Cobbled together PV in W.Sussex

[Outtasight]

ROFL... Nice one.

No, she's just a kind soul that lets me do what I want in my space. The power centre is in the corner of my "cinema" with one wall painted grey as a screen for the video projector and otherwise generally full of AV gear.  The inverter and battery bank are under / behind one of the comfy chairs.  The other half of the living room is separated by a half partition. It used to be two rooms but the previous owner knocked it through apart from the chimney & fireplace that was in the middle of the wall so that had to stay.  That half of the room is "her" side so has the usual dining room / living room accessories and no "weird stuff". 

Besides, she's quite keen on the idea of solar. Her sister and father have both got big grid tied systems on their houses (3kW & 5kW respectively).  Her dads house has the inverter in the dining room on the wall and a massive air source heat pump in the utility room, so it kinda runs in the family...

She probably spotted I was a mad inventor type and was attracted by that  .

That said, when it eventually stops raining / snowing / being dark as night in the day, she'll probably want to be able to see daylight through the patio doors and be able to open them and go outside...  .

[Outtasight]

Some examples of what the built in web server gives you...





Apparently, with an adaptor, the SunSaver MPPT-15 can be connected to the EIA-485 serial bus on the TriStar and then bridged to the Ethernet so that you can remote control and instrument the SunSaver, even though it has no native IP capability.  That way, I'd be able to log data from both controllers and do stuff like intelligent load management.

Even though it was cold outside today, it was at least very sunny and so I could finally make some progress on painting the wood preserver on the frame for the BP panels sitting in the window.  With clear skies above, I managed to run the water heater and the computer room from solar and still had 8A to spare to keep charging the battery.  In about 90 minutes, the water went up by about 9'C.  So, rather conveniently, each 10 minutes of solar power provides a 1'C boost in water temperature.

[RichardKB]

Because you have an MPPT controller you could do with a higher array voltage for your 24V battery setup. I have just checked the datasheet the maximum voltage is 150V this would make it more efficient if you had a higher array voltage, say 3 panels in series or 4.

Rich

[EccentricAnomaly]

Clearly you want the array voltage to be high enough that in low but usable light conditions the Vmp is still enough above the battery voltage that charging can take place.  Also, higher voltages will result in smaller losses in the cabling.  Beyond those considerations, are there any advantages to higher array voltages?

[RichardKB]

Clearly you want the array voltage to be high enough that in low but usable light conditions the Vmp is still enough above the battery voltage that charging can take place.  Also, higher voltages will result in smaller losses in the cabling.  Beyond those considerations, are there any advantages to higher array voltages?

No you are quite right with your statement but with an MPPT controller running with only 32V which is only 4V above the charged battery voltage the controller has very little room to work with.

[Outtasight]

I did try reconfiguring the arrays for 36V nominal a while back but there were two problems... The SunSaver could only go up to 75Voc and I have some amorphous panels that put out 26Voc, so three in series could put out 78V.  I played with a voltage clamp at 69V but the controller didn't like it.  The second problem was that groups of three panels suffered from shading more, especially as the panels are just sort of spread around the garage and garden wall.  What with nearby trees and encroaching shadows from walls and fences and so on, it's easier to keep the more of the array productive if it's broken up into small "cells" that can function independently in patches of light.  If the whole lot was connected in series there'd only be an hour or so in the day when the whole thing was shade-free.

If I get round to mounting the whole lot on the house roof then the shading problem will go away and all the panels will be angled at the same bit of sky so the array will be balanced.  Also, it would be ok to run at high voltage up on the roof - away from the risk of cats chewing the cables and human folk coming into contact with the panels. 

It's worth thinking about though... The 9 main panels on the garden wall could be connected with the 160W 35V panel in series with the others connected in two parallel strings of four to give me 800W at 105Vmp.

