Creating power for your home, off grid.
Emphasis on nuts-n-bolts, hands-on projects.
I knew these things have high surge starts but didn't figure it would be so high... Nearly 1,100W! The inverter beeped and the power gauge went right off the scale for a second or so before settling. If I were to run the fridge from my inverter I'd have to not run any other loads as the start-up surge plus any other loads would blow a fuse (or worse).
So, I'll have to content myself with running the upstairs firewall / browser PC on solar power. It also uses about 110W (140W with the screen on and the extra hard disks whirring) but at least it doesn't surge so I can use my work laptop and the small kettle at the same time and still keep within the 1,000W continuous power limit.
You'd probably have to run at least a 3kW inverter to be able to run a fridge-freezer reliably and allow for the compressor surge while using other things at the same time. Just an ordinary 2.2kW kettle running at the same time that the fridge decides to start up would still cause an overload though...
I'm going to run a single extension into the kitchen though (just by the counter) so that we can use the small kettle in the kitchen and while the fridge is a no-no, I did discover that the rice cooker only uses 400W for about 20 minutes so that would also be a good candidate for solar powering.
We're currently enjoying a run of clear blue-sky days (up to 27C today) and the battery bank is bursting at the seams by lunchtime so I'm getting about 1.8kWh per day out of the system (up to 15% of our daily power requirement).
They must have originally been used for some telemetry or something as they came on heavy frames which looked like they had been U-bolted to a post. Anyway, they made useful 25 degree angles and are heavy enough to stop the panels blowing over so here they are on the roof.
Despite being over 4 years old, they still put out about 82W in series in the strong sun we had yesterday.
It's amazing how much smaller they are than the amorphous panels for the power.
Sadly, the amorphous panels are degrading much more than they should. Amorphous panels naturally have a degradation curve that is initially much higher than that of traditional silicon wafers (called the Staebler-Wronski Effect or SWE). Panel manufacturers are supposed to declare the power output for amorphous panels after at least 100 hours of irradiation. This is the minimum time they are generally considered to have stabilised, after which the output declines in line with traditional wafers over a number of years. Some unscrupulous makers (and now I know that Shenzhen Topray Solar, who make mine are amongst the worst offenders) over-state the power rating of their panels before stabilisation. So what is described as a 12W panel is only 7.7W after a few months in the sun. They have been banned from the Kenyan market, where they consistently underperformed compared to other amorphous panel makers. The report here makes interesting reading...
http://www.humboldt.edu/~aej1/aSiKenya_ ... ly2_05.pdf
But these panels are now finding their way into the European and US markets through sellers like Maplins and Amazon that carry their kits for caravan / camping users.
I've now got 24 panels made by this company and after 6-9 months irradiation, the average output is only 64% of the stated maximum output. I know now that this is the case because the Sharp and Kyocera panels put out their rated power (or higher) under the same lighting conditions. The Shenzhen Topray Solar panels actually degrade much more than the SWE would allow for (usually no more than 10% degradation to the "stabilised" output). This seems in line with the reports findings of 2005 so their products don't seem to have improved since those tests were made.
I'm not going to return them though as I got them so cheaply and I can't return a "whole kit" as I've sold off the bits that I didn't need on eBay. Plus, the thought of the argument I'd have to have with the customer service department is not worth thinking about - "No sir, there's nothing wrong with these panels - it's just not sunny enough" and so on. It's just annoying about the amount of space they take up for the power output.
From now on I'd rather buy used polysilicon panels from known reputable makers (like Kyocera) than buy any suspicious Chinese panels (even if they are dirt cheap).
Ahh... well. You live and learn.
I spotted another bargain on eBay today... A pair of Rolls Surrette RB/12CS11P 12V 357Ah batteries on sale for 500 Pounds. They're about 715 each when new and I almost put a bid in until I realised there's no way on Earth I can move the things! They're 125kg each. I had enough trouble lifting the 35kg 110Ah batts I'm using now into my car...
- Dennis The Bus Dweller
- Seasoned Nomadicista
- Posts: 1861
- Joined: Mon Nov 15, 2004 9:33 am
- Location: Southold N.Y.
