Green Insulation

Normally, I don't spend much time factoring in the amount of carbon generated by manufacturing something that I'm going to use for carbon mitigation when figuring out how much carbon it will reduce. My reason is that we have to use the carbon-based technology we have to build the carbon-free society we want. Most calculations I've seen for, say, the amount of carbon generated by manufacturing solar panels v.s. the carbon they save over their 30 year lifetime indicate that the scales are heavily weighted towards massive savings in carbon emissions for solar cells. Even renewable energy sources with marginal carbon saving, such as corn-based ethanol, have become more efficient over the past 5 years. If you don't take indirect carbon into account - the amount of carbon generated by land that is put into production to make up for the land devoted to corn for ethanol - there is a savings, though not much. Indirect carbon accounting is controversial, however. If the same accounting method were used for other technologies, for example, the carbon generated by nickel mining to make up for the nickel that goes into a Prius battery, almost any technology would come out losing in the end. The people who came up with indirect carbon accounting were just looking for an argument against corn ethanol because they don't like the idea of trading off food for fuel with the population still headed upwards, and I agree totally. There are plenty of other ways to achieve a carbon free transportation network which don't involve that tradeoff, and we would be pursuing them more forcefully if it weren't for politics.

But a couple of nights ago, I came up hard against this issue of global warming gas emissions from technologies that are good at reducing energy use and thereby carbon emissions. I'm not talking about carbon here, but rather HFCs. HFCs are gases containing fluorine, carbon, and hydrogen that are used most commonly in refrigerators and air conditioners. They replaced CFCs after the Montreal Convention was passed, because CFCs were causing reductions in the ozone layer, an even more serious - but now thankfully receeding - environmental  problem than global warming. HFCs have no impact on ozone but they are a global warming gas. Their impact on global warming can be many thousands of times as powerful as carbon dioxide. On the positive side, they do not remain in the atmosphere very long, around 7-10 years, unlike carbon dioxide which has a atmospheric lifetime of around 450 years.

I was interested in what impact our spray foam insulation might have so I sent email to Paul asking what product Ponzini used for their closed cell spray foam. As usual, he didn't reply so I checked Ponzini's web site and found a link to the product, Johns Manvill Corbond III SPF. I had done some research before on spray foam, and there are some green products out there, advertised as made from recycled plastic and soy oil, like Demlec. I had used Demlec in a previous open cell foam job to seal the wall between the attic/garage and the house. JM's Corbond  doesn't appear to use soy oil, it's petroleum based, but, as mentioned above, using food crops for nonfood applications has some powerful arguments against it. JM's Corbond claims around 16% recycled content, that is about what Demlec claims too, even though  JM doesn't make a big deal about being a green manufacturer. Quite to the contrary, their page on sustainability is filled with a lot of high minded rhetoric, but when you look at the pages on their products the information on ecological effects, such as about recycled content, is buried deep in the data sheets (for example, this data sheet on the B component of Corbond). Demlec, on the other hand, calls out this information in a completely separate page. Both Demlec and JM Corbond use HFC 245fa, a powerful global warming gas, as the blowing agent to blow small bubbles into the foam. So if you drill down on it, the only real difference between a "green" product such as Demlec and JM Corbond is the fact that the green product uses soy oil instead of a petroleum derivative.

I should say something about spray foam insulation here (unfortunately, there's no good Wikipedia page on this). Spray foam insulation comes in two basic types: open cell and closed cell. Open cell foam has a R-value about the same as fiberglass batt, R-3/R-4 per inch, while closed cell foam  has a much higher R-value, R-6 per inch. Closed cell foam is the "gold standard" in insulation. Only aerogel does better, at R-10 per inch, it is the platinum standard. "Platinum" and "gold" here refer not just to their insulation potential but also to their cost. Spray foam, even open cell spray foam, is about 2-3x as expensive as fiberglass batt, while aerogel is completely unaffordable except for very limited applications such as my planned thermal bridging treatment. Both closed cell and open cell foam reduce air penetration, unlike fiberglass batt, but open cell foam is permeable to moisture while closed cell foam tightly seals against it, thereby acting as a vapor barrier.

The kind of foam insulation that we are having installed is polyurethane, a commonly available product  used not only in home insulation but also for insulating refrigerators, hot tubs, coolers, etc. Prior to installation, polyurethane spray foam comes as two separate liquids, an A component and a B component. The A component contains a mixture of isocyanates while the B component is primarily a polyol with some other chemicals. When the two components combine, the two chemicals react to form polyurethane  plastic. A blowing agent pushes the two chemicals together and expands to form tiny bubbles in the plastic. The tiny bubbles give the plastic its insulation value. Depending on the blowing agent, the result is either soft open cell foam or rigid closed cell foam. If the blowing agent is water, the result is open cell foam because the water reacts with the chemicals to form carbon dioxide, which expands rapidly resulting in a less dense foam with less insulation value. If, however,  closed  cell foam is desired, the blowing agent is HFC245fa. HFC245fa expands more slowly than carbon dioxide so the bubbles are smaller. But both types of spray foam insulation generate green house gases, however closed cell foam generates a gas, HFC245fa, with 1000 times the greenhouse gas potential of carbon dioxide.

The fact that the insulation would result in greenhouse gas  emissions was, of course, somewhat depressing since the whole point of this work is to eliminate greenhouse gas emissions. However, as pointed out in this article,  the net impact depends on how much global warming gas emission is eliminated by the treatment over the lifetime of the product. So I went about calculating what the rough impact was of the global warming gas emissions v.s. the amount of gas eliminated. I took the crude architectural model of the house I created a few years ago and measured the depth of the studs in all areas so I could calculate the volume of the foam that would be blown into the stud bays. For the floor, I estimated the area as 2500 square feet. Then I calculated out the volume of foam, correcting for the volume of the walls, floor and ceiling that are structural members, as described in the thermal bridging post. I came up with a volume of 1435 cubic feet, or around 1500 cubic feet.  This is probably a bit high but should be a good enough estimate.

Closed cell foam has a density of 2 lb/cubic  foot. This gave me the weight of the foam, around 3000 lbs. Of that, half is the B resin which contains HFC 245fa, or around 1500 lbs. HFC 245fa makes up between 7-12% of the B resin fraction. Taking the higher number, the amount of HFC 245fa released should be around 180 lbs. Since HFC 245fa has a global warming potential around 1000x carbon dioxide, this would amount to around 180,000 lbs, or 90 English tons, of carbon dioxide. Then comes the question: how much carbon dioxide will be eliminated? That is harder to calculate, but I've estimated that R-6 per inch foam insulation in our walls should reduce gas usage by around 30%. Using that as the figure, the insulation should eliminate 106 therms/year of gas. At 11.7 pounds of carbon per therm, that's 2907 lbs of carbon per year, or around 3000 lbs rounding up. So it will take 60 years for the house to break even from the HFC245fa released during the insulation.This was truly distressing. Here I am trying to do the right thing and reduce the carbon footprint by making my house energy efficient and the technology I've used will take 60 years before there is breakeven in the carbon emitted by the insulation!

So what were the alternatives? We had originally planned to install a geothermal heat pump with  blown cellulose insulation in the walls and spray foam only in the ceilings and the floor. Geothermal heat pumps use  HFCs too, and  can leak them slowly out in the atmosphere, but not as much as spray foam. But the geothermal heat pump turned  out to be too expensive and too complex. We could have continued with blown cellulose, but it does not achieve an  air barrier and  is therefore much less effective than spray foam. It is possible to make a house air tight with blown cellulose, but it requires going over the entire house and sealing up all the cracks in the siding. In addition, most newer houses have complicated vapor barriers on the exterior between the siding and plywood sheathing. Our house has no plywood sheathing, there is just the exterior plywood siding and the interior drywall, with a  thin layer of tar paper under the siding. This is a rather poor vapor barrier, even in California's Mediterranean climate. So the spray foam  should give  us a much better vapor barrier. Blown cellulose also has a tendency to sag over time, and completely loses its insulation value if it becomes wet. Finally, spray foam increases the structure integrity of the building envelope, an important factor when considering earthquake. 