I'll stick with my 35Vmp set-up for now.  If you can keep the wiring losses down, the controller actually operates at its highest efficiency at lower voltages.  This is especially true at low powers.  At 100W (common on a heavy cloud day like today), the converter efficiency for a 98Vmp array input is only 90% compared to almost 97% for a 33Vmp input.  So, if you can keep the wiring loss down to 2% or less, running at lower voltages can be advantageous in winter.  Even at the other end of the scale it's better to run at low voltage (33Vmp at 1kW gives 98% efficiency compared to just under 96% at 98Vmp).

Luckily, most of my panels are very close to the controller as they are effectively just on the other side of the wall and the cable run from the outside junction box to the controller input is about 3m.  From the junction box, each pair of panels has its own feed using differing cables from 2.6 to 6mm², depending on the power of the panels on that leg.  Now that I'm pulling over 20A instead of just 11.5A from that junction box, I'll have to upgrade the feeder to 25mm² to match the rest of the new wiring I installed for the new controller.

The big Sharp 340W string on the garage roof fares quite well on the SunSaver controller as the pair put out 46.4Vmp. That helps keep the losses down as they are at the end of 13m of 4mm² wire to the house.

[EccentricAnomaly]

Outtasight, thanks for your detailed and numerate discussion of your experience with your panel setup.  It's really helpful to me, at least, in understanding the practical issues with designing a system.

[Outtasight]

Today's Episode:  Putting the Navvi into Navitron 

I finally got a couple of dry enough days to get the frame together for the new BP380U panels and so out they went, replacing the last of the old 12W modules.



This then presented a couple of problems. 

Firstly, the array was now making so much power (I've seen 1,047W logged on the new TriStar controller) that the feeder from the junction box outside has been getting a little warm... Not good.  Both in terms of it (and the chocblocks in the the junction box outside) getting hot and also in the power losses.

Secondly, the change from amorphous to crystalline cells has made the array much more susceptible to partial shading.  Covering one cell with my hand causes the amps to collapse from a string pair.  The overhead wires to the garage cast a moving shadow across the six panels that make up 480Wp of the array and now at least one of the three blocks always had a line shadow across it.

I bought a bigger junction box and used some solid earth blocks to make up positive and negative bus bars.  This connects to the gubbins indoors by 25mm² jump start cable, eliminating the volt drop and heating problems at even the new controllers 60A maximum.



The second problem involved some push fit waste pipe and a lot of digging.  The push fit joints to link the pipes and make the right angles are easy to use and should be water tight with their built in o-rings.  Pretty cheap too.  I was considering using pond hose but it's more expensive and you'd have a hard time feeding the wire through 6m of the stuff.  With the waste pipe I could push the wires through it section by section and then join the sections up when I'd finished.  A bit of oil on the o-ring makes it easy to push them together.

I decided to replace the two other scrawny wires that linked the small strings on the garage roof with a bit of 6mm² twin and earth cooker wire that I got from a car boot sale last year.  Being only chunky stranded, it was no good for an overhead link as the swaying in the wind would damage the copper but buried underground it's in its element.  I twinned the positive and earth conductors so that (in one direction at least) it is actually 8.5mm².  That should be enough to allow some modest expansion of the array on the garage roof  .



Here you can see it all fitted together.  I considered putting the CCTV Cat5 cable down it as well but it's not a good idea to run data in the same trunking as high power cables.



In order to keep the peace with 'er indoors, I had to devise a way to do the construction work so as to not make a mess of the lawn. 

The tricky part was cutting out the turf in sections to allow excavation under it.  Laid to one side for later and the soil taken to a hidden part of the garden in buckets to keep it off the rest of the grass.  Here I haven't started digging out yet.  I went about 25cm down in the end.



I was amazed the first time I visited Japan.  I was walking to the hotel after a night out and saw a massive hole in the ground swarming with workers and football stadium flood lights and a huge noise.  I pitied the poor residents their lack of sleep.  But walking back that way the next day, I was stunned... The roadworks were all gone.  Filled in.  Finished... In one night.  British Gas and Thames Water, take note  .

