Cobbled together PV in W.Sussex

[Outtasight]

Hiya everyone!

Here's where I'm at today after just over a year of playing with some solar panels in my back garden.

It all started because I got turfed out of my company office and they said I had to work at home (no biggie as I work in IT and just need my laptop and a internet connection to earn my crust).  Besides, I can work in my pyjamas at noon and nobody is none the wiser 

I baulked at providing the electricity to run my work laptop at home and there was a sale on of cheap amorphous panels at Maplins, so I thought, "How hard could it be to run my laptop from solar power?"...



From the top left there's:
1x BP 3160S 160Wp (35.1Vmp)
2x Sharp NE-80 80Wp (17.3Vmp) in a series pair

From the bottom left there's:
6x TopRay 15Wp amorphous (17.5Vmp) grouped in 3 parallel pairs
8x TopRay 12Wp amorphous (17.5Vmp) grouped in 4 parallel pairs

There's more on the garage too... 



From the near end there's:
2x Kyocera KC40 40Wp (16.9Vmp) in a pair
2x Sharp ND170E1F 170Wp (23.2Vmp) in a pair
6x TopRay 15Wp amorphous (17.5Vmp) grouped in 3 parallel pairs

The big 170Wp Sharps are on one Morningstar 15A MPPT controller and the rest are on a second identical controller that parallel charge 4x 110Ah 12V leisure batteries (yeah, I know, but they were really cheap) wired for 220Ah at 24V.  The noxious gasses get piped outside by fish tank tube that fits into the common vent holes on these batteries... Lucky, as they bubble noisily like a witches cauldron when it's really sunny 





I've got some DMMs that read battery charge Amps on embedded shunts (made by just sticking pins into the 10mmsq feed wires at measured distances) and a SmartGauge (see http://www.smartgauge.co.uk/smartgauge.html) to accurately read the bank state of charge in %.  A small variable lab PSU, that has been wired to work from the 24V supply rather than mains, powers an assortment of bits and bobs with any voltage between 3V and 15V (mostly an AA/AAA battery charger and a car phone charger - the DMMs run on rechargeable AAA cells).  These particular DMMs are often on sale at Maplins for just £7 and run for weeks on a charged set of AAAs even when turned on 24x7.

I've got a couple of 12V LED lights in the room with the solar gubbins but the rest of the house runs off a Cotek SK1000 24V 1kW pure sine inverter with a remote control and load / battery monitor in the living room. I also have a home-made wireless remote (fashioned from a cheap wireless doorbell!) upstairs so we can turn the power (and so the lights) on without having to stumble downstairs to the inverter controls.  You can tell if the switch is working because it still goes "bing-bong" 

Another embedded shunt on the inverter provides info about net battery charge / drain with a little mental arithmetic from the other two meters.  The batteries and inverter are all wired up with 35mmsq truck jump-start cable.  It was really hard to solder... had to use a PipeMaster plumbing soldering iron.

Finally, a pair of cheap plug-in kWh meters measure yield and Winter charge from a grid charger to stop the batteries rotting. It's only a 3A 24V electric bike charger that I bought at a car boot sale for £2, but it does the job and I haven't had to use it since February .



My system is still pure off-grid but with the house lighting circuits switch-able between grid and solar by a change-over socket. The grid socket (right) is just fed from one of the old 6A lighting breakers and the solar socket is plumbed into the inverter via the plug-in kWh yield meter and a RCD breaker.  The two lighting rings are just paralleled in the fuse box and go to the 13A plug to select the power source.  I've not had to run the house lights on grid since February 



The rest of my solar appliances run on a completely separate "ring main" that consists of some semi-permanent trailing sockets round my living room, into the kitchen for some low watt appliances and upstairs to the computer room.

Back in June I broke 100kWh offset from the grid and in July I broke 5% lifetime offset as a percentage of all electricity I've used since starting measurements on December 8th. In June/July my weekly offset averages were 9.9% of all electricity consumed (solar plus grid). On a daily basis I might get up to about 15% or 2kWh.

My goal is to be able to run the computer room 24x7 off-grid. I work at home so the internet router, wireless, and mail/firewall/browser/file&print PC is a challenge.

I'm also thinking about converting half my array (the 35V one) to run on a SunnyBoy 1100LV grid tie inverter but I'm not sure what implications that will have for my metering and whether I have to contact the utility company or whatever.  It seems easy enough to install - just plugs into a grid AC socket and I'd just have to swap the DC wires over from my existing battery charge controller but I've got one of those digital (LCD) utility meters and I don't know if it will read right if I export power through it.  Some meters run forwards even when power is going out of the house and that won't do  .

For those interested in the gory details of the past year of tinkering to get to this point you can see it at another forum I hang out in.

http://www.mrsharkey.com/forum/vwtp.php?t=542

My family are all solar mad...  My in-laws in Japan have "proper" grid tied solar systems.  My sister-in-law has a 3kWp system









My father-in-law has a 5kWp system





He's also got a air-water heat exchanger called a Sanyo EcoCute ("Cute" being a play on the English word "cute" and the Japanese word for hot water "yu").  It uses a reverse air conditioner to suck heat out of the hot and humid air outside the house and pump it into a big insulated water tank.  It uses electric but much less than if you used a traditional resistive heater.

[Justme]

Sweet.

Tell me more about how you make a shunt by sticking two pins in a cable?

You must be getting over 5kwh per day with that lot of panels if not more on a good sunny day (remember them?)

Justme

[wookey]

That's a fine array of random pannelage. Do you know how much the whole system as cost so far?

[EccentricAnomaly]

Well, I'm impressed.

Any comments on is the relative performance of amorphous and crystalline panels in bright and diffuse sunshine?

[Outtasight]

As for cost... I was keeping tabs on this but it's kinda evaded me in the last few months.

The amorphous stuff I bought new but at various discounts.  Whenever Maplins were doing a special on the panels I'd buy one or two.  I started with some single 12Wp ones that they were doing for £40-£50 (but I get these discount vouchers they send out so if I spent more than £50 I usually got £3 off one or £7 off two together).  The 15W ones came from 60W kits that they sell but I'd wait until they discounted them to £200 and then use a £15 discount voucher so they cost me £185.  So in all I spent about £1,070 originally.  But some kits came with 4A controllers and some came with 10A controllers.  Some kits came with batteries and LED lamps as well and I sold these on eBay separately and then sold the first four I bought (that had diodes in them so weren't so good for 24V series use) at a car boot sale.  So in the end they maybe cost me about £800.  Not so great for 276Wp of power.

Then I got serious... 

I bought the new pair of Sharp ND-170's and that cost £1,250 (with the 13m MC3 extension leads) plus about £100 to build the aluminium mounting frame (although I bought too much aluminium so there's stuff left over for other projects).  The Morningstar controllers (that replaced one of the cheapo 10A controllers) cost £186 each (and another £100 for the battery temperature probes and the serial interface to programme them).

The batteries were £296 inc. quick release wiring posts.

The wiring was mostly left-overs in my garage but the truck jump start cable set was £30 from Halfords.  Seemed the cheapest way to get red/black flexible 35mmsq cable in my town as they were 5m long each.  I'm still using the off cuts on the spare 1kW inverter as those bits are about 2m long and still have the massive crock clips on them.

The pure sine inverter was £540 (inc. the remote panel) but I ended up buying two as the first was 12V and then a few months later (when I'd got too much PV for the Morningstars to handle at 12V), I decided to up the ante and go 24V.  I've still got the 12V inverter set and was playing with it this week to see if I could run my fridge-freezer from it to up the utilisation of the PV array (which often tops off the batteries by lunchtime and then lots of power just gets wasted).  Here's my hairbrain idea for running the fridge-freezer from an auxiliary inverter (because the inrush surge is too big for a 1kW inverter to handle on it's own when also running about 200W of continuous load):



The SmartGauge was another £150.