Scouring the pages of eBay, I've been lucky to find some gems amongst the sea of cheap Chinese panels that are flooding the market.
Well, apart from this first item, which is an unbranded and probably cheap Chinese panel . But it was only 67 Pounds and a useful 40Wp mono-crystalline panel and seems to work ok. I bought it from a Hungarian dude who was building a massive power plant in the shed in his garden. He'd just bought a bunch of 200Wp panels (hence selling the little one) and had an old 3kW UPS and batteries coming out of his ears. In fact, he had so many batteries that he threw in a bunch of little 7Ah SLA batteries for free with the panel... He was a real mad inventor type.
I then started looking for another one to make a 24V pair but instead found a guy selling a used Sharp 80Wp panel. I had to pay a bit over the odds for that one (257 Pounds), having got carried away with the bidding. Still cheaper than a new one though. So, I reconfigured the two 40Wp Kyocera ones as an 80Wp parallel and put that in series with the new Sharp. When I went to collect it he said he's got five in all and so I might try to bid on another one as he's selling them one by one on eBay. The Sharp is 17.3V and the Kyoceras are only 16.9V so there are some mis-match losses in that string.
There was another half-price sale on those test meters a couple of weeks ago and so I bought one more to read the battery voltage. Happily, a Maplins shop has finally opened in my town so I don't have to trek 25 miles to the next nearest one. I can cycle to the shop now to get diodes and wotnot.
Then I struck gold (err... solar panels are actually blue but you know what I mean). Someone on eBay sold me this BP Solar 160Wp 35.1V panel for just 375 Pounds! This one hadn't even been used. The guy had bought it and then propped it up in a store-room at home, forgot about it and then found it when moving house. He put it on auction at 375 and nobody else bid for it. I think most people are only really interested in buying "12V" type panels and this one is only good for 24V systems or you have to use a MPPT controller. Happily, I'm using both!
Sharkey, you'd like this guy... He was an ex-airforce engineer and drives a 2.4 litre Audi diesel that he makes his own bio-diesel for. He collects used vegetable oil from fast food outlets and then processes it through two refining plants. The first plant he bought but the second one he made from scratch, copying the design of the first one! This guy is a real mad inventor's mad inventor .
As I now had nearly 500Wp of solar power on this one 400W controller (even allowing for the amorphous panels being only 64% of their rated power), I decided to take down the four oldest of the amorphous panels on the wall. These ones are the least well suited to 24V operation anyway as they each have a blocking diode built in so in a series pair they suffer from two diode drops in voltage. Off the wall, they actually performed quite well, still putting out over 700mA in short circuit and at least 10W (12V into a 10W auto bulb) so maybe they aren't all that bad after all - just not suited to 24V battery charging. I might try to sell them or keep them - for any other project. I've still got a couple of 10A controllers and a mountain of 12V SLA batteries and a gaggle of 12V inverters kicking around doing nothing
I put the new BP panel above the patio doors as a sort of roof. I got the idea from a German solar panel maker that sells solar porch roofs. The BP panel wasn't too deep (78cm) so could be supported without vertical poles to the ground (just wooden stays) and was coincidentally 160cm long so was just about the same width as the double sliding doors so it even looks like it's supposed to be there (or at least that's what I convinced my wife was the case )
It's also got two sets of mounting holes on the long edge so in the Winter I'll make shorter stays to change the panel angle to be steeper. The stays are fixed to the panel and wall by hinges so I can unbolt it from the panel and fold up the "Summer" stay when not in use.
I bought a second 80Wp Sharp from the same guy. He put it on eBay and this time I had less competition for it and secured the second one for just 205 Pounds.
Now that I had another 160Wp set, I decided to put them both on the wall with the BP one.
I lose some ambient light generating capacity from them not being able to "see" the whole sky above them and the late afternoon brings a shadow from the roof and wall as the sun creeps round to the NW side of the house but it looks neat and does mean that the DC wiring to the controller is much shorter than if it was on the garage roof. I reconfigured the two Kyocera 40Wp panels back to series and they can stay on the garage as the much lower current means less of an issue with voltage drops on the overhead line.
