Looking to run home office off grid

[KingCnut]

Hey all,

Look forward to reading your comments and other posts.

My APC UPS gives me wattage readings of between 175 and 250 watts total office load. Everything is currently running through the UPS with the printer peaking my power consumption to 250 watts while in use. I would like to DIY sustain 400 watts (for growth) 24/7 on a solar array. Can I do this with (2) 80 watt panels and (4) Trojan 105 AH batteries?

Live in Houston area. Clear shot to sun all day.

I will install transfer switch for back up and 4 or 5 consecutive days of clouds.

Appreciate your thoughts...

KC

[russ]

Hi KingCnut - Welcome to Solar Panel Talk!

Let us see what kind of feedback you get from various members!

Again, Welcome!
Russ

[greenHouse]

I don't know how many hours of sun Houston gets, but we average 5.1 here in Austin. To backup a 250 watt load (assuming 24/7) with solar you'd need 6KWh / day of AC, derate by 77% gives 7.8KWh / day DC. Divide that by 5.1 hours (for Austin) and that's a 1,560 watt array.

[Sunking]

Hey all,

Look forward to reading your comments and other posts.

I would like to DIY sustain 400 watts (for growth) 24/7 on a solar array. Can I do this with (2) 80 watt panels and (4) Trojan 105 AH batteries?

Live in Houston area. Clear shot to sun all day.

No way Jose. Time to get realistic on cost and equipment.

OK 400 watts x 24 hours = 9600 watt hours. Houston winter insolation = 3.7 hours best case. So the solar panel wattage needed = (9600 wh x 1.5 efficiency factor) / 3.7 hours = 2600 watts

Battery capacity needed = 9600 wh x 1.5 efficiency factor x 5 = 72 Kwh or 52 Trojan T-105 batteries.

Estimate DIY cost = $20,000 If I were you I would just pay Reliance their 12 cents per Kwh and thank my lucky stars electricity is so cheap and plentiful.

[Sunking]

I don't know how many hours of sun Houston gets, but we average 5.1 here in Austin.

You cannot use your average insolation sun hours on a battery system, you have to use winter worse case or you go dark in winter. Further battery systems at best are only 65% efficient unlike 77% of a grid tied system. Realistically battery systems are 50% or less efficient over all.

[greenHouse]

No way Jose. Time to get realistic on cost and equipment.

OK 400 watts x 24 hours = 9600 watt hours. Houston winter insolation = 3.7 hours best case. So the solar panel wattage needed = (9600 wh x 1.5 efficiency factor) / 3.7 hours = 2600 watts

Battery capacity needed = 9600 wh x 1.5 efficiency factor x 5 = 72 Kwh or 52 Trojan T-105 batteries.

Estimate DIY cost = $20,000 If I were you I would just pay Reliance their 12 cents per Kwh and thank my lucky stars electricity is so cheap and plentiful.

You forgot to upsize the array to make up for float losses

The OP is simply being unrealistic in the first place. Designing for 5 days of "no sun" is very expensive, in part because a 50% design DoD means 10 days of batteries, and that's actually 96KWh, not the 72KWh you spec'd

Also, very few "no sun at all" days ever actually happen -- much less 5 in a row. What actually happens is "bad sun days" are intermixed with "good sun days", even in the low insolation months. One day you might get 2 hours of "full sun", and the next 4 or 5. As long as the system is designed to capture the sun when it gets the chance, without wasting energy floating batteries that aren't needed, the system can survive the "poor sun" months just fine.

[greenHouse]

You cannot use your average insolation sun hours on a battery system, you have to use winter worse case or you go dark in winter. Further battery systems at best are only 65% efficient unlike 77% of a grid tied system. Realistically battery systems are 50% or less efficient over all.

Yes, but using "average" was more than enough to show that a pair of 80 watt panels weren't going to do the trick.

As for the "at best only 65% efficient" followed by "50% or less efficient over all", that's true in the Great White North, it is not true down this end of the world. The conversion from DC straight off the panels, through the inverters, and out the other end as AC is about 80%. With Houston being as far south as it is, even the winter is relatively sunny, and that means fewer amp-hours make the array-battery-inverter trip and more go array-inverter directly. Schedule more loads during daylight hours and system efficiency goes up.

[Sunking]

The OP is simply being unrealistic in the first place. Designing for 5 days of "no sun" is very expensive, in part because a 50% design DoD means 10 days of batteries, and that's actually 96KWh, not the 72KWh you spec'd

Um 5 days reserve is the minimum which in real application only gives you 2.5 days reserve before you hit the 50% DOD mark. With 5 day reserve capacity only bears 95% availability which means you spend quite a few days of the year in the dark waiting for the sun to recharge the system after 2 cloudy days which is quite common in Houston area. To get to 99% you move up to 10 day reserve (5 day real ) which is rarely ever done as installing a LP genset to recharge the batteries in 4 hours is more cost effective and gets you better than 99% availability..

As for the 72 Kwh capacity is spot of for 9.6 Kwh per day. Battery conversion efficiency is 67%. so with 5 day reserve (2.5 day real reserve) is 9.6 Kwh x 1.5 efficiency factor x 5 days = 72 Kwh.

[Sunking]

As for the "at best only 65% efficient" followed by "50% or less efficient over all", that's true in the Great White North, it is not true down this end of the world. The conversion from DC straight off the panels, through the inverters, and out the other end as AC is about 80%.

You know better than that 80% is the best a grid tied system can achieve. In a stand alone battery system the battery charge efficiency is 80% by itself. Add all the losses up and you get to 66% using MPPT CC.

100 watts output from the panels, 2% wiring loss between panels and CC. 98 watts into CC and occur 5% loss to the batteries. 90 watts to the battery with a charge efficiency of 80% now puts you at 72 watts. From the battery to inverter another 1% which I will ignore and say 0% wiring loss. So now 72 watts at the inverter input assuming 95% top of the line inverter loss now puts you at 68.4 watts usable to the load. So best case stand alone battery efficiency is 68.4%.

My reference to realistic 50% overall comes from the fact you would have to utilize every single watt hour the panels can deliver. In reality that is highly unlikely unless you have constant predictable load 24 x 7 .

FWIW I live in the DFW area and design for TX, NM, AZ, and OK

[greenHouse]

Um 5 days reserve is the minimum which in real application only gives you 2.5 days reserve before you hit the 50% DOD mark

The OP spec'd 5 days, not "5 times the amount of power used per day so I can have 2.5 days of reserve."

Try sticking with what people ask about and I'll stop poking holes in your numbers when you tell me I'm wrong.

Also, at 5 days or 10 days or any number larger than 1 days of reserve, you start to get into Peukert territory and all the numbers go straight out the window. My bank is rated at 522Ah at C/20. But my discharge rate is much closer to C/40 or even C/50. Compensating for the reduced discharge rate, the capacity climbs to 620Ah or more.

Finally, that 67% is the full-cycle efficiency, not the efficiency of what comes out of the batteries. A fair amount of full-cycle losses are caused by elevated charging voltages relative to discharge voltages. A 100Ah battery bank, at rest, has already lost 10% or more of it's full-cycle energy, just falling from absorption level voltages to resting voltages. Consider that an amp-hour added at 57.6 volts (48 volt system) "costs" 57.6 watt-hours to add, but only produces 51.2 watt-hours if removed at 51.2 volts -- which is a pretty standard-ish resting voltage for a 48 volt nominal bank. Since the last 10 to 25Ah of that hypothetical bank all went in at the absorption voltage, you're losing (split the difference) 112 watt-hours out of 4,800 nominal watt-hours just because the absorb stage ended on just the absorb stage charged amp-hours. AND, they don't get lost a second time around ...

[Sunking]

Also, at 5 days or 10 days or any number larger than 1 days of reserve, you start to get into Peukert territory and all the numbers go straight out the window.

I am very versed with Peukerts LAW and that is why we design using C/100 specs, not C/20

[greenHouse]

You know better than that 80% is the best a grid tied system can achieve. In a stand alone battery system the battery charge efficiency is 80% by itself. Add all the losses up and you get to 66% using MPPT CC.

Right, but you don't lose on battery charge efficiency on power that comes straight off the panels and goes through the inverters -- which is what I said. That's why timing loads to run during sunlight hours makes a system more efficient -- you're avoiding the battery cycle losses by not cycling the batteries.

The most efficient use of a system is maximizing (within the system limits) loads during the daylight, and minimizing them at night.

FWIW I live in the DFW area and design for TX, NM, AZ, and OK

That's okay. I have clients on every continent but Antarctica, and I'm working on getting them there as well.

[greenHouse]

I am very versed with Peukerts LAW and that is why we design using C/100 specs, not C/20

Right, but most batteries are spec'd with C/20 not C/some-other-number which is often not equal to C/100 anyway. And for the OP's request, 5 days of no-sun is 10 days of capacity, which is C/240.

As I said, once you move away from one day, throw the numbers out because you're going to need to recalculate them anyway. And we ain't even touched on float losses, which in C/100 or C/240 territory can become the single largest load in the system.

[Sunking]

Right, but most batteries are spec'd with C/20 not C/some-other-number which is often not equal to C/100 anyway.

Then you do not know about Rolls Surrette batteries Here is a sample spec from their RE line. Rolls/Surrette RE line products are the real deal and come with 10 year warranties (5000 series) where the first three years is 100% and the last 7 pro-rated..

[Jonathan Cole]

The problem when people who are only technically oriented comment on such questions as this, is that they leave out the obvious economic issues which relate to efficiency and low-waste strategies that are required for affordable solar energy systems. In this case, greenHouse and Sunking have glossed over the obvious fact that if your use peaks out at 250 watts with the printer running (you did not say what kind of printer) then you probably are not using 250 watts 24/7. I run an office with computers, printer, scanner, DSL modem, etc but guess what folks, when I sleep, so can the machines. The computer in standby uses a few watts to keep the battery charged, the printer and scanner are on a switched power strip and modem. Solar is very expensive when you import the normal wasteful habits that Americans have become accustomed to. But then again its also expensive to be flushing your money down the toilet while paying the utility for the wasted energy....yet millions of people do this.

I suggest that you reframe your question. Get yourself a Kill-A-Watt meter (Ebay $30) and measure the exact usage that you need during operational hours. (The meter measure cumulative kwHrs over time) You can do this by plugging your UPS into the Kill-A-Watt meter. Then look at where you can reduce wasted energy by using timers or switched power strips to reduce the number of hours that your equipment is in operation or standby mode.

As far as the technical issues go, you can find your average peak insolation hours at my web site (see below) under useful links at the top right of the page. As far as the technical "he said, she said" stuff here, I would say that there is one substantial thing that I would disagree with. With temperature controlled deep-cycle batteries you can expect 90% in/out efficiency as long as you keep them between 40 and 85% state of charge. This is according to Sandia National Laboratories who have done extensive research on the matter.