Standalone Solar Installation....

We decided it would be good to have some electrical power in the greenhouse, primarily for growlights and at least one of our electric propagators during the winter months.

Taking power via a spur from the house mains is not so easy - there's nowhere to run a cable underground without cutting slots across at least one concrete pathway.  The best alternative solution would be overhead, from the house wall via a cable supported by a catenary wire.

However, the amount of power we're looking to draw is not huge, and so we've simply relocated three of the solar panels from our experimental array onto the greenhouse roof.


The first step was to trim the hawthorn bushes behind the greenhouse, for easy access.  This is a job we usually leave to the depths of winter - it's a jungle behind there at this time of year...

We built a timber frame and then fixed it to the structure of the greenhouse.  The panels face south and overhang the rear - they are also angled up at 30 degrees elevation, for a slightly more direct aim towards the sun to increase the amount of energy we can capture, and also to reduce the amount of panel obstruction over the glass roof, which would otherwise block the light from above.   We could have made this angle much steeper for better alignment with the sun, but then I'd have been concerned about the wind loading on the structure.  

finished timber frame before painting

The frame was designed so that the four verticals align with those of the main greenhouse structure.  Where the front edge of the frame rests across five of the greenhouse roof beams, we screwed it down at these locations too.

painted, erected and panels installed.

The solar panels are built within old static caravan windows, with a flange around the outside of the rectangular frame so they simply drop into the frame openings and are then secured to the timber faces by screws. 

at the rear

This will be a nominal 300 W standalone installation - we'll need to source a charge controller and feed a couple of 12V lead-acid 'leisure' batteries which we already have.  Since the open circuit voltage of the panels is around 32V, we'll connect the charging circuit in series - experience shows us they put out ~26V when under full load.  We'll likely draw the power down at 12V by connecting the load circuit in parallel across the two batteries.

We have an old 300 W 12V DC / 220V AC inverter which we've previously used for charging the laptop and cameras while camping, so we'll try this unit out initially although the LED growlight panels we intend to buy may require a cleaner supply from a pure sine-wave inverter.

So at least all the heavy work is now out of the way - I shouldn't really be doing this because of the ongoing problems with my neck and shoulder !

I'll post an update when we've completed the electrical installation....


Update 28-Aug-14 - PANEL TESTING

Today I hooked up the solar panels to the distribution box that contains the meters, switches and fuses etc, the output from which will be connected to the batteries and load via the charge controller.

There are a few different types of charge controller available.  The cheapest, for less than a tenner on eBay, have the load continually connected to the batteries and the output is regulated at the battery voltage.   The controller keeps the batteries charged, but if the load is either switched off or consumes a lot less energy than the panels could actually provide, then the available panel energy is not fully utilised*.

Another type of controller is that which charges the batteries until they're full, and then diverts the solar panel output to a 'dump' load*.  When the battery voltage has dropped beyond a set point, then the controller stops the diversion and resumes battery charging.  This arrangement gives two load outputs - the primary load attached to the battery and the secondary, dump, load which comes straight from the panels.  

With a charged battery these two loads can draw energy simultaneously until the battery voltage drops sufficiently to require further charging and the diversion to the dump load is cut off.  The dump load circuit is not regulated, coming straight from the panels, and therefore its voltage will vary with the prevailing light conditions, but this variation is no problem if the dump circuit is only used for fans and simple resistive heaters - these load types become self-regulating, i.e. if the panel voltage drops because of cloud cover etc then the fan simply turns more slowly, or the heater output is proportionally reduced.

I think this second type has more advantages for our installation - in the summer, when the growlights are not required, then the controller will be diverting continuously and I can use this 'direct' panel power via the dump circuit to drive extractor fans to ventilate the greenhouse.  In the spring when it's still cool, I could use ceramic fan heaters of the type you can buy for car windscreen demisting, although with a nominal 24V panel output I'd need two units in series or a type specially designed for trucks.

However, before I decide exactly which type of charge controller arrangement to install, I thought I'd carry out some testing on the panels in their new location.   By switching panels in or out on the distribution box I can test any combination; individually, in pairs or all three together.

The open and short-circuit tests were carried out at 12:40 pm BST, or around 28 minutes before today's solar noon, and in bright sunshine.  On average, the individual panel open-circuit voltage was 32 V and the short-circuit current 3.0 A.  All three panels operating together in parallel produced around 8.9 A when short-circuited.

I then connected a 100 W rheostat in series with my multimeter and switched the meter to measure DC current, before hooking it up to the output from the distribution box.  

Using just one panel (No. 3) because of the power limitation of the rheostat, the intention was to adjust the resistance and measure the current at 2V increments, to plot the 'characteristic curve' of the panel and determine the maximum power point.  

All solar cells exhibit a similar curve, with zero current at the open-circuit voltage and zero voltage at the short-circuit current - under either of these conditions, the cell cannot generate power.

However, this being England, the clouds rolled over and then it started to rain, and so it was 4.00 pm before it cleared again and I could resume the testing, by which time the sun had moved westwards and it was lower in the sky.  At this time of day, the short-circuit current from the panel had reduced to 1.68 A.

But anyway, here is the panel characteristic curve plotted at the conditions prevailing at 16:00 BST.

And the corresponding power curve, showing the maximum power point (36 W) occurs at around 26V.

To establish the maximum panel parameters without repeating these tests on another day at noon, the data series collected for the above plots were adjusted by factoring the measured current at 16:00 by 3.0 / 1.68 (the ratio of the short-circuit currents), to reflect the earlier conditions nearer noon today.

From this data, it can be seen that the maximum expected panel output is 65 W, at least in late summer, and therefore in our energy calculations we'll need to de-rate our assumed 300 W installation to just 195 W.  

We'll update again when everything's connected and working....



Update 06-Sep-14 

The system is operational, and now includes a fourth solar panel that we've relocated.

We built an extension to the roof frame for this fourth panel - the frame needed angle bracing because it overhangs the edge of the greenhouse roof to the west.

braced extension to panel frame.....

So the total panel area has been increased by a third, so we'll say it's now around a 260 W system based on the testing we carried out late last month.

I decided on using a 12 V charging circuit because we already had the 12 V home-built charge controller from the workshop.  This isn't an ideal device for an array which can provide 26 V at full-load current, because it means the panel voltage is always held down to around 14.5 to 15.0 V by the controller when charging, and the cells can't provide any more current to compensate so it only gives us around 150 W of charging power.  

However, when the batteries are fully charged, the controller switches to the dump load circuit from which we can draw whatever panel power is available under the prevailing light conditions, up to the full 260 W.

So, in our installation the batteries are connected in parallel for both charging and discharging - we've added an extra battery which is easier when parallel charging because it's not so important if the batteries aren't all in the same condition or have different amp-hour ratings.  For a 24 V controller charging in series we'd be limited to multiples of two 12 V batteries, and for each pair connected in series they'd both need to be of the same capacity and in very similar condition. 

I'm still tweaking the charging set points via adjustment of the mini-potentiometers on the controller circuit board.  I think the cut-in voltage is now set OK, at around 11.6 V, but I need a few more charge / discharge cycles to ensure the drop-out voltage is OK.  I'm aiming to set this to around 14.7 V.

This means the controller will regard the batteries as fully charged and switch to the dump load at 14.7 V battery voltage and only resume charging when the voltage drops to 11.6 V.

So here's the installation on the back wall inside the greenhouse - the wiring needs tidying-up a little and an insulated cover making to conceal the batteries.

The top box is the solar distribution unit with the individual switches and fuses for the solar panels, as shown in the photo from the previous update.  The voltmeter and ammeter measure the panel voltage and current.  

The lower box incorporates the charge controller and also has meters for the battery voltage and load current.   The charge controller switches the loads via a standard automotive relay, which means that the rating of the whole system depends on the rating of the relay fitted, currently 40 A.

There's a large isolation switch fitted in the primary load circuit from the batteries, so if a problem arises we can immediately switch off all power to the connected loads.  We haven't yet built junction boxes to connect the primary and dump loads - at present, there are just 30 A terminal blocks screwed to the side of the upper greenhouse shelf. 

You can also see our 300 W inverter sitting on the lower shelf in the photo - I've been using this with a 240 VAC 60 W tubular heater to discharge the batteries and adjust the lower set point on the controller.

There's a blocking diode required between the solar panels and the battery feed from the controller, to prevent the batteries from feeding power back to the panels during low light conditions and at night.   I had the original 20 A diode fitted, but when testing with the panels switched off I noticed the panel voltmeter was registering around 7 V, and so the diode was obviously leaking.

I'll need to source a replacement next week, but as a temporary measure there are now three 5 A diodes in parallel in there, which I had in the workshop. 

temporary 5A diodes....

So, the system is basically operational barring a few tweaks, and we're continuing to work on the load circuits and devices.  From the primary load circuit, we're think of individual switched feeds to a greenhouse overhead light, growlight panels and also to the 300 W inverter.

For the dump load circuit*, for the winter at least we're looking at electrical heaters - see our next post.....


* a little note on the dump load circuit.  The charge controller is a generic type and was built to work with either a wind turbine or solar panels (or both together).   For a simple wind turbine, when the batteries are fully charged and the controller drops out the battery feed, it's very important that an electrical load is maintained on the turbine generator otherwise the blades and machinery can overspeed and potentially damage themselves.  

However, on a solar panel installation, a dump load is not required at all for safety or operational reasons - if there's no dump load connected then the panels can happily just remain open-circuited until they're needed once again to charge the batteries.  The downside here is that there's a lot of energy available from the sun which is not being captured at all during times when the batteries don't need charging.