Here's a neat angle shot:
I can just about clean the BP one from my wife's study with a mop from the window. The two Sharps will require a ladder.
All the mounting hardware was just some brass hinges to make flexi joints to bolt on the panels with 8mm bolts and then screw the other side of the hinge into wall plugs.
Now I've reached a limit again... The Morningstar 15A controller has maxed out in medium / strong sun and is applying current limiting. It does mean that even on overcast weather I get almost double the Amps of the other controller that just has the 340Wp Sharp pair on it. Combined, I can make at least 5A even on a grey day (provided it's not actually raining)... just about enough to run the computer room upstairs without using any battery power.
Then I spotted something on the Morningstar web site... From September they are going to start selling a MPPT version of the TriStar 45A and 60A controllers. These will have much higher voltage limits than my controllers (150Voc) and will be able to charge 12, 24, 36 or 48V batteries with up to 3.2kWp of solar input (for 48V). It also claims to have the highest conversion efficiency of any MPPT controller at 99% in a fanless design. The MPPT15 manages 97.5%.
Check it out:
http://www.morningstarcorp.com/en/suppo ... 14.091.pdf
Hahahaha... I'm locked into a vicious spiral of more panels, bigger controllers and bigger batteries.
I also reached another milestone this week. On the lifetime offset statistic (total solar kWh as a proportion of total solar+grid kWh consumed since last December) I've reached 5%. On a weekly basis, I'm averaging 9.9% for June and July... Which is, co-incidentally, about the amount that the news was saying that residential power bills will go up by 2020 due to CO2 taxation and energy companies clawing back their investments in renewables in the UK.
I've been pondering what to do with the excess power that my solar arrays make on very sunny days. It gets to about lunchtime and the battery bank is about 75% full and so the charge controllers start to throttle back to their absorption power levels. Good for the batteries but a waste of lots of solar power that can't be shoe-horned into them any faster.
Up until now I've gone upstairs and switched the computer room over to solar power. But on a good day there's still power left over. So I wired another extension to my wife's study and she then runs her laptop and desk lamps for craft work on solar power... but there's still power left over (who'd have thought you could have too much solar power )
So, I started looking at the big loads that I'd discounted in the past as "difficult" to run on solar power.
First up was the fridge freezer. Uses about 1.5kWh per day and presents a fairly manageable load when running - about 125W. Better than that, it only comes on for about half the time and consumes about 5W when idle, so it's really a 60W load when averaged out over a long time.
The reason why it's a "difficult" load is that you have to watch it like a hawk because if the power fails (dead battery) all your food will defrost. Also, fridge compressors take HUGE amounts of start-up power (maybe 10x the running power) so despite being a 125W nominal load, you need at least a 1kW inverter to get the thing going.
My main solar system has a 24V bank of batteries and the 1kW pure sine inverter that feeds my work laptop, the house lights, etc.
With the usual suspects turned on, the main system inverter sits around about 200W output constant load. I once tried running the fridge-freezer on it with nothing else plugged in and it just about worked. It overloaded for about half a second (more than 1150W on the power display) and it beeped. But after that it sat at at the bottom of the scale. I gave up on it because the inrush current meant that I couldn't risk running anything else at the same time.
Then, one especially sunny day, I decided to see if I could run the fridge-freezer and the other loads at the same time.
I had a few spare bits kicking around: A really old 100Ah battery (that just refuses to die even though I often neglect it in the garage for months), the spare 1kW pure sine inverter (left over from when my main system was 12V) and the 30A variable lab PSU that I use for charging batteries, starting small cars, blowing stuff up
Here's what I did:
The fridge-freezer only needs 125W when running, so the main solar inverter can deliver this with the other constant loads it has to supply. The fridge-freezer only needs the 1-2kW surge for a fraction of a second and the 1kW inverter can deliver a 2kW surge. So I used the spare inverter and battery as a buffer. The 12V battery is only there to act as a big capacitor. In fact, I considered trying one of those 1 Farad car hi-fi capacitors instead of the battery (but I happened to have a battery kicking around).
Once the fridge-freezer is running, the 30A PSU runs the load from the house solar mains (drawing about 9-11A @ 13.8V, so it can recharge the 12V battery at the same time). I'm using an old deep cycle battery but a car battery would do fine as it only has to deliver a "starting" current before going back to bed... floating to 13.8V after a couple of minutes recovering from the "start".
Obviously, the wife wasn't too happy about a big boat battery next to the fridge in the kitchen but it was just an experiment for a day.
One thing I discovered was that if the inverter can't muster the power for the surge then the compressor can fail to start and I did have this happen once (while doing the dishes in the evening). The charger was sitting at 15A and the fridge was doing nothing. The compressor had stalled on start-up, drawing 180W while not moving (hope nothing was damaged). I cycled the power and the fridge started up as normal after a couple of minutes.
It almost made it though the night on its dedicated battery (quite a feat for the 12 year old & often neglected battery). After sun-down, it ran on the house bank via the charger and then at about midnight I turned off the house inverter (with the bank depleted to 60% DoD).
At about 4am, I woke up and had a look at it... still working. The compressor was running. About 6am I woke again and this time the battery had died with the inverter locked out on a low voltage alarm. I fired up the house inverter and the charger sat pegged at 30A for about 30 minutes (with the compressor running) but I had to stop it as the house bank was now drained down to 80% DoD. The fridge carried on with its own battery and I let the sun start to charge the house bank.
So I managed to run the fridge off-grid for a whole 24 hours but the inefficient conversions meant I killed all the batteries.
Next time on this thread... The water heater.
So next there's the water heater problem... Uses about 4kWh per day heating up a tank of water from 22'C to 60'C for our bath.
The solar system makes lots of spare power in the afternoons but I've only got a 1kW inverter and besides, the array only makes about 1kW with all other loads off... remember we want to just find a use for excess power.
Some charge controllers use dump loads for diversion regulation but that would require a separate tank to the normal one upstairs and a special 24V heater element. The standard immersion heater element in the tank is 230V 3kW.
It would seem there's no easy way of getting the spare power into that water tank.
One way to solve the problem is to throttle the power to the 3kW heater - effectively using a "dimmer" on it like a light bulb. Only you can't get 3kW dimmers. So I thought about the supply side. If I were to reduce the voltage to the heater to 115V, it would reduce the power drawn by the heater. The heater has a cold resistance of about 20 Ohms and so at 115V it will only draw 5.75A (or 660W). Perfect!
I thought about running a separate wire up to the heater so that I could feed it 115V AC. Trouble is, that would mean running a special 115V inverter as well as the 230V one.
Then at a car boot sale the other weekend I saw this:
A 4kVA tool transformer. Very rusty and neglected but the guy sold it to me for Â£15 (assuring me it hadn't been used for scuba diving).
Inside it wasn't so bad and worked ok. The transformer itself is potted in cement of some kind. Made it very heavy but at least it's quiet - doesn't even hum noticeably.
These things are usually a regulation bright yellow colour to denote they are 110/115V (230V isolation transformers are blue). I didn't plan on selling it or using it on a building site so there was no need for it to be restored to it's former colour so I sanded the outside down and gave it a generous helping of black Hammerite (wot I bought half price at a supermarket clear-out a few months ago - 'cos "It'll come in handy sometime later").
I couldn't find anyone locally that sold 32A "commando" plugs (most transformers and extensions come with 16A plugs/sockets these days) and I actually needed a regular 13A 230V socket for the heater so I just drilled a hole in the thing and connected up a spare extension lead (through a cable grip grommet).
And here she is... Incredibly, it fits in the airing cupboard. Fired up, you can see it powering the water heater with about 630W.
Today, it was on for 1hr15 and used 0.76kWh and raised the tank temperature from 28.4'C to 34.0'C. I gouged out a hole in the foam insulation and stuck the outdoor sensor of one of those "in/out" digital thermometers against the tank wall (about 1/3rd the way down from the top of the tank) so I can monitor the thing. After glueing the sensor in place I repacked the hole with foam and even found a short bit of pipe insulation to snap over the hot water pipe that exits the top of the tank.