So the missus expected the same service.  Later that night by garden flood light...

[Outtasight]

Free hot bath today!

Nearly wall to wall blue sky today and with the added charge capacity of the TriStar, I was able to run round and turn everything on!  The battery had done it's last 20% by about 10am, I turned on the 115V water heater and put all the computers on solar and even the fridge-freezer too.  The 3kW inverter can handle all that lot and a 850W kettle without throwing a wobbly.

The water in the tank is never mains water cold as there's always some dregs of heat left in it from the previous night and so it started at about 29'C at 9.30.  Five hours later the thermostat clicked off at 61'C.  My first full tank of piping hot water without recourse to any mains electricity .

The 650W 115V water heater is just about the right size load for the array as it can service that load and still put up to 8A into the battery to finish off its absorption phase, keep it floating at 1-2A or recover from a short gap in solar power from the odd passing cloud.

By the end of the afternoon, the TriStar read 4.15kWh collected and that didn't include the other controller that I didn't data log today but was also pumpin' out around 320W for quite some time.  Most of that power went directly through the inverter and to the loads as the battery was full and just floated in the afternoon, filling in for the odd cloud or bit of haziness. 

Tomorrow's set to be another bumper harvest day too! 

[Paulh_Boats]

Free hot bath today!

Nearly wall to wall blue sky today and with the added charge capacity of the TriStar, I was able to run round and turn everything on!  The battery had done it's last 20% by about 10am, I turned on the 115V water heater and put all the computers on solar and even the fridge-freezer too.  The 3kW inverter can handle all that lot and a 850W kettle without throwing a wobbly.

The water in the tank is never mains water cold as there's always some dregs of heat left in it from the previous night and so it started at about 29'C at 9.30.  Five hours later the thermostat clicked off at 61'C.  My first full tank of piping hot water without recourse to any mains electricity .

The 650W 115V water heater is just about the right size load for the array as it can service that load and still put up to 8A into the battery to finish off its absorption phase, keep it floating at 1-2A or recover from a short gap in solar power from the odd passing cloud.

By the end of the afternoon, the TriStar read 4.15kWh collected and that didn't include the other controller that I didn't data log today but was also pumpin' out around 320W for quite some time.  Most of that power went directly through the inverter and to the loads as the battery was full and just floated in the afternoon, filling in for the odd cloud or bit of haziness. 

Tomorrow's set to be another bumper harvest day too! 

Great news Outtasight, did you know you can now qualify for FITs if you had grid tie and an Ofgen meter? See:

http://www.navitron.org.uk/forum/index.php/topic,10035.0.html

cheers
Paul

[Outtasight]

Hmmm... had a look though a couple of the referenced threads.  It's a good change for Sean and a few others but the likes of Justme, Billi and me won't get anything out of it, being off-gridders.  I've got the same issues as them in that sometimes I resort to a mains battery charger in the winter to save the bank from rotting and that power is then recycled though the inverter to loads, so a few kWh here and there aren't "renewables" but would go through the total generation meter none the less.  I keep tabs on the grid charger with a separate kWh meter attached to the socket it leeches from and then for my own records, I subtract the charger kWhs from my total generation meter kWhs so I know the net PV generation.

Today was another 4kWh+ day though so I'm still as happy as a person with free hot water 

[billi]

being off-gridders

 

Never would like to miss what i learnt so far ( i am not a technical advanced person   ), but independancy fourced me to that route  

Billi

[Outtasight]

Ahhh... missed a bit... of wiring that is.

Having upgraded the whole path way from the junction box outside to the TriStar and the fuses on the other side of the new 100A battery disconnects with 25mm² cable, I'd thought I'd cured my unintentional heating elements... But I'd forgotten to upgrade the bit that goes from the fuse to the battery.  I have two fuses, one for each controller but both were still connected with about 50cm of 10mm², soldered into a single 8mm ring terminal at the battery post.  One is fine as it only has to carry 15A at most but the other is now seeing peaks of 40A and getting a bit warm.

