The 1909 Franwall Solar Story

It started inauspiciously ...

with an email from some do gooder group announcing "yes, you can have solar power. Just call xxx". I think Solar City I had thought about installing PV panels, but not seriously. The house site was surrounded by trees especially on the south side. However, I figured a contact wouldn't hurt, so I sent them an email.

The email reponse was within a day. They asked for my address, and within a few hours sent me a link to a page containing a google map of the neighborhood. I clicked on my house and submitted. A day later the reply: "sorry, due to the trees we aren't able to install panels on your property" (not enough sunlight-- not enough electricity produced). I sent a joking response about mounting the panels on 50 foot struts, but otherwise forgot about it.

A few months later I decide to do an energy audit, only $100 -- Pepco (the local electrial utility) subsidized, (details). I went ahead with many of their recommendations, starting with the most cost effective (attic insulation)... and then thought that maybe their install solar panels from Sungevity might make sense. For I had been on the roof and thought "I don't want to cut down the trees in front, but the back North facing roof seems to be out of their shade".

But how to check? I looked online and found a service that will examine the solar prospects of your property, using google images, etc. And, as I expected, a North roof mounted solar system was reasonable (details)

Time to get serious and call solar installers. I called around; the big guys (like Sungevity) and some more local companies. And they all said, either immediately or after a site visit, that they were not interested in a north facing roof install. I also got contacts from one of the parents in my kid's bnei mitzvah group, a solar advocate named Scott Sklar. He gave me two. Both were much more willing to tackle this job, and I chose one that seemed to be the most interested: Kenergy of Takoma Park.

They did their own on site measurements (details), which confirmed the feasiblity of the project.

Time to design ...

(details) My goal was to maximize the average cost of PV power (installing enough panels to spread the fixed costs out, but not so many as to require panel placement in not-very-good spots). After some back and forth, I contracted for a 6.5kw nominally rated system: consisting of 25 made-in-Wisconsin Helios PV panels (the label!) (Kenergy typically uses German made Mage, but with a $10k fed rebate expected, I figured I should buy American). So at the end of June I am ready to go! Or so I thought.

A PV system basically has two components: the panels, and some way of inverting DC power produced by the panels to AC power used by your house's loads.

Roughly speaking, there are two ways of accomplishing this inverting.

Central inverters. This is an older technology. DC power from panels is fed into a central inverter, typically located in the basement near the circuit breaker box.
Micro-inverters. A relatively new technology. Each panel has a small inverter attached to it, that converts DC to AC and then sends the power to your circuit breaker box (I am abstracting from safety switches and combiner boxes...).

The costs of inverters tends to scale up with the size of the system -- directly so with micro-inverters (for central inverters it is more of a step function, since central inverters come in different maximum kw capacities). The costs of these systems are similar, with central inverters tending to be a bit cheaper. As an older technology central inverters may be more dependable (who knows how long these micro inverter will really last). And, central inverters tend to be more efficient at converting DC to AC (efficiencies range around 95%).

However, central inverter PV installations have a design issue. For several reasons (such as increasing voltage to reduce line losses), PV panesl are usually serially wired in strings. This makes them vulnerable to shade -- since the output of all panels in a string can only be as great as the worst (i.e.; most shaded) panel in the string! That means a branch partially shading just one of a dozen panels can cut the output of the whole array of panels by a lot!

Micro-inverter avoid this problem, since each panel is independent. There are ways of limiting this problem with central inverters, by having several inputs (so that each string is directly wired to the inverter), or by adding DC optimizers (at about a 20% cost increase per panel).

With an odd installation -- shade, or orientation at different angles -- these kind of problems are more important.

Fate intervened in the form of the July 1 derecho, which knocked out power for 4 days. Which brought home a limitation of most household PV installs: during power outages they don't produce electricity (details)!

Even though the total electricity consumption of a house is likely to be greater than the PV power produced, most of the time power produced from one's panels is not used. Power useage at night or during very cloudy days isn't provided from panels, and large loads (AC, stove, etc) can only partiallly be powered from the panels. Conversely, during the middle of the day when no one is home, only a fraction of the power from the panels is used.

What to do? One could just waste the power, but that would be kind of nuts. If one was off-grid, battery backup is used. But batteries are expensive (often greater than the cost of the panels).

Instead, most PV installs in urban settings use a grid-tie architecture. Basically, PV power is fed directly to the main circuit breaker box. From there, it either goes to loads in the house (lighting, AC, etc) or if more power is being produced than needed -- it backfeeds to the grid. This means your meter spins backwards -- your house becomes an electricity producer, not a consumer.

In a sense, the grid acts as an infinite battery, able to take whatever excess power you may produce.

In most states, the utility is obligated to accept PV power produced by households, and to credit this power to your electrical bill. In other words, they pay retail for the power you (the PV equipped home) provide to them. The utiltities don't like this very much, they would much rather buy it from you at wholesale!

Actually, this grid tie architecture does complicate the utitlity's power management regime. The utilities have to be adjust production to demand across the grid. They are used to doing that with their centrallized generating facilities. Adding a whole bunch of small distributed facilities (homes with PV panels) makes the balancing more difficult. So long as the fraction of power supplied by homes with PV panels is small, it isn't a huge problem. But should this fraction grow, it could become problemmatic (i.e.; how to adjust centrallized generator output as clouds move in front of the sun).

There is one concern: what happens during a power outage? If a bunch of PV panels continue to produce electricity and backfeed it to the grid, that means the lines (in the street) may be live during an outage. For utility workers, that is a major safety hazard!

Therefore, grid tie systems must have a safety feature: during power outages, no power is fed from PV panels to the grid. The simplest way to do that is to simply shut down PV production.

What does that mean? During the times when PV power would be very useful -- during outages -- it is not available! That seems wrong, but it is an inevitable result of using the grid as a battery. And since 99% of the time the grid is working, that's a tradeoff that can make sense.

Still, it does seem just wrong. It seems sensible to consider ways to finesse this problem, ways of building a hybrid system that uses both your own batteries and a grid-tie connection to the utility.

In Jan 2012 (after suffering 3 several day outages in 10 years) I had installed a generator with appropriate critical loads GenTran sub panel/transfer switch (manual), so I had power. But it was noisy and stinky and a pain in the neck. Maybe it is worth considering some kind of hybrid system that would provide power during an outage from combination of batteries, PV panels, and a generator? Such a system could also act as a most-of-house uninteruptble power supply. That means power to most devices (TV, computers, etc) would not be effected by transient outages. Thus, I wouldn't have to provide desktop UPS's to my several computers. That's a cost savings: these desktop UPS's wear out after 2 or so years, and I use several of them.

Interesting, but I didn't want to spend a lot on such a system. So I thought about it for a month, did a bunch of research (details), discussed some options with Kenergy ... and modified the contract in mid-August. Now, about 11kwh of battery backup would be added to the system. That meant changing the PV installation, and increasing the cost. We were ready to go -- 6 weeks or so for the installation (early October)!

Types of hybrid systems -- systems with both battery backup and grid-tie.

The basic idea is to somehow keep PV panel production going during an outage. This requires some means of integrating grid power, PV power, and battery power. And it means somehow safely keeping PV production going during an outage.

The basic idea is to seperate the PV power from the main circuit breaker box. PV power is fed into an intermediate device which can use the PV power in 3 ways: charge batteries, feed to a sub-panel, or send to main breaker box (and hence to the grid). When an outage occurs, this intermediate device can quickly switch to battery power (it "inverts" DC power from the batteries to AC power). This battery provided power is sent to a sub panel that supplies the most important loads; such as refrigerator, heating system, and basic lighting. This sub panel is isolated from the main panel, hence from the grid. Furthermore, the PV panels "see" the AC power from this intermediate device and will continue to produce electricity.

This means two additional components are needed. First, an intermediate device -- such as an off-grid inverter. Secondly, batteries.

Batteries are not cheap. A good quality 200 AH 12V battery costs $600. Four of these yields about 9.6kwh of power. Actually, these are usually only discharged to 50% -- which means something like 5-6kwh of power. If you are frugal, that's about 6 hours of power.

Details galore: I ended up choosing an AC coupled system using Magnum off-grid inverter and about 11kwh of batteries. This was minimal: I anticipate using my generator to charges these batteries at night. Still, if it works this setup provides UPS protection for much of the house, protection against short outages, and a great reduction in generator run time.

Articles describing AC coupling: Solar Pro, Renewable Energy World, or a Magnum white paper,
A table listing some options
Thoughts on design choices (July)
Thoughts on design choices (August)
A slide show diagramming some possible layouts
Thoughts on design choices (August)
A slide show diagramming actual possibilities (based on some of the equipment I finally chose).

And then things got slow.

The components arrived piecemeal, with panels delivered in early December

Installation was in dribs and drabs; with the rough install mostly done by February

By the end of February, most of the install is completed, but ...

While panel installation almost done, thes system was not producing power (Feb photos).

Wiring (grounding and checking continuity) was incomplete.
A battery enclosure is needed, but is not on site.
And what is up with the permitting?

A flurry of work occurred in Feburary and early March (early March photos). After a few bugs (i.e.; shorted wire from inverter to AC cutoff causing scary short circuit)... power was being produced!

But there were problems...

The back row of 4 panels on the flat roof was being partially shaded by the front row (the sun not being all that high in March. This degraded production from those panels, and hence the entire string
One of the strings was not producing at all.
I needed a battery enclosure.(since the wood one was unacceptable)

Basically, Kenergy failed. I did obtain a battery enclosure (with minimal Kenergy help), but they solved nothing else. So, I ended up cancelling. Time to find a new contractor. (details)

Why couldn't Kenergy do the job? My current thoughts are there were 3 complications, and that was too many. (details)

Replacing the contractor ...

It isn't easy to get someone to step into a mostly finished job. I tried the electrical contractor who had worked with Kenergy (Plus Electric). I got one response email from him, and then after I sent several replies -- nada. Who needs to deal with more of that?

Per usual, the larger companies had no interest. So I figured I would try Capital Sun Group -- the other company that had expressed interest in May 2012. After some negotiation, we settled on a cost plus contract, with a max of $9,200 (note that I owed Kenergy $5,000, so this represents up to $4,000 extra).

This included installing Tigo maximizers. From what I could tell, the uneven shading on the roof seemed to depress output, something that Tigo maximizers might alleviate. And, these maximizers provided the ability to monitor each panels output. Lastly, the controller for the maxiziers had a DC cutoff switch, something that was a good safety feature.

Capital Sun wasn't that fast, but just about the time I would get really annoyed, they would show up. By Sept 1 the PV panels were fully installed. Inspection, and Pepco approval still needed.

What work did Capital Sun do...

The non-functioning string was easy. It turns out that the Aurora 1 inverters require each string to have 200 Volts. Well, the non-functioning 5 panel string only achieved 150 volts -- not enough. The solution was to re structure the arrays into two 8 strings and one nine string array. These would easily reach 200 volts. Moreover, their input would rarely exceed 2000kw, the maximum any input on the (dual MPPT inputs) Aurora could handle.
Solving the shading. I would of tilted down the front (southern) row on the two 4-panel rows on the flat roof. But Capital Sun convinced me it would be more elegant to raise the back row. It was a bit problemmatic, but it did work. The problems stemmed from the torque of the elevated panels. Once raised, they tended to sway a bit -- they were not rigid. Worrisome, after a while they could get loose. This was solved by cross bracing the elevated mounts; a cheap and easy to install solution. In fact, we also cross braced the front row (that was not raised).
Now both of these rows are good and stiff!
Installing Tigos. Twelve dual and one single Tigo maximizers, a management unit (MMU), and a Gateway were ordered. It took a while to install them, and the MMU and gateway arrived late. Still, it was done.

PV install completed and legal...

Just because everything is working doesn't mean it's all done. County and Pepco inspections etc were still needed.

The first step is to get Montgomery County electrical and building inspections. And wouldn't you know, the original (Kenergy and Plus) electrical permit never mentioned the heavy up. So a new electrical permit and building permit had to be obtained.
The county needed the electrical site plan before inspecting. So we had to obtain a copy of the Kenergy plan from the county.
When the inspector came (for both building and electrical), he was stymied by the fact that the Kenergy plan described the PV installation, and the battery backup installation. But we had put all the battery backup stuff away -- it was complicating things too much, so the plan was to get the PV install fully/legally done, and then consider installing the battery backup. After the inspector called his boss, he was informed that dropping this portion of the plan was an acceptable modification (the PV portion was unchanged). So after looking things over, the inspector issued gave us an approval!
Not so fast... later that day the boss decided that a corrected set of plans was needed. Kind of stinks, changing his mind like that.
So Capital Sun (Carlo LaPorto) created some plans, I vetted them, and we went to the county to get them submitted (in early October). The county got picky, and wanted to see notations of what was changed. After some wheedling, they agreed that penciled in notes were good enough. Carlo left, I did the penciling, and submitted the revised plans.
Two days later they were approved. So I went back to the county and picked up the approved electrical plan (hi resolution) -- for only about $400! A cost of switching contractors (thanks Kenergy!).
A few days later the inspector came back, spent about 5 minutes checking the equipment out, and put issued the inspection approval sticker.

That means the installation is county legal! Almost done, still need Pepco approval -- you are not supposed to turn on your inverters (hence not supposed to produce PV power that is sent to the grid) until Pepco approves.

An application for a Certificate Of Completion
And three weeks later (end of October 2013), the final approval.
Which means: the PV installation is complete and legal

I can now turn on my invx:erters, and use the electricity from the solar panels, or sell it back to Pepco (and watch the meter run backwards)!

Solar Renewable Energy Credits ...

Solar Renewable Energy Credits (SRECs) are designed to help utilities meet regulatory requirements to produce a specified fraction of their power via renewable means. The utilities can achieve these requirements either by producing solar electricity themselves or by purchasing SRECs from other parties producing solar energy.

What does it mean to sell an SREC? The SREC is not the same as a watt of electricity! Rather, it is a fact about how the watt was produced. Thus, the watt and the SREC are seperate.

Which means that I can consume a watt of energy produced by my solar panels (say, but running this computer), or by selling it back to the utility (say, on a sunny day where I am producing 10x more power than my house is consuming). The SREC can be sold seperately-- to whatever utility might need credits to meet their regulatory goals. Or, I could choose not to sell the credit -- to not sell the fact that I produced a watt of electricity using renewable means.
The typical small producer (i.e.; a residential facility) will sell SREC's to an aggregator, who will then sell it to utilities. There is a formal market for SRECs, I believe an auction mechanism is used. Regardless, the aggregator takes care of this stuff for the small producer; so that the small producer does nothing except cash checks sent to him by the aggregator (usually on a quarterly basis).
I chose Sol Systems as my aggregator. Of course, this is not a simple process. One has to fill out a form describing your installation, which is then sent to various state regulators for vetting and approval. And (of course), this is not a quick process

I recieve confirmation of a succesful application in mid November 2013.
And then (what else) I wait ... until March 2014. I start calling and asking "you claimed a 3 month wait, it's been over 4 months".
After a few more calls, the system is officially approved.
But what about the lost months of production? Turns out that they will retoractively start counting from December 1 2013.

How is the quantity of solar energy computed?
There has to be some way for the aggregator to measure how much solar electricity my facility produced. Note that this is the total wattage -- not the wattage sold back to the utility (that is, it includes both electricity sent back to the grid and electricity I directly consume)
Often, the aggregators use a model to estimate power produced. However, through my Tigo maximizers I can monitor & measure electricity production an a minute by minute -- or day-by-day or month-by-month -- basis. It would seem to make sense to use that number, rather than an estimate.
It took some insisting, but the aggregator agreed to do that (I give them my Tigo login information). I tried to be honest and point out that the Tigos report DC production -- and conversion from DC to AC is about 95% efficient (my inverters can report total production over any time period, so it is straightforward to estimate this efficiency factor). But they just don't want to hear about this further complication, so I let it be.

What is the payment rate? As a marketable commodity, the value of an SREC can fluctuate.
Presumably, as more people install solar panels, the SREC price will drop (more supply). On the other hand, Environmental Economics 103 dictates that regulators should set requirements so that the marginal cost of producing an ecosystem service is equal to its marginal benefit. Thus, as the supply of solar power increases (hence its marginal cost decreases), the regulation should be come more stringent. Perhaps one could accuse the solar power producer of advocating this regulatory adjsument for selfish reasons (keeping SREC prices high), but it is basic economics!

One can sign a multi-year contract specifying a fixed payment per SREC (they are sold in megawatt chunks). Or one can gamble and sell it on the spot market. Since I am generally risk averse, I figured a gamble is worthwhile, especially since the spot market price in November 2013 was several times greater than multi-year contract price.

Are SRECs a good idea? SRECs are a very good deal for the residential solar facility, since the per kilowatt value of an SREC is about what utlities charge per kilowatt. In other words, by selling SRECs one (more or less) doubles the value of every kilowatt your panels produce! But is this good public policy?

On one hand, SRECs are an answer to the question how to use market forces to reduce GHG emissions. A more direct approach would be to tax (or cap and trade) GHG emissions. But, subsidies for non-GHG emissions have a similar impact -- they increase the relative attractivness of non-emitting production technologies (such as solar panels). So maybe taxes would be better, but in lieu of taxes subsidies make sense.
On the other hand, when the environmentally conscious homeowner installs solar panels so as to put his money where his mouth is, his goal should be to cause an overall reduction in GHS emissions. But, if this homeowner then listens to his wallet and sells the SRECs to a utility -- his net effect is zero. Since whatever SRECs he sells the utilities means the utility doesn't have to directly purchase green electricity. In other words, all you are doing is saving the utility some money, and making some yourself. You are not directly reducing GHG emissions! One could argue that reducing the overall costs of GHG reductions will lead to an increase in GHG reductions (i.e.; via economically efficient strengthening of regulatory requirements discussed above), but that is an indirect impact.

As of mid-May 2014, still waiting for a first check.

The original plan was to have a battery backup system installed, which would automatically kick in and provide power to selected critical loads in the event of an outage....

The battery backup system would be integrated with my current generator, so that I could recharge the batteries using the generator. Furthermore, some of the PV panels would run during outages, providing power to the house and helping to recharge the batteries (as noted in the AC coupling discussion).

Once the PV installation was legal (mid October 2013), the next step was to install the battery backup. I had the necessary equipment (batteries and boxes and inverter). Shouldn't be too hard, shouldn't take too much time? Of course, it took 7 months.

As batteries are expensive, the system is a bit undersized, so one has to choose what loads are critical.
I had some ideas about the architecture of a system. Ideas that would allow the homeowner to selectively choose which loads were critical.
If too many loads are battery backed up (hence are powered via the Magnum inverter), you might overload the Magnum. Especially if some of these loads are mid-sized loads (such as the +1kw used by toaster ovens and microwaves)
Overloading the Magnum means it shuts off -- which means no power to the critical loads (even though the other, non-battery backedup loads, are uneffected). That's not a disaster, but it is a pain in the neck; as it it requires resetting the inverter, which is not something most people have any idea how to do.

I put together what I thought was an interesting architecture ...
An interesting solution would be to specify optional critial loads (such as the microwave) that are not battery backed up during normal times, but can be switched over to the battery backup system during outages.
The idea is to reduce the probability of overloading the battery backup system (the Magnum inveter) during normal times, but give you the convenience of powering optional critical loads during outages. Behind this idea is the notion that during outages one is highly cognizant of ones electrical useage, hence will avoid overloading the system. Whereas during normal times one doesn't want to think about that stuff. The ability to choose which of loads (especially ones that tend to be largeish, such as microwaves) are subject to battery backup allows this happy compromise.
- overview and rationale
- line drawings
Alas the electrician (who needs to sign off on the permit to the county) didn't like it. Too complicated. They proposed a simpler system, that wasn't as robust -- their are no optional critical loads. Kind of pissed me off, but (after a few tweaks by me) I accepted it. And truth be told, I suspect the electrician is correct -- the plan I had would of been real hard to convince the county to accept.
After some back and forth, a plan was drawn up (full version of plan).
Basically, the Magum would recieve power from the the set of breakers that were generator & supported; and the Magnum would send power to a new critical loads panel. Thus, when necessary the the critical loads and to recharge batteries.s

I took them to the county, and the first clerk quickly called in a higher level guy, who then went to an even higher level guy. I was able to explain the rationales (the fact that I noted the addition of a subpanel to deal with multi-branch circuits seem to make a positive impression), and they accepted the plan.
Time to install the equipment. Somewhat suprisingly, that took a week to complete (no time at all).
Alas, A few mistakes were made in the wiring, including bonding of neutral and ground in the critical loads panel. So it failed inspection, and some fixes were needed.
This failure made me nervous, so in addition to correcting this flaw (along with some not quite clear wiring of some multi-wire branch circuits), I reconsidered the DC grounding.
- Basically, the Magnum manual recommends a method which requires a connection to the grounding rod equal to the largest connection in the DC system. Which is the connection between the battery pack and the Magnum. Which is 3/0 -- wire used for welders. Which is not cheap. To avoid this cost, we had used the existing AC grounding connection, which might be pushing code.
- The reconsideration, and some pricing (3/0 is $2.50/foot) that wasn't utterly crazy, I purchased and installed 50 feet of the stuff. Had to reopen the hole in the basement wall that leads to the grounding rod, and entertain myself with good 2 hours of pulling heavy cable and drilling holes through floor joissts... not so bad, really.
These fixes did the job -- reinspection went smoothly, and as of May 2014, I now have a battery backed up system legally installed. Did some testing, and it works -- covers 1.5kw of critical loads for four hours (I ran a space heater to get that much constant load, as typical load on the critical circuits is about 600w). And the generator was able to recharge the batteries in 3 hours.
Hooray!