Then I ran out of space in the garden for amouphous panels (they take up lots of space for quite low power outputs) and started buying used good quality panels from eBay.  I decided to do this on the results of the first 8-9 months of using cheap Chinese panels... don't do it.  Amorphous as a technology is ok if you've got lots of space and you buy good ones.  The thing with amorphous panels is that they have to be rated after "running in".  In the first few months of exposure to light they degrade in output but stabilise at a value and then they degrade at the same rate as traditional poly/mono cells.  An unscrupulous manufacturer ( ) might quote the panels as 12Wp or 15Wp but this is on the day you buy them and then after a few months in the sun, gradually they degrade to their stabilised value which seems to be 65-70% of name-plate in this case (hence why in the diagram I only rate the 35V array at 600Wp).

The Sharp ND-170's quite happily outperform the amorphous panels in all light conditions, even though they are traditional poly panels.  They have more cells so their norminal Vmp is 23.2V with a Voc of over 29V so in series they kick out 46.2Vmp into the MPPT controller (this is a must or else you'll waste lots of power) and I often see them putting out 340-365W into the battery.  That's between 100% and 107% of name plate even after losses through the wiring and MPPT controller.

Good panels (BP, Sharp, Kyocera) are more expensive for a reason.  They put out their name plate power and they are built to last 25 years.  I bought one used unbranded (read cheap Chinese) 40W mono panel from someone on eBay but it's a very low voltage one (Voc ~19.5V) and I've never seen it put out more than 32W.  Luckily it was only £65 and it does lonely duty topping up some of my spare batteries but nothing else.  It has the typical paper thin aluminium frame and waterproofing that looks like it was done by a bathroom fitter with a hangover.

The Sharp panels don't use any silicone bath goo to keep the water out (because even in a bathroom it rots within 5-10 years) and they use frames made of sturdy stuff (with anti-torsion beams and such to aid rigidity).  But watch out kids, there are some fakes out there claiming to be used "quality" panels because they know that used quality brand panels fetch the same prices as new cheap Chinese ones and are much sought after by the likes of me...

Fakes can perform poorly and even catch fire under partial shade (one I saw had tiny 3A bypass diodes fitted to what claimed to be a 175Wp panel with a 5A nominal output).

At the moment the going rate for used quality panels is about £100 per 40W (so I paid £200 for my pair of Kyoceras and £375 for my BP 160W panel and £455 for my pair of Sharp 80W panels - ok I got carried away on the bidding on the last one  ).  The Kyoceras are the oldest ones I have, being about 4 years old so plenty of life left in them yet.

My father-in-law has Sanyo HIT hybrid mono/amorphous panels on his roof and they probably are better than regular mono panels.  My sister-in-law has regular Sharp poly panels.

The amorphous panels I have do put out a lot of voltage in low light (but not much current).  The Sharp ND-170s easily match them in low light as they have such a high nominal voltage that they also work in overcast conditions.  In full sun, the whole array can put about 28A into the battery bank (actually the 35V array has overloaded the Morningstar controller it's on and on clear days it sits pegged at 14.9A as the controller has current limiting protection). In overcast (but not threatening to rain darkness) the whole array can muster about 3-5A.  The 35V array wins at the moment because it has (in theory) 676Wp installed but in reality under cloudy conditions, the Sharp ND-170 pair aren't far behind (I might see 2A from the Sharps when the 35V array is doing 3-4A).

As for the "embedded shunts"... it's very easy.  I just cut off the test probes from the cheap DMM leads and soldered them to brass drawing pins.  Then I pushed one into the cable (these are 10mmsq or 35mmsq so the pin can go right in and not come out the other side).  Then I pass a known current through the cable (with another DMM Ammeter in series) and just poke the other drawing pin into the cable a bit further down until the mV reading on the new "Ammeter" is equal to the known current.  You might have to have a couple of goes to find the right length from the first pin but the holes are small and you'll cover them up with electrical tape soon anyway.  The DMMs all can read DC Volts down to 0.1mV on their 200mV range. It's not hugely accurate as I just calibrate the new meter to read 1mV per Ampere (giving a resolution of 0.1A). On the fat 35mmsq wire, I calibrated for 0.1mV per Ampere (giving a resolution of 1A for the inverter that can run at 46A at full tilt so I don't care about anything less than 1A accuracy).  The latter meter reads with the decimal point in the wrong place (0.1 = 1A and 1.0 = 10A) but that's not a problem.

Why didn't I just calibrate the 35mmsq cable the same as the 10mmsq cable and have it read 0.1A to 200A (as the DMM can read to 200mV)?  Simple... the voltage drop on a 35mmsq cable is very low and the pins are about 20cm apart to read 0.1mV/A so to get 1mV/A I'd need the pins 200cm apart and the inverter wires are only 90cm long!

This method has two advantages... it's cheap.  And it means I don't have to cut the wire and insert a commercial shunt that would mean contact resistances and more joins.  I usually solder the probe wires on to the pin when it's sticking in the fat wire as this melts the plastic insulation around the pin and so makes an air tight seal.  Then I just use electrical insulation tape to cover up the heads of the pins (you can see the inverter shunt on my battery pic as two bits of black insulation tape on the positive cable going to the inverter with grey wires going off to the DMM on the wall above.  The wires to the DMM can be tiny as they carry next to no current (just measuring Volts). It's ok to do this kind of embedded metering at 24V but don't try it on a 600V DC feed or the mains output of the inverter or else bad things are gonna happen. 

[kristen]

"I can work in my pyjamas at noon and nobody is none the wiser"

I work from home too, but it never occurred to me to dress for work!

Great account, and thanks for posting.  Have an Applaud!

Do the panels on the garage roof shade the ones behind them (particularly in December)?

"I baulked at providing the electricity to run my work laptop at home and there was a sale on of cheap amorphous panels at Maplins, so ..."

So the panels cost you less than the electricity would have? 

Now then ... I am self employed. I wonder if my company could buy panels for the house instead of paying for the electricity. (I know the answer to that, its based on area-of-house-used-as-office, but that's no use to me). I'll have a word with my accountant about the disproportionate amount of use the PCs and Servers use, whereas for Light and Heat I'm stuffed by the area-used-as-office proportion.

[sjaglin]

Hi,

Really detailed and exciting post! I found the Japanese houses very nice too, it seems that PV is quite developed there then?

Do you keep a tracker of your production?

Stef

[stephend]

Really detailed and exciting post! I found the Japanese houses very nice too, it seems that PV is quite developed there then?

Second only to Germany for installed PV: http://en.wikipedia.org/wiki/Photovoltaics#Worldwide_installed_totals

[Justme]

 I'll have a word with my accountant about the disproportionate amount of use the PCs and Servers use, whereas for Light and Heat I'm stuffed by the area-used-as-office proportion.

You need a better accountant. Whilst there are accepted "norms" for what you can claim, in fact you can claim what you like as long as when / if you get an inspection you can justify / convince the inspector your right. With your heating for instance, If you was at work the heating would not be on at all. As you are at home you need the heating on, If you dont have any zone controls its all on. Thats all business use (just the same as they dont make you claim 1/5 the mileage on a car cos the other 4 seats are empty & so not in business use) . Same with lights & elec. When we had a VAT inspection as long as I could justify what I had claimed for they were fine. Our fuel costs % of biss to personal use were much much higher (like 80+% biss) than normal for the industry I was in. But as I could show them how far & how often I went places it was fine.

Justme

[dhaslam]

If the solar panels are only used to power office equipment then they would qualify as being "wholly and exclusively" for business use.   In the longer term PV will be less expensive so the house portion would be less expensive if done later.     PV gets very interesting at  one euro, or less,  per watt  and this sort of price may be commonplace within five years. 

[EccentricAnomaly]

Thanks for the remarks on amorphous vs crystalline.  Having now read your Mr Sharkey thread I see you touched on the matter there but didn't really make a specific comparison.  Pity the whole matter is confused by the overrating of the cheap amorphous panels but it's also interesting what you say about using them at high voltages.  I think amorphous are traditionally used at higher voltages so maybe that's part of the reason people think they work better at low light levels.

Does anybody know of a source of reputable/properly rated amorphous panels?

My thinking is that crystalline is probably better for any sort of retrofit because of their smaller area/watt but on a new build where you can design for suitable space on and around the building the lower cost per watt and supposed somewhat better low-light performance of amorphous is attractive.

It's interesting that you use a number of charge controllers.  I worry about that from a number of points of view: that the charge controllers will trip over each other in controlling the charging process and also that they tend to throttle back the output of the PV array when the batteries are charged rather than putting the power to better use.  Does anybody know of an MPPT which isn't a charge controller: that is, one which just throws as much power at the battery as it can for whatever the battery voltage is at that moment?  I imagine that chargers like the Morningstars can be programmed to do that just by setting the various cut-off voltages high but it seems a pity to pay for and have potential unreliability from functionality which is not being used.

[stephend]

Does anybody know of a source of reputable/properly rated amorphous panels?

All of the below have the required IEC61646 certification.  Some are guaranteed to 80% at 25 years, others 20 years.

• Kaneka k60 or GEA60 about 2 euro/watt (they also have a new hybrid panel which looks very interesting but could be pricey)
• Sulfurcell www.sulfurcell.de
• QS Solar 1.67 euro/watt, although there were recent announcements that they've dropped further
• solartechnics www.solartechnics.de 1.4 euro/watt just a few months ago, these seem to be rebranded QS Solar panels
• Inventux: www.inventux.com don't know much about these.

Reasonably priced suppliers:

• www.b5-solar.de
• www.nature-solar.de
• www.ibc-solar.com
• Solartechnics sell direct

[Justme]

It's interesting that you use a number of charge controllers.  I worry about that from a number of points of view: that the charge controllers will trip over each other in controlling the charging process and also that they tend to throttle back the output of the PV array when the batteries are charged rather than putting the power to better use.  Does anybody know of an MPPT which isn't a charge controller: that is, one which just throws as much power at the battery as it can for whatever the battery voltage is at that moment?  I imagine that chargers like the Morningstars can be programmed to do that just by setting the various cut-off voltages high but it seems a pity to pay for and have potential unreliability from functionality which is not being used.

Each controller will work on its own relative to the settings it using. You need sep controllers where the Vmp's dont match very well.

All chargers throttle back as the battery fills. Its that or over charge the bats. The thing to do is have a system that monitors the charging & when that stage is met to turn loads on to use the excess power. You can do this with a FM80 Outback controller or an controller with a suitable system. You could even set the inverter to do it with a Victron unit (virtual switch). I think its been mentioned in here about using every electron if not try the Outback forum.

Oh & its the battery that controls how much charge is accepted not the controller. Power is drawn not pushed.

[kristen]

"You need a better accountant. Whilst there are accepted "norms" for what you can claim, in fact you can claim what you like as long as when / if you get an inspection you can justify / convince the inspector your right. With your heating for instance, If you was at work the heating would not be on at all. As you are at home you need the heating on, If you dont have any zone controls its all on. Thats all business use (just the same as they dont make you claim 1/5 the mileage on a car cos the other 4 seats are empty & so not in business use) . Same with lights & elec. When we had a VAT inspection as long as I could justify what I had claimed for they were fine. Our fuel costs % of biss to personal use were much much higher (like 80+% biss) than normal for the industry I was in. But as I could show them how far & how often I went places it was fine."

My Accountant has agreed with the revenue a number of things that surprise me, but are justified on the basis that "You need to have that resource available to be able to do such-and-such when you need to, so although you only use 10% its OK to claim 50% [say]", so I reckon he is on the right tack generally. There are issues about capital gains on the property, which will be based on the proportion of office-as-business claimed, which I am keen not to over-egg.

However, I have not had a specific discussion with him about how the fact that Electricity use is disproportionate [between home and office] and how that might impact payment for Wind/PV installation, and I'll be interested to see what he can agree with HMR&C

[Justme]

Yes the CGT is a big issue now for home workers.