So having foam insulation has some really desirable properties which we don't want to give up. In a situation like this, my feeling is the best solution is to pay for carbon offsets. It's like flying, in some cases there are alternatives to flying (taking a train or driving for example) but they are often impractical. We offset our  our familial carbon emissions for housing and transportation every year anyway. We use carbonfund.org, a nonprofit, which means our carbon offsets are tax deductible. As it turns out, the carbon emissions from our insulation job should be about what a family of 4 in the US emits in one year, which in and of itself is kind of sobering. The cost at $0.004/lb is around $810. So next year, after the insulation  job is finished,  I plan to send a check to carbonfund.org for the carbon offsets. Carbonfund.org lets you select the type of offset, so I'll select  forestry to establish a long lasting sink for carbon that will span the 60 years that I hope our house will still be around (though of course I'll be long gone).

I've asked Paul to get the amount of the B resin used by Ponzini so I can refine the estimate, and I will report on the exact amount when I have it.

Yet More Drywall Removal

Last week, I suddenly realized that we might not actually have all the drywall removed from the thermal envelope. The front of the front upstairs bedroom sticks out slightly from the plane of the house, like a dormer. You can see it in this photo below:

This is about a two foot section where the bedroom has an exterior wall rather than sharing it with the attic.

When I went upstairs to check, sure enough, the drywall was still on. So the drywall removers had to come yet again. Here's what it looked like on the west wall when they were finished:

The round white target shaped thing on the top is the HRV vent.

Sigh. Hopefully, this will be the very last time we will need to have the drywall removers here, because Ponzini has started doing the insulation underneath the house.

Thermablok and Thermal Bridging

While mold most often occurs where there are roof or siding leaks, it also occurs when warm, moist air contacts cold structural members, causing water to condense out. We found mold on the face of a stud in the south hallway wall. The south hallway wall receives no sun in winter so the outside wall doesn't warm up during the day and, as a result, likely gets much colder at night. When mold occurs on a stud and there is no evidence of any siding leakage (as in this particular case) it is probably a result of thermal bridging. Insulation won't help because it is installed between the studs, and it won't stop heat transfer through the face.

I discussed thermal bridging in this previous post. The conclusion was that it would be too expensive and difficult to treat the whole house for thermal bridging, besides which, the insulation and drywall contractors probably wouldn't know what to do. However, I do want to treat the south wall of the hallway, to eliminate the condensation problem. So I ordered 50 strips of Thermablok , which I described in my previous post, enough to face the studs on the south hallway wall.  Thermablok is an aerogel insulation that comes in 4' by 1.5" strips. The back of the strip has adhesive on it. Installing it couldn't be easier: you peel off the plastic strip and paste it onto the face of the stud (making sure of course there is no drywall dust or other dirt on the stud to clog up the adhesive). Aerogel is the most effective insulating material after vacuum, at around R-10/inch. These strips will add R-4 to the studs, effectively doubling the insulating value of a 2x4. But aerogel isn't cheap, the strips are about $4 apiece.

On Fri., a package arrived with the Thermablok. The guys at Acoustiblok (the company that sells Thermablok) like packaging. The entire thing was packed in heavy cardboard with duct tape on both ends. Inside that, the Thermablok strips were bundled together in plastic garbage bags that were also duct-taped:

Inside these, the Thermablok strips themselves were packages in shrink-wrap:

Here's a closer view of the label:

I took the strip and set it against a stud to see how it would look when installed:

Here is a closer view of the bottom:

The old drywall was not removed under the molding, so you can see how the Thermablok strip will line up with areas where there is still old drywall. Only about 0.1" of space separates the face  of the old 0.5" thick drywall from the 0.39"  thick (after the plastic strip is removed) Thermablok strip.

This is a problem. The thinnest drywall you can get is 0.25" and that is not recommended on framing with 16" centers The recommendation is for 0.5" thick drywall, probably because drywall has very little structural strength, and 0.25" therefore has much less strength that 0.5". So if 0.5" drywall is installed, the new drywall will stick up 0.4" beyond the existing drywall and various other stuff on the wall (electrical sockets, window sills, etc.). I looked around for alternative wall covering material that is thinner than drywall,  but there doesn't appear to be much of anything. If we were planning on doing thermal bridging treatment for the whole house, this problem could cause a lot of additional work. All the molding would need to be taken off, maybe the electrical boxes would need to be moved, etc. But since we are only planning on doing one wall, the amount of work should be fairly limited.

Hydronic Radiant Heat Systems and Foam Insulation

Our  radiant heat system is what hydronics contractors call a "staple up". We have a crawlspace and the house originally had forced air heating. Unlike houses with radiant installed from the start, where the heating pipes are threaded through a poured concrete floor or slab prior to pouring the concrete, the pipes in our case are fastened to the subfloor. The contractor who installed our radiant system crawled around under the house for 2 months threading pipe between the joist bays beneath the floor.

For insulation, he stapled polyethylene/foil bubble wrap radiant barrier insulation immediately under the pipes, with a small gap to allow  the radiant barrier to function properly. Below that, he installed 4" of fiberglass batt. The house previously had fiberglass batt insulation under the floor, but, as I've mentioned before, fiberglass batt is a poor choice for insulation, and even worse in an application such as this. It has a tendency to sag, and in this case, because there is nothing across the bottoms of the floor joists, the tendency is even worse. In a wall-based application, at least the drywall is holding it in place, but I think the insulation contractor used fishing line or something to hold it up.

So as part of our quest to tightly seal the thermal envelope in our building, we decided to have closed cell foam blown into the spaces between the floor joists, just like in the walls and ceiling. But the situation is more complicated there, we can't simply foam on top of the pipes, since there needs to be some clearance. The foil radiant barrier provides that clearance, but will it withstand the pressure of the foam expansion? At first, Forrest, our architect, thought not, so we discussed installing some metal plates in place of the foil radiant barrier. But Paul said that would result in a much larger expense and delay. I asked Paul whether Ponzini, our insulation contractor, could do a test item that would determine whether installing the foam on top of the radiant barrier would work. Here's a picture of the test item:

The test item consists of two 2x8's with a plywood top simulating the floor assembly, and some radiant barrier stapled partway down. Below, you can see the radiant barrier and the gap:














The foil radiant barrier seems to hold up well, the foam pressure isn't enough to collapse it.  So we are going forward with foaming on top of the radiant barrier.  I have, however, asked Paul to check whether there is enough clearance on all the pipes and to ensure that the barrier is firmly stapled in place and hasn't collapsed in the 4 years since we had the hydronic system installed.

One issue that does concern me is leaks. A large earthquake could separate the PEX tubing from the manifolds and zone pumps. Leaks within the PEX tubing are less likely, but still possible if a pipe fails due to long term deterioration. With closed cell foam insulation, the water would accumulate between the insulation and subfloor, ultimately leaking out upwards and possibly warping the subfloor. The only indication of a leak is that the hydronic system starts drawing a lot of water out of the domestic water supply and the system pressure gauge indicates a drop in pressure from the normal 7 psi. But the pressure gauge is in the hydronic closet and we mostly don't look at it (though of course if there is a large earthquake we probably will check it to make sure the system is OK).

I briefly searched the Web for some water sensors that we could put into the floor to detect a leak directly. They all require replacing AA batteries at periodic intervals, a nonstarter since the sensors would be sealed into the floor by the insulation. This is a general problem with current sensor technology. If you want to put a sensor in an inaccessible place, even if it could have a wireless data connection, you still need to supply it with power through a wire. I called Kevin Smith, our hydronic contractor (he now has his own company, IntuitiveClimateControl) to discuss options and he said there are a variety of sensors we could put on the water line to the hydronic system that would measure and report if the hydronic system began  to draw water at an accelerated rate. So when we finish the job, I'll probably look into putting in a sensor, but not now. There are too many things to finish up, and we just want to get this job done so we have our living space back.

Neither Paul nor the insulation contractor seemed to know much about our situation, which leads me to believe that they haven't run up against a hydronic stapleup that needs closed cell  foam. I wonder if anyone else has tried closed cell foam insulation together with a staple-up hydronic system?

Duke Comes Through (Sort of...)

A couple weeks ago, our good friend Duke came through and fixed the problems with the HRV venting... sort of. In the front, he cut a new vent, which is supposed to be for the HRV exhaust, below the attic vent. He hadn't switched the actual ducts yet, so the intake duct was connected to the new vent rather than the exhaust, but he knew about the problem and was planning on switching the ducts around when he came back to install the fitting on the condensate drain. Apparently, Fantech forgot to ship a condensate drain fitting along with the HRV unit. Duke ordered it, it was supposed to be here that Fri. Unfortunately, our good friend Duke is lacking in, shall we say, aesthetic sense. The concept of "curb appeal" is lost on him. Consequently, the front of our house now looks like this:

The new duct is about 3 inches lower than the old one, instead of nicely aligned. Rather than try to fix the problem, which would involve filling in the hole, we're simply going to paint the ducts the same color as the house and leave it at that. They should blend in enough with the house that they won't be all that visible. We also needed an electrical outlet to plug the front HRV in, but that's the electrician's job. He was supposed to install it that week.

Speaking of aesthetic issues, the HVRs require piping to drain off any condensate that might occur on the heat exchanger. The missing fitting is what was holding up the front HRV. Since the HRVs are installed in attic areas without any plumbing, there needs to be some way to get the water out of the HRV and drained away. Here's what our back condensate piping looks like:

The white pipe coming out through the wall is the condensate drain. You can also see the solar panel for one of our solar garden pumps near the top. This white pipe doesn't look so nice either, and will ultimately be painted. The pipe is unfortunately PVC. PVC is not a particularly environmentally friendly material. I asked  Paul why they couldn't use ABS pipe instead, but ABS doesn't come in the 1/4" diameter stock that is needed for this application.

On the back, we now have a veritable proliferation of hoods:

Only the one on the lower right and upper left will ultimately stay. The others will disappear when the siding is replaced in the triangle on the right side of the picture. Then they will recut the vent on the lower right, which will be an attic vent. The upper left vent is the HRV intake. The HRV exhaust has moved to the roof:

It should be the tall one, the small one is the bathroom (high speed) exhaust. Below you can see how Duke rerouted the bathroom exhaust duct out the roof:

It is the hard, elbow-shaped pipe running up  the middle of the picture. The exhaust is the soft duct on the upper left.


As they came from Fantech, the SH704s have no low voltage switch. You just plug them in and they start up. I'm not sure what Fantech had in mind as far as control. A Web search turned up the Aubetech 1702-3W which is a three wire, high voltage (i.e. 110V), 7 day timer switch with maximum 2 settings per day. I ordered and received 2  Aubetechs for the HRVs. They are rated for up to 1 HP motors, the Fantech motors are less so there should be no problem. This will allow the HRVs to turn off during the day when we are at work, and on again in the afternoon. The electrician needs to wire them up yet. We are going to put them into the closets in the bedrooms nearest the HRV units, so they are not visible from the room.

I also started the HRV above the master bedroom and, to my surprise, there was no problem with vibration at all. While the actual HRV unit itself vibrated due to the fan motor, the vibrations were not transmitted through the shock absorbers to the house frame. There was a very faint noise in the downstairs bedroom, about what we hear when the solar thermal pump goes on, which I assume was from vibrations transmitted through the air to the ceiling. With closed cell foam in the ceiling and additional soundproofing, I think the noise level should be OK.