Being a 4kVA transformer, it can handle running at 650W indefinitely and didn't even get warm doing so. I was worried that it might overload the inverter with its inrush current but it doesn't seem any worse than my hi-fi amp with its massive toroidal transformer that makes the house lights "blink"...
Now all I need is a way to make the solar charge controller automatically and wirelessly remote control the transformer so that when the sun goes behind some clouds, the dump load is disconnected... That and I need my new 3kW inverter that is out of stock for a couple of weeks. With the dump load running we can't put the kettle on at the same time on my 1kW inverter or it overloads and trips out.
I'm glad people are getting something out of this. I now cross post to a UK forum where there's lots of home-made bodgery going on.
I masquerade as "Outtasight" over there (on account of having changed my car to a Honda Insight).
This last weekend has seen probably the last of the really sunny days of the year here. Today it was so dark and rainy all day that the battery bank ran right down. So, it's probably just as well that I got me another huge haul of PV and batteries at the weekend!
Another eBay job. This guy was selling three of the same Sharp 80W panels as I bought a couple of months back. I won two of them in a last minute (literally) bidding battle but even so they were very cheap at only Â£340 for the pair (I paid Â£455 for the last pair).
When I arranged to pick them up (from some 250 miles away!) the guy asked if I be interested in some other kit (namely another four unbranded 80W monocrystalline panels and 16 deep cycle batteries!). I told him I'd take a look - it'd be rude not to
Got there and the Sharp panels were ok but somehow he'd managed to put big dents in both the frames - still they were very cheap. The unbranded ones were not bad (a bit flimsy frames but the output is fine). And because they were unbranded he only wanted Â£150 each for them so I said I'd take two. Even the super cheap Chinese ones that you see on eBay now are about Â£200+delivery each so two for Â£300 was a good deal.
Then he showed me the batteries... WOW!! What a find! They were Deka 8GGC2 180Ah 6V gel batts - proper solar ones. These are rated for 1000 cycles at 50%.
Most were dated 2008 but I managed to dig out four that were almost new (just five months old, May 2009). I gave them a load test with a 60W lamp and my voltmeter to see what they were like open circuit and then with the load on... Read 6.50 / 6.46V - These things really don't self discharge much! He let me have them for just Â£50 each, pretty good considering they go for $250 each (about Â£150) new.
So, here are the new panels, washed down and sitting propped up on the plastic chairs... another 320Wp added to the pot.
It's getting a bit crowded in the patio area
The solar controller was already pegged at 14.8A so I couldn't get any more through it so I just wired these panels into a 24V nominal 2x2 series parallel string and wired that into the terminal blocks for the mains charger - direct to the battery. With that, I could make about 34A of solar power in the afternoon sun - enough to run the new 640W dump load and still have almost 6A charge current left over to finish topping the battery or run some of the computer gear.
Just need some chunky links for the new batteries now.
The terminals aren't particularly well placed for a 24V bank but if you arrange them in a circle you can use short links and the output terminals can be any two adjoining (+) and (-) posts.
180Ah is a bit smaller than my notional 220Ah pack now but the cheap flooded leisure batteries are showing their weakness after being cycled to 50% every day for over a year now. They're probably only rated for 200 cycles at 50% but these Dekas are rated for 450 cycles to 100% discharge and 600 cycles to 80% discharge.
That, combined with the higher charge acceptance of these batteries should make the whole system more efficient. Flooded cells take about 130% charge to complete (you have to put 130Ah in to get a 100Ah charge stored) whereas these gel cells are supposed to only take 110% charge to complete. They charge at a lower voltage (2.35V per cell) and don't need (or rather are damaged by) the EQ charging that flooded cells do.
The very low rate of self-discharge should also help, as flooded cells can lose 10% per week doing nothing at all whereas the gel cells lose a claimed 2% per month. That's an extra 21Ah (0.5kWh) a week you can use for AC generation rather than battery maintenance!
Then I discovered that SMA make a low voltage SunnyBoy that can work on my 35V strings (so I could swap between battery charging and grid tie with just a DC switch).
