musingsheader image
9 Helpful?

Solar Thermal is Dead

It’s now cheaper to use a photovoltaic system to heat domestic hot water

Posted on Mar 23 2012 by Martin Holladay

In the northern half of the U.S. — and even much of the South — installing a residential solar hot water system doesn’t make any sense. It’s time to rethink traditional advice about installing a solar hot water system, because it’s now cheaper to heat water with a photovoltaic (PV) array than solar thermal collectors.

In short, unless you’re building a laundromat or college dorm, solar thermal is dead.

The idea has been percolating for six years

In the early days of PV, when PV equipment was much more expensive than it is now, homeowners with PV systems (especially off-grid homeowners) were instructed not to use electricity for heating. After all, since electricity is precious and expensive, and since PV power usually costs even more than grid power, it made sense to save electricity for uses like refrigeration, lighting, and home entertainment.

For decades, we all assumed that the greenest way to heat domestic hot water was to use a solar thermal system. But then two things happened: PV equipment got cheaper, and heat-pump water heaters became widely available.

The logic of using a PV system to heat water was first explained to me in early 2006 by Charlie Stephens, a policy analyst for the Oregon Department of Energy. I reported the details of that conversation in an article, “Heating Water With PV,” published in the May 2006 issue of Energy Design Update.

“If you want to do solar water heating and solar space heating, solar thermal remains too expensive,” Stephens told me. “It’s not as cost-effective as using an air-source heat pumpHeat pump that relies on outside air as the heat source and heat sink; not as effective in cold climates as ground-source heat pumps. coupled to a PV array. In our climate, a properly sized solar thermal system can provide 100 percent of your hot water in the summertime, but it won’t do diddly in the wintertime. So you paid $4,000 for a system that provides 40 or 50 percent of your hot water needs. If instead, using the same money, you just add an extra kilowatt of PV to the roof, you could heat all of your hot water year round with an air-source heat pump.”

You can quibble with the details used in Stephens’ argument — it may take more than a kilowatt of PV to meet your hot water needs, for example, and his 2006 price estimate for installing a solar hot water system is now much too low — but his conclusion is even more valid now than when it was first made.

Some solar-heated water goes to waste

Solar thermal proponents know how to calculate the number of gallons of hot water produced by a typical 4' by 8' solar collector in a variety of climates. After calculating the thermal energy that this represents, they usually concluded (before PV prices dropped, anyway) that solar thermal collectors were a better bargain than a PV array.

But the number of gallons of hot water produced by a solar collector is always less than the number of gallons actually used by the homeowners. After all, if great quantities of hot water are produced on a day when it isn’t needed, you can’t really count the energy production in your annual tally.

Solar thermal energy is inconsistent, and during the long sunny days of summer, most solar thermal systems make more hot water than the typical family can use.

Although Charlie Stephens (pessimistically) estimated that a residential solar thermal system in the Pacific Northwest would only supply about 40% and 50% of a family’s annual hot water needs, the so-called “solar fraction” will be higher in other climates. In a 2006 study, researchers from Steven Winter Associates monitored two residential solar thermal systems for a year, one in Wisconsin and one in Massachusetts. Each house had two solar collectors. The solar fractions of these two systems were 63% and 61%, respectively.

Comparing solar thermal and PV systems

Compared to a PV system, a solar thermal system has several disadvantages:

  • Unlike a PV system, most solar thermal systems have moving parts (pumps and solenoid valves).
  • In freezing climates, solar thermal systems are sometimes subject to freeze damage.
  • Solar thermal systems require regular maintenance, including antifreeze replacement.
  • Unlike owners of a grid-connected PV system, who can be credited for their excess electricity production during the summer, owners of a solar thermal system can't sell the excess summer production of their hot water systems.
  • While a pole-mounted PV array can include a tracking mechanism to follow the sun's path across the sky, it's virtually impossible to install solar thermal collectors on a tracker.
  • On average, PV systems probably last longer than solar thermal systems.

There are far more stories of troublesome solar thermal systems than there are stories of troublesome PV systems. Solar thermal systems sometimes develop air bubbles that interfere with the circulation of fluid, suffer from leaking pipes, or experience problems from summertime overheating. PV systems, which suffer none of these headaches, look attractive in comparison.

Let’s do the math

Since 2006, when Stephens first proposed that it was cheaper to heat water with a PV array than a solar thermal system, two factors have emerged that greatly strengthen his case: more reliable heat-pump water heaters have become widely available, and PV modules have gotten dramatically cheaper. (During the same time period, sales of solar thermal systems have also been hurt by a third factor: dropping natural gas prices. But that's a topic for another article.)

Although it's always difficult to predict future price trends, there are reasons to believe that the price of PV modules will continue to drop, while the price of the copper tubing used to make solar collectors will continue to rise.

In northern states, a typical residential solar thermal system includes two 4' by 8' collectors and a 120-gallon solar storage tank; the installed cost for such a system is about $8,000 to $10,000. Instead of spending $10,000 on a solar thermal system, what would happen if you invested $3,000 in a heat-pump water heater and $7,000 in a 1.7-kW PV array?

The 1.7-kW PV system would produce 2,114 kWh per year in Boston or 2,093 kWh per year in Madison, Wisconsin. Let’s be conservative and use 2,000 kWh for our example. Assuming that the average COP of the heat-pump water heater is 2.0 — a fairly conservative assumption — it takes 0.0855 kWh to raise the temperature of a gallon of 50°F water to 120°F. So 2,000 kWh can produce 23,392 gallons of hot water a year, or 64 gallons a day — exactly equal to the amount of hot water that the U.S. Department of Energy assumes is used by the average American family. So, once you’ve paid for the system, you get “free” hot water.

That’s a much better deal than a solar thermal system that produces only 63% as much hot water — even if you do have to buy a new heat-pump water heater in 12 or 14 years. (Trust me — if you have a solar hot water system, you'll have to invest in maintenance and replace a few parts over time, too.)

Of course, if your family uses less than 64 gallons of hot water a day, or your heat-pump water heater has a higher average COP than 2.0, or you live in a state with more sunny days per year than Massