Your situation does not sound very different than mine. I just completed my first month under EV-TOU2 and it is working out very well for me. I just posted some numbers here, but the executive summary is that with TOU, my 3.1 kW system is on track to produce enough to cover 600 kWh of monthly demand, split about 55%/45% between baseline usage and EV. In this first month, if I had been on the tiered plan I would have had a $15 bill. Not bad right? With TOU, instead I had a $46 credit that will be used in winter when the array won't produce enough.

Offhand, I'd guess that a 4 kW - 4.5 kW system would be enough to cover your bill under TOU, but there are some caveats. Where in San Diego are you, and what is your roof orientation and pitch? EV-TOU2 has a very solar friendly 12 pm -6 pm peak period right now, which is likely to turn into 2 pm - 9 pm at some point in the next few years. I can share a spreadsheet that shows some whatifs, although it is helpful if you download your "green button" data from SDG&E's website so that you have a reliable consumption history to model.

Welcome to the neighborhood.

Since you ask:

What's your anticipated alt./azimuth ? A south facing, min. shade 5.4 kW system will put out something like 9,000 kWh/yr. around here.

Things will be changing w/ rate reform, probably soon.

If you have T.O.U. and charge the EV at super off peak, and a good part of the rest of your billing will be at semi peak or off peak, particularly in the winter, cost effective solar may be a hard sell as a good part of your billing will be at about $0.20/kWh or so at current rates.

I'd start with the T.O.U. SES tariff. I'd estimate hourly use patterns with the EV thrown in, load it onto a spreadsheet along with the SDG & E hourly rates, and figure an annual bill. That will take a spreadsheet with 8,760+ rows (not as bad as it sounds). Add the PVWatts hourly output estimate in another column and multiply that by the same SDG & E hourly rate for an estimate of long term ave. annual revenue ( at current rates). Run the system size up/down until the expense = the revenue. Adjust as you think future needs and rate guesses dictate.

Depending on your use pattern, you may find that T.O.U. can allow a smaller system to produce as much revenue as a larger system on tiered rates. The bite on that is a smaller system may cost a bit more /Watt.

If/As you get more experience, and you have some west(ward) roof space, you may consider splitting the array - most south, a few west - to take advantage of the favorable bias T.O.U. billing adds to a more, but not entirely west facing layout. You'll be looking for the particular south/west split that maximizes revenue for any total size. That's a bit easier with micros, but usually possible with string inverters with a few extra considerations for string size and inverter type. Each array will need its own PVWatts run, but after a few runs, and some judicious choices helped along by experience, things can converge fairly rapidly.

LG will probably be more cost effective than Sunpower. SWAG on my part, and an educated guess given some life cycle costing/est., if a S.P. system can be found for about a 5% price premium on a $/Watt basis, it may be worth it. More than that, not worth it. I'd suggest not putting much stock in the S.P. "Most Efficient" hype. That's an area efficiency and useful as long as you place a high value on roof real estate. Actually, in terms of bottom line life cycle $$ efficiency - cost effectiveness - S.P. is way down the list.

Get your roof inspected and serviced if/as needed. A PV system can last a long time. Give the roof under it a better chance of a long, leak free life that's as trouble free as the panels can be. You won't regret it. Cheap insurance.

Edit: Check out the other T.O.U. rates as well - just watch the hour/day diff. for peak/off peak/weekend rates.

Take what you want of the above. Scrap the rest.

Thanks for the fast replies. Some more info. I am just west of the 5 in Carlsbad near the beach, the likely roof placement would be mostly south (approx. 12 panels) with the remaining on the west side. Forgive my ignorance of pitch, but it seems to be a standard roof pitch, not sure how I would measure it to be certain. Same with finding the altitude and azimuth. Also for Sensi, I have the xml of my usage, but it only includes the last two months since my EV and TOU started as a true baseline. The remainder would be my typical non EV usage for the last year. If you know of an easy xml converter please let me know. My xml skills are limited.

Any suggestions for an installer would also be greatly appreciated. My initial quotes are over $4.

Cmarschi:

For my part, you're welcome.

Check your PM's. I don't think you can respond yet.

With that low usage, I figured you were somewhere the ocean. Morning fog may shift an optimum orientation a bit to the west.

On tilt, sometimes it needs an inventive outlook common roof pitch is ~ 18-23 deg. or so. Avoid horiz. if possible due to poorer annual performance and increased fouling problems.