I quizzed a couple of vendors but they said that the SunnyBoy 1100LV had been discontinued (leaving the normal high voltage SB1100).
Then the main distributor said that they can convert a WindyBoy 1100LV to work with PV instead of turbines. Turns out the two models are exactly the same but at the factory they set a firmware "switch" that makes it either work in turbine mode or MPPT PV mode. Why you can't set this in the field with a DIP switch I don't know.
I started looking into the rules and regulations on grid tied PV in the UK but there wasn't a lot of info available from my local company and some regional distribution network operators (DNOs) have been difficult to date.
Then there was the cost... The SB1100 is a 1.1kW inverter but for less money I could upgrade my off-grid inverter to a 3kW one.
If I bought the SB1100LV it would have been nearly maxed out with the PV I have now, leaving not much scope for expansion. Grid tie also means that in the future you are still vulnerable to power cuts as the anti-islanding software in the inverters means that if the grid fails, your inverter shuts down too (for grid safety).
With the state of under-investment in UK generating capacity, I figured it was more useful to have a backup supply of power. In recent rain storms (and not very severe ones, just very wet, not even thunder or windy) we had quite a lot of brown-outs that caused the computer room UPS to cut in.
So, with my new haul of PV and batteries, I decided to stick with my off-grid route and called the Antares guys again for a 3kW inverter. They were out of stock for a month .
Then this week, on my last day before a weeks leave from work, I happened to be working in Reading (near their store) and decided to give them a call to see if they'd got any stock yet. I got lucky and they had one!
And so here it is:
It was quite easy as the technician calculated that my short (90cm) 35mmsq power leads would still be up to the job. I had thought I might have to double up to 70mmsq (as it mentions this in the manual) but he said it was ok as the manual assumed up to 3m long connections. So I was able to just connect it straight up using the existing 8mm bolt rings. This model has nut and bolt fixings rather than 8mm machine screws that allows tighter clamping of the ring terminals - handy when the full rated DC current is 138A. I uprated the three in-line fuses from 30A to 40A each (the original fuses I had for the 12V inverter). Should be plenty for now but if I can find 60A fuses, I'll put those in. The manual recommended a 200A fuse in the DC circuit.
The remote panel is common to all the 24V models and so all I had to do was plug it in. The LED power display on the panel reads % power output so it automatically adjusts to each model. Now each step on the ladder is just 300W load instead of 100W load.
In this picture I was just testing the install. Since then I've re-wired the output using the hardwire terminals inside the front panel. You just undo four screws and the output panel comes away, revealing the terminal block and earth bonding post. I used 1.5mmsq (good for 16A) cable as far as the distribution block (via the RCD and generation meter) and then re-wired the extension to the kitchen with 1.25mmsq cable (good for 13A). By the fridge, I upgraded the worktop single socket to a twin (it was getting annoying having to swap the toaster and kettle plugs all the time). I'll probably run another socket down to near the fridge power point so I can run the fridge on solar on good days. Now I don't need my wacky power buffer system any more . This inverter has more than enough muscle to start a compressor (up to 6kW surge and 3.3kW for 3 minutes).
The original battery bank isn't powerful enough to run a 3kW load for more than a couple of minutes and even at 2kW (our usual kettle, not the solar one) the battery bank sank to 23.2V when delivering 90A.
This inverter has a higher threshold for when the fan comes on. Up to about 150W load it doesn't use them at all. Above that but below about 250W one fan comes on. Above that, both come on.
Now I've got two spare 1kW inverters... The 12V one is doing duty running the 24V battery charger from the old 12V battery for emergencies when the main bank is getting dangerously low. I suppose I could try to sell the spare 24V one but it's also maybe useful to keep around as a spare in case the main one develops a fault.
If the 3kW inverter idle load of 1.5A is too much to bear in the depths of Winter (1.5A x 24hrs is 36Ah of battery!!) then I might have to use the small inverter as a Winter converter (it only draws 0.65A x 24hrs = 15.6Ah).
You can never have too many inverters, that's what I say .
Users browsing this forum: No registered users and 1 guest