KHz. Although the circuit was vastly under powered, the intention is to put 6 to 8 switching FETs, complete with their own flywheel choke and series diode, in parallel. This will allow continued use of standard devices and standard mica washers. Also the ferrite flywheel choke can remain similarly standard (an RM10 core). See proposed circuit


Battery replacement (Jan 2001) 

Actually not a replacement, but a much larger capacity unit put in parallel. These were 3 cell (ie 6v) 160A/H units rescued from a former Ionica radio site.

So, 24 of these very heavy units were placed in a stout timber framed housing.

This capacity (160a/h at 150v) will allow me to run my cooking needs as well as the lighting. And the PC’s and monitors (plus amateur radio gear) can all be fed from this source.



Some links

The solar system that Jenny has at jsquared

Similarly, Nicks system at Milton Keynes