This week, Paul, Christine, and I checked the HRVs and it seems like everything has been properly installed. The ducts are routed out the right vents, the front HRV has an outlet, and the condensate drain is in on the front HRV. The electrician still needs to install the Aubetechs yet, but he's started the wiring. All in all, it looks like our adventure with HRV installation is slowly drawing to a successful, if not entirely satisfactory (from an aesthetic standpoint) close. In a future post, I'll give a retrospective.

The Last of the Drywall Removal

We are hopefully finally, finally done with drywall removal. Due to a screwup by The Project Manager Who Shall Not Be Named, the original drywall removal effort failed to remove drywall from 4 different areas on the thermal envelope. Two of those were small enough that they could be removed without calling the drywall removal contractor back, but the other two were not. One was fairly obvious, it was the back wall of the southeast upstairs bedroom (Bedroom3), and we called the drywall removal contractor back to do that in September. But the last area was a 4' wide section on the southeast wall of the main hallway that Christine caught when she was in the house a couple weeks ago. We had it removed this week:

Our house has lots of small wall runs along the thermal envelope like this one. While they give the house a kind of quirky aesthetic (actually, one reason why the house appealed to us in the first place), they do increase the surface area for heat leakage and complicate insulation. If the house were just one big, four walled box sealing it would be simple but it would be less interesting from an  architectural standpoint. Proper air sealing becomes even more important for these small runs, and especially the joints between the short runs and the rest of the wall. You can be sure I will be looking for that when I do the inspection after the insulation contractor is finished.

PG&E and Electric Car Chargers

If you recall, we applied to PG&E for a 200 amp service upgrade at the beginning of November (we currently have 100 amps). The request has been grinding its way through PG&E's process. A few weeks ago, we informed them that we would be getting an electric car next year. This week, The Lovely Wife got a call from a Glenn, from PG&E. He asked to speak with me. Because he called TLW on her cellphone, I wasn't there. But he never called me back. On Thurs. at our weekly project meeting, Paul told me that he had spoken with Glenn (I guess he thought Paul was me) and what PG&E wanted. What they wanted was to know whether we wanted a second meter for the electric car.

I was puzzled. Why would we want a second meter? Paul said that PG&E wanted us to charge the car at night, to avoid overloading their transformer and the circuits that supply our neighborhood. The claim is that an electric car establishes a load equivalent to another house. The charger draws maximum 40 amps at 240V (but most likely less than that, since breakers are typically overprovisioned). That's 4.8 kw, a pretty hefty sum, and that over 6 hours, for a total of 28.8 kwh. If I'm the only person in the neighborhood charging, then there is probably no problem. But if electric cars catch on, PG&E is in trouble. They will quickly run out of overprovisioned capacity and need to start upgrading their transformers and other "last mile" circuits. As we know from the telcom world, utilities hate to upgrade "last mile" infrastructure because there is a lot of it and therefore upgrading is expensive.

I don't think PG&E has much to worry about, personally. I suspect my neighbors with houses built in the 70's like mine (some were built later) all have a 100 amp mains feed like ours does. In our neighborhood, the mains feeds are underground rather than coming from  a pole, which makes the neighborhood look nicer but makes upgrading a real challenge. Rather than simply restringing a wire, PG&E needs to bring in a backhoe, dig up the conduit, and replace the 100 amp line with one that will handle 200 amps. Backhoe work is expensive and PG&E's process is time consuming. I can't see any of my neighbors going through that kind of grief, unless gas gets up to $6.00 a gallon. Then maybe even the two neighbors across the street with two SUVs each might have second throughts.

Anyway, I pondered briefly if we did want a second meter for the electric car charger. PG&E offers a special tariff for electric car charging service, called E-9. The cost works out to around $0.06/kwh at night and $0.30/kwh during the day. So there is a real financial incentive not to charge during the day when more power is being used by businesses and residences. But we are already on a time of use tariff,  E-7W, for our solar PV. This has the same kind of incentive: $0.09/kwh at night most of the year and $0.30/kwh during the afternoon in summer (the winter rates are much less). The rate on summer afternoons is high because PG&E needs the power for all the air conditioners in the Central Valley. We can save big time by not using power then and banking the credits for the winter, when our solar panels get little sun and the folks in the Central Valley have their gas heaters cranked up instead of their air conditioners.

But the $0.03/kwh differential in price caused me to haul out my calculator. Would we save by taking a second meter an the E-9 rate? Our new solar PV system has 2000 kwh/year capacity designed in for the Nissan Leaf. This amounts to around 8000 mi/yr, the amount we usually drive our around town car, at 0.25 kwh/mi, the amount of energy a reasonably aerodynamically efficient electric car uses. Since we cannot connect up the solar PV system to more than one meter, the generated power would not offset the power we use for the electric car on the second meter. Instead, PG&E would reimburse us for it, which they will start doing next year. The rate we get is paltry, $0.08/kwh, even for the expensive power we generate in summer (this is a ripoff in my opinion, but that's the subject of another post).

We would end up paying $120/yr for power on the E-9 rate and getting back $160/yr for the solar power we sell to PG&E on the main meter, for a grand total of $40 to our credit, enough for a nice lunch for two at a fancy Palo Alto restaurant. Looks good, right? Think again! PG&E also charges you $75/yr for renting the meter. So we would end up paying $35 more with a second meter than with a single meter. And, in addition, we would be constrained to always having to charge at night, even if we occasionally need to charge in the afternoon on a weekend because we have taken the car somewhere (this happens sometimes with our current plug-in hybrid). So our answer to PG&E is a polite: "Thanks but no thanks."

According to PG&E's online schedule, the 200 amp upgrade begins Dec. 24. Think of that: PG&E as Santa Claus!

More on MPP Balancing v.s. Microinverters

 IEEE Spectrum has a short article this month discussing MPP balancers v.s. microinverters for optimizing power from solar PV systems. If you recall, in this post, I discussed the reasoning behind my decision to go with Tigo's MPP balancer solution v.s. a microinverter solution such as Enphase. Suprisingly (at least to me) the author of the article confirmed the reasoning. The author found  the MPP balancers, in particular Tigo, better than microinverters because the electronics are simpler and cheaper and there are no electrolytic capacitors to fail. The complicated part is not up on the roof, it is down next to the inverter doing calculations, so it is much easier to replace. REC Solar (the company we are having install our system) tried out both microinverters and MPP balancers at its corporate headquarters for a year before going with Tigo, something we in Silicon Valley call "eating your own dog food."

The REC sales guy also mentioned something about an infrared picture of a system with a microinverter where the microinverter was in the center of a huge red area of high temperature. I searched a bit but couldn't find it. I found a forum where there was some discussion of microinverters v.s. MPP balancers. The claim from the microinverter faction is that microinverters are mounted on the panel racking and therefore don't touch the panels, but even so, they increase the temperature by around 6F. That shouldn't be enough to cause major problems if properly installed, but a slight bit off and the panel could get cooked. Temperature increase was another one of my concerns. It can cause premature panel aging and will in any case reduce the amount of power produced by the panel. MPP balancers use DC-DC conversion to maintain the voltage at an optimal level for maximum power output. This results in less heat loss near the panel, though, of course, the centralized inverter still is not 100% efficient so the inverter can heat up. But it is easier to put the inverter in a shaded area, or an area with some wind to dissipate the heat.

Really, the only advantage I can see of microinverters is that they can be installed by someone with essentially no training, except in standard AC wiring. MPP balancers require training to install. This of course makes the labor cheaper, but - the professionalism of American construction technicians being in general what it is - you get what you pay for. If the guy installing it decides, hey, why not just clip the microinverter to the panel nobody will notice, it will end up prematurely aging the panel. Then they get to come back and replace it, for a fee. We've had plenty of experience on this job with that kind of reasoning.

I also discovered I need a CAT5 cable from the monitoring box to where my DSL router is located (technically it is a DSL modem cabled to a Linksys router, I have an older model modem). Luckily, the walls are completely open,  since the electrician will need to run the cable through the walls from the front of the garage to around the middle of the house on the second floor, where our home office is and where we keep the DSL modem. I asked Tigo about a WiFi wireless bridge. These devices are not access points, but act like a wire without the copper. Tigo discouraged me from using one. They claimed that the wireless bridges often go down and then they need to call up the homeowner and ask them to reboot it. The Tigo box sends measurements to the Internet every 10 minutes, and can buffer up to 2 days worth of data. Tigo monitors the condition of the panels and I guess lets the homeowner know if there is some problem.

I think the assumption behind the wireless bridge manufacturers is that they will be used in an office setting and therefore someone will notice if they go down because YouTube isn't coming up. Then somebody will reboot the bridge. For machine to machine communications, though, the reliability must be much higher, otherwise, you end up with an expensive "truck roll equivalent" in which some person must find the box (could be difficult if it is in an out of the way place) and reboot it. Perhaps wireless links are better for control systems, where a person is dialing temperature up or down or setting a timer on a fan. In that case, if the wireless link is faulty, it should show up on the remote and the person operating it can take some action, like replace the batteries in the remote.

So we'll go with the CAT5 cable, and, in addition, get another from the solar hot water tank so that we can at some point put Internet data recording for that too. With data on both, I can get a much better picture of how much of the house's energy is coming from our renewable energy systems.

More on Optimal Solar Thermal Backup

If you recall, in this post, I discussed why we chose an electric on-demand hot water heater for backup to our solar thermal hot water system. The conclusion was that electric on-demand was best because:

• On-demand allows the solar hot water to contribute the maximum amount of energy. If you have a tank-based backup, either integrated with the solar hot water tank or in a separate tank, the backup energy source will come on whenever the tank goes below 120F and will heat the whole tank, whereas the on-demand heater only comes on to heat the water you need.
• Electric on-demand is better than gas on-demand because the electricity can be directly offset by installing more solar PV. But even if you don't consider the offset, electricity is better because the electric grid is incorporating fossil carbon-free energy sources faster than the gas grid.
• Gas on-demand requires complex venting whereas electric on-demand is just another appliance you can fasten to the wall in a closet (which is what we are doing).

An article in the most recent edition of Solar Today goes into much more detail about the optimal backup for a solar thermal hot water system. The criteria in the article are somewhat different, emphasizing seamless transitioning between solar thermal and backup. We are not so picky, we turn off the backup in summer and on again in late fall,  so if we get any clouds in the summer we may end up with a day of warm water. But the article confirms that on-demand is better than tank backup, for the same reason as I concluded, and that electric on-demand is better than gas. The author goes into much more detail about why electric is better than gas, but it boils down to the lack of real modulating capability in gas on-demand and the need for a larger draw of water volume (0.5 gps v.s. 0.25 gps for electric) before the gas on-demand unit kicks in.

On the downside, electric on-demand units draw an enormous amount of current. Our unit has 3 220V/20 amp heating units. When all units are active, it will draw more current than the hot tub does. The utilities don't like that kind of draw because it causes transformers to heat up, and can cause lights to flicker in the house, especially if there is more than one on-demand heater active in a neighborhood at a time. You need to have a 200 amp residential service in order to install one, we have a 200 amp service upgrade on order from PG&E.

But the utilities are going to have to upgrade their transformers and circuits for electric cars anyway. Gas on-demand units also require a large (1" I think) gas main, which many residences don't have, so even with gas-on demand you may need a residential service upgrade from the utility. And, for solar backup, the on-demand heater must only heat the water from 80F to 120F, not from 60F which is the usual water main temperature here in winter (our solar hot water tank is now around 80F due to cloudy weather), so it is less likely that, in our case, all three heating units will fire at once. All things considered, electric on-demand seems to be a better option for solar thermal backup.