For azimuth, get a Google shot and measure off the screen w/a protractor for a decent swag.

Most of this stuff doesn't need super precision.

I'm on the conservative side. Don't size your solar with TOU plan, anything could change year after years. I'll try to get at least 90% coverage. Solar will degrade overtime and there might be less rain with more AC usage. You might have a less production year compare with PVWatts (20 yrs average) For example, my solar output during May is about 10% less comparing with last year. It could be (+ or-) 10%. One additional kW isn't going to break you or hurt your wallet especially when the cost per Watt has been really low now. I have 2 friends just start their solar project with $3.20/Watt before inventive using LG 285W + Enphase M250 on concrete tile roof.

While respecting Silver's opinion and not directed toward him, but for other readers:

PVWatts is NOT a 20 year average. It uses either TMY2 or TMY3 data which is not average data, or something called SolarAnywhere data which is also time series data but still not average data. Reader's can spend 10 min. or so w/ the PVWatts help screens for further information on why it's inaccurate and perhaps misleading to call it average data.

Over a long time, say 5,10,20,30 years, systems that match PVWatts user inputs may well get reasonably close to PVWatts estimates as a long term average output. However, each day, month and year's actual output will be different from the PVWatt's estimate. As Silver correctly (IMO) points out, and I and some others have been writing for some time, any year's actual output can be +/- 10% or more different from the long term PVWatts estimate. Individual months can show more variation than that and individual days are pretty much useless as a comparison tool. Most of that variability is due to the variability of the energy input - solar radiation - which is as variable as the weather.

For the most part, the economic viability of most any solar, PV and other applications relies on a long time frame to have any hope of cost effectiveness. For that to have any chance of being reasonably close to reality takes long term estimates of possible energy input to a system. One tool to help estimate that input is weather data of the TMY or SolarAnywhere type that PVWatts uses. That data is not average in the usual sense of the word.

On system size: I'll be the first to agree with what Silver is writing that the future is unknown. However, I'd suggest another conservative outlook might see conservative as not committing too many resources to an unknown and seemingly very uncertain future as far as energy prices and availability are concerned. Gas prices are/were something like a buck/gal. less than a year ago. I'd bet not many of those cynical about energy prices as an ever increasing type situation may have been surprised. Same for CH4 Prices. Those who bought/leased EV's solely for economic reasons may have been (temporarily) disappointed. Things change - both ways.

At the end of the analysis, it's a matter of opinion as to the future and each person's perception of that future. More information and understanding of the situation, and the strengths, weaknesses and the limitations of the available tools increases the probability of a better set of choices, but guarantees nothing.

Kind of a do your homework and take your best shot thing.

Take what you want of the above. Scrap the rest.

One consideration with two exposures is to do just the south exposure now, keeping the cost low. Especially if the EV late night charging rate is low.

Realize that under $3/watt pre-incentive will be common this decade. At some point this decade SD will have too many solar installers.

I've lost track of how long net metering is expected to be available in your area.

As of now (or at last a month ago), the SWAG on the 5% cap expiration is ABOUT 12 months or so +/- some, MAYBE. As of 05/30/2015, the allocation was at 3.37%. Update may be out over the next day or so, usually issued around the end of the month.

There will probably be some form of net metering after the 5% is reached, but details are not firmed up at this time. Stay tuned.

I don't understand this suggestion. Imagine you have a stack of panels in your driveway, waiting to be installed. Each panel is an independent revenue producer... it doesn't care what any of the other panels around it are doing (except for maybe a minor thermal effect). As each panel goes up the ladder and looks for the roof face that will allow it to generate most revenue, the decision will always be the same... one roof face will be best (or maybe two faces will be tied). While I agree that a SW oriented face will do best for most TOU plans, if you have 2 panels SW, that is not the same as one panel S and another panel W.

Some numbers, using SDG&E's EV-TOU2 plan, and assuming the array is 12 panels x 260 W, with available faces at 180 deg and 270 deg. The PVWatts model was run with 92111, San Diego TMY2, premium panels, roof mount, 18.5 deg tilt, 8% loss. Also considered is a proposed change in the plan to make peak from 2 pm to 9 pm, instead of the current 12 pm to 6 pm (all else remains unchanged).

S panels / W panels / Annual credit $ from current plan / Annual credit $ from proposed plan
12 / 0 / 1554 / 1346
11 / 1 / 1549 / 1342
10 / 2 / 1545 / 1338
9 / 3 / 1540 / 1334
8 / 4 / 1536 / 1330
7 / 5 / 1531 / 1326
6 / 6 / 1527 / 1321
5 / 7 / 1522 / 1317
4 / 8 / 1518 / 1313
3 / 9 / 1513 / 1309
2 / 10 / 1509 / 1305
1 / 11 / 1504 / 1301
0 / 12 / 1500 / 1297

As far as I can tell, with the existing net metering system, installing on multiple faces is only beneficial if the best face fills up, or as a hedge against even more dramatic changes to the TOU structure than what I've tested here. If the choice is between 180 and 270, 180 still wins for me. For orientations that are skewed (say, 160 / 250), it may be closer and SWW will quickly overtake SSE.

With regard to the net metering cap, the numbers are updated monthly in this thread, and it looks likely to me that the SDG&E cap will be hit before this time next year.

My suggestion was that since T.O.U., as far as SDG & E is concerned, is believed to shift the optimum revenue azimuth, and since most roofs are not optimally oriented with T.O.U. pricing, and since PVWatts runs take about a min. or less per run, that it might be possible to come up with, in effect, a "blended" azimuth that is a mix of say, south and west facing portions of 2 arrays instead of one south or west facing array with the hope of maximizing revenue. As you have confirmed, for as long as it takes, it seems a worthwhile exercise.

Well, I wasn't sure how the #'s would work out, but I believe you got the sense of what I was getting at. While I'm not sure, I suspect the results will look more sensitive to azimuth as the tilt angle increases (With horiz. obviously showing no diff. and vertical probably showing the most $$ change as f(azimuth)). I'm initially a bit surprised, but much less so once I think about it for a min. or two, at the relative insensitivity with respect to azimuth at what is a fairly common, but at the same time, fairly shallow tilt, but I believe your #'s.

BTW, I'd like to point out, that your findings, again, which I believe, tend to put a nail in the coffin or at least raise some doubt on the idea that west facing arrays are better revenue producers when used with T.O.U. tariffs than south facing arrays, at least in your neighborhood and at a common tilt angle.

I appreciate you input. Honest.

Thanks, this is timely info for me and saves me a lot of time wrestling with similar question.

This is interesting. Doubling the tilt to 37 deg, and staying with the 2 - 9 pm peak...
all south = $1364 credit
all west = $1306 credit
(incrementally moving panels from south to west linearly drops to the all west value)

Proving the main concept, though...
all southwest = $1419 credit

The falloff from southwest to west looks steep...

The problem I have with precise analysis is that , IMO, solar electricity in five years in southern California won't be worth much. What residential incentives will be in place for solar when the wholesale price is near zero? I agree with the idea that, by 2020, solar won't be installed without storage. Net metering simply can't last.

I see inverter choice being particularly tricky for the next 6-12 months. I see enphase as roadkill.

I obviously agree it's interesting - and somewhat complicated, not only because optimum tilt is f(azimuth), but T.O.U. time/rate and different weekend rates for some schedules will affect array revenue. I'll think some more on the situation.

My current educated guess is that as tilt increases, westward orientation becomes less deficient in ENERGY output (saying not much or anything about REVENUE). This makes sense given the geometry since solar zenith angle increases as the solar az. angle distance from south increases. For example, by my reckoning, a vertical, 270 deg. az. orientation beats annual ENERGY output of a vertical 180 deg. az. by about 25-30% around here. However, a 20 deg. tilt, 180 az. beats a 20 deg. tilt 270 deg. az. by about 10% or so. Given that, one other way to look at this is that given a fixed azimuth, for any location (lat.) there is an optimum tilt, which is actually a trivial result - kind of like "no kidding, really??".

Until I have clearer thinking on the subject and/or do or see some number crunching, if anyone asks me about shifting off south to take advantage of T.O.U. time shifting for revenue, I'll say that, speaking of T.O.U. revenue considerations alone, and aside from shading situations, it appears that some inquiry on the subject suggests that unless the tilt angle is relatively high - probably much higher than practical or common around here - south or mostly south is still best, or at least a 270 deg. az. is not as good as some less informed folks on that bandwagon would think.

This one puts a lot of balls in the air to juggle and is not only way off topic, but perhaps/probably of little practical interest to most readers. Maybe you or I should start another thread after a bit more thought on this and in the meantime, get out of this one. Fun for me, but a bit inconsiderate on my part anyway.