I didn't have any more of the car jump lead, having used it all up already but luckily, I had a bit of 16mm² that I picked up at a car boot last year. So I made up a new link for the 15A controller with that and used the two existing links combined together for the TriStar fuse, giving it 20mm². 

Seems to be running cool now, although following the pattern it'll probably be the turn of the battery to overheat now from over-zealous charging currents  . 

[Outtasight]

The last couple of days have just been outstanding for clear blue skies!  In the last 24 hours to noon today, I used 5.3kWh of solar power (mostly on water heating and the fridge)... Just shy of 47% of my total electrical power needs by solar today. Over the whole week to Sunday I averaged 22.6% of requirement.  The combined array made a peak power of 1.36kW today (some way off the design peak of 1.44kW but it seems to drop off as the panels heat up compared to those very frosty days).

The cheap-o no-brand (Chinese) "mono" 80W panels were noticeably poorer than the Sharp and BP poly panels (putting out only 32.0V @ 3.88A in the heat of the noon sun today, compared to 33.0V @ 4.5A - 4.6A for the good brands).  The cheap-o panels Isc was on the money at 5.1A but their Vmp is lower than the other panels so they struggle to put out power in the combined array (dragging the overall Vmp down).  I have to look at upgrading the blocking diode on them to super low Vf Schottky types to claw back some of the voltage loss compared to the other panels.  I'm also guessing the black cells actually get hotter in the sun than the blueish ones and this also makes them put out lower volts under the same conditions...

Even so, I made more than enough power to keep the water heater and the fridge chugging the whole day and keep the battery topped off so it's now not so important that I optimise the array to the n^th degree.

Roll on the summer! 

[Outtasight]

I was (well still am actually) considering getting the RSC EIA485 adapter for the Sunsaver charge controller.  This would allow connecting it to the TriStar and bridging the Ethernet IP data to it, so that I can data log both controllers from the PC upstairs.  Nobody's got them in stock though and they cost about £60 (a bit of a rip-off for a serial port buffer).

While at a car boot sale (the first of the season locally) I spotted just what I needed for my aged Toshiba laptop for it to do the business downstairs...  A pair of old laptops for £5!!!  Actually, I didn't want the laptops but one had an ancient PCMCIA 10Mbit Ethernet card in it (the Toshiba can't use modern CardBus types).  As it turned out, the two laptops actually worked and could both run XP so no doubt they'll find some use around the house.  It's cheaper to pick up old PCs than buy PIC controllers from Maplins 

Now with the Toshiba able to read the two charge controllers and log them, I could measure total system power.  The Toshiba is a useful laptop as it's an old 33Mhz Win98 machine that can run on just 11-14W of power (11W with the display off) so it runs from the unmetered AC socket on the inverter but doesn't consume too much to do it's work.  In fact, as the generating day comes to a close, I unplug the AC adapter and it runs from its internal battery for a couple of hours.  I managed to find a replacement battery for it on a "laptop batteries R Us" type site and so the machine works as new now.  The laptop cost me £5 but the battery cost £25! 

The logger software can't do the arithmetic to combine the separate figures for charge current but that's easy to fix in Excel as the logger creates simple CSV text files.  I can transfer the days data to the PC upstairs and add a couple of columns to the data (adding the currents together and then multiplying by the battery voltage to give system power).

time (UTC)                                 Battery (V)   Current1 (A)   Current2 (A)   System Current (A)   System Power (W)

2010-03-09T10:30:30.090+00:00   27.29           10.82            3.94             14.76                       402.9

2010-03-09T10:30:45.090+00:00   27.29           12.97            4.61             17.58                       479.8

2010-03-09T10:31:00.080+00:00   27.33           14.13            5.52             19.65                       537.1

2010-03-09T10:31:15.080+00:00   28.33           35.77            11.15            46.92                      1329.3

The Morningstar MSView software can then open the modified CSV file and display graphs of the real and calculated variables: