Planning new house, planning on designing as efficient of a house as we can afford. Will be doing solar. Main question is, would it be better to avoid paying to have a gas line put in (pg&e) and just go all electric appliances. Part of the savings from installing gas line could go to a larger solar system, better insulation, and more efficient appliances heating cooling etc. House will be around 2500 sqft in nor cal, will spend more on cooling 105 in summer vs 50 in winter.

The cost of gas in in constant flux; the cost of fuel for your PV system will be the same in 100 years as it is now: -0-; so yes; you are better off spending that money on a better envelope. I say better "envelope" rather than more insulation because insulation is but one of the important factors in the building envelope. You'll need to understand the factors in building a high performance house - most of which are typical of Passive House. I'd suggest starting with Joe Lsiburek's book: A Builders Guilde To XXX* Climate. *There are three climate zones in Northern CA: Cold, Mixed Humid, and Hot-Dry; find the one closest to your house. Joe is a principal in Building Science Corp; their website has tons of freely available information on building good buildings.

I should have specified but yes, a better envelope is was I was meaning to say. I will have to take a look at that book.

It really depends on the size and quality of the house you build. Where the balance of heating and cooling tend to cooling, a heat pump looks more attractive if you have local support.

http://www.eia.gov/dnav/ng/hist/n9190us3m.htm

Typical dry zone N.CA 99% outside design temps are in the mid-30s to low 40s and 1% outside design temps in the high 90s to low 100s. A high-R house will have a heating/cooling balance point of about ~50-55F, which makes it very much a cooling dominated situation. There's no real rationale for hooking up gas, since it's going to save AT BEST 10s of dollars per year of marginal operating cost over going with heat pump solution to heat & HW, and that's even without considering the low lifecycle e cost of solar compared to standard grid rates in CA.

With the current direct subsidy & rate structure incentives available in CA for solar the lifecycle cost of PV power is well under the cost of residential retail power. In CA there will soon be significant incentives for on-site battery storage as well. Cooking with induction ranges is very efficient, but introduces high peak draws, which may eventually incur a "demand charge" on the billing to pay for grid infrastructure, but that can be offset significantly with a modest amount of on-site battery with smart controls. Heating hot water with standard electric HW tanks is inefficient but not super-high draw, and newer models will now capable to (and be compensated for) providing ancillary services to the grid operator in the form of demand response & frequency control under the recently passed Energy Efficiency Improvement Act of 2015. (See: http://www.greentechmedia.com/artic...er-heaters )

Alternatively, from a raw efficiency point of view, there are now heat pump hot water heaters operating with an EF greater than 3, which pulls ~2/3 of the heat going into the hot water out of the room air, reducing the air conditioning load, putting the heat purged from the house into the tank. (See: https://www.energystar.gov/productfinder/product/certified-water-heaters/details/2230929 )The peak draw of a heat pump water heater is a fraction of that of even a standard electric tank and an order of magnitude lower than that of a tankless HW heater, thus limiting demand-charge issues arising from hot water use. The total energy cost is less than 1/3 that of a standard electric tank, when you account for the cooling load reduction.

There are now several mini-split heat pumps with a cooling efficiency SEER north of 30 (more than 2x the current minimum efficiency requirements), which would reduce the size of the solar array required to get you to Net Zero Energy (which will be code-min in CA starting in 2020 when the updated Title 24 regs become active.) If you're not Net Zero at the time you want to sell this place you will be at a market disadvantage, but since it's financially rationale to build to that standard NOW on just the energy cost issues, it makes sense to roll all that in at low mortgage type financing rates, and never have to worry about it going forward.

Given these factors, what would be the point of hooking up to the gas grid?

The mini split is interesting, I will have to look into that more. It is hard to understand how a ductless system would be able to cool/heat a house efficiently that way.

With modulating mini-splits using temperature set backs usually works against you, since they have to run full-speed during the recovery ramps,

How hard it has to work during recovery depends on how much the interior temperature changed (not much in a very well insulated house at typical loads) and the ramp rate that you can live with. Dumb algorithms might do otherwise.

It would be interesting to see data on the efficiency of using setback to avoid running during the hottest/sunniest (cooling mode) and coldest (heating mode) times of the day vs no setback at all. Steady interior temps negate the benefits of internal thermal mass.

Insulation even at code minimums makes a surprisingly small difference in cooling loads- it's primarily about solar gain, not conducted heat.

Letting it modulate with load is with few exceptions more energy efficient ( and WAY more peak-draw efificient, from a demand-charges point of view), since a large fraction of the cooling is delivered during a cooler part of the day, when efficiencies are even higher. Efficiency benefits from both the cooler outdoor temp, and much higher efficiency if of running at part load. See the cooling efficiency curves with both modulation & outdoor temp in this third party testing of a couple of representative 1-tonners of a handful of years ago (the newer versions are more efficient still, but the in efficiency across modulation level & temp hasn't changed.

http://www.nrel.gov/docs/fy11osti/52175.pdf

In particular note Figure 14- at 85F outdoor temps it's nearly 3x as efficient at low speed than it is at max speed, 1.5x as effiient at mid speed. Even at 95F it's 1.5x as efficient when cruising rather than sprinting (the results for the Mitsubishi unit weren't as consistent, and the authors postulated that there may be some defect with that unit, since they could not reproduce the manufacturer's HSPF & SEER test data.)

Under the current rate restructuring proposal by the CPUC all investor owned utilities will be running time of use rates starting in 2019 (see: http://www.utilitydive.com/news/cpuc-rate-reform-proposal-two-tiers-tou-rates-possible-fixed-charges/389914/ ) which also affects the cash-efficiency. During the middle of the day, before the peak AC load or the peak grid load, the local PV would be providing most of the power.

So far CA has steered clear of residential demand charges, but other states are flirting heavily with the concept.

http://www.natlawreview.com/article/colorado-puc-considers-distributed-energy-storage-challenges-and-opportunities

One utility in AZ is already been allowed to apply demand charges for solar customers (whether net zero electricity or not):

http://www.azcentral.com/story/money/business/2015/02/26/srp-board-oks-rate-hike-new-fees-solar-customers/24086473/

If CA takes the demand-charges route modulating or even pre-cooling rather than setting the temps higher during the middle of the day makes even more sense, since it maximally leverages the peak PV output (which occurs before the cooling peak, and well before the grid load time of use peak) and minimizes the absolute peaks, as well as the peak time of use energy charges.

It's also important to note that using temperature set-back strategies REQUIRES oversizing of the equipment to have reasonable recovery ramp times. While a modest amount of oversizing is good for efficiency, it takes a substantial oversizing factor to get reasonable cooling ramp times in a low load house of even typical thermal mass, let alone a high mass low load house.

Thermal mass helps deliver steadier interior temps, which would actually allow you to get by with an undersized AC system without much loss of comfort. But it does nothing for the efficeincy of a mini-split.

The fuel savings in any moduling appliance, more especially in those designed to reduce latent heat, are in setting and forgetting the controls.

Fixed setback is nearly obsolete strategy adopted by the forced air industry and finally the power companies.

The future is in smart thermostats where appropriate.

We use WiFi Honeywell, Nest and Ecobee 3, but our new Fujitsu runs on factory control.

Smart thermostats that are allowed to be remotely controlled by a third party and aggregated into a cash compensated demand-response program can also be a way to minimize cost.

But aggregated demand-response is not available everywhere, and until/unless the circuit court's pushback on FERC 745 gets overturned allowing demand response programs to participated in capacity markets, the remuneration of demand response won't be nearly as good.

http://www.powermag.com/ferc-order-745-and-the-epic-battle-between-electricity-supply-and-demand/

This may be going WAY off into energy policy geekdom, but it matters. Today's utility business models won't be viable in another 10-15 years, and there's a whole lot o' movin' & shakin' goin' on in the mean time.

...or even pre-cooling rather than setting the temps higher during the middle of the day makes even more sense,...Thermal mass... does nothing for the efficeincy of a mini-split.

But such an efficiency ($/btu) improving setforward/setback strategy depends on thermal mass.

It's also important to note that using temperature set-back strategies REQUIRES oversizing of the equipment to have reasonable recovery ramp times.

Unless you aren't at design day loads (the usual case) and therefor already have lots of excess capacity. Then recovering from setback might push you from intermittent low speed operation to even more efficient continuous low speed operation. There is no need to rush a 2F (or with enough mass, .2F) recovery.

Classic residential setback is going away, but setback controlled by smart algorithms (ASHP COP optimization, time-of-day energy cost, PV output, peak loads, weather prediction, etc) will certainly increase.

All houses have thermal mass.

Turning the AC on after that thermal mass is above the setpoint imparts a cooling load that is whatever the instantaneous heat gain factors are PLUS whatever the thermal mass of the house is emitting/convecting.

And even if the AC can bring the air down to temp in a reasonable amount of time, human comfort is about the average radiant temp, not the air temp. As long as the mass of the house is still radiating more heat onto the humans, it almost doesn't matter what the air temps are.

Pre-cooling during the middle of the day when there is surplus electricity and possibly turning the AC off during the peak grid load period is a cost-savings feature, not an energy savings feature. Grid peaks tend to lag the cooling load peaks by a few hours, so the rate at which a pre-cooled house heats up when the AC is interrupted is relatively slow, and will have less of a comfort hit.

The algorithms by which mini-splits set the modulation rate are proprietary, but when at 2F from the setpoint they are definitely NOT loping along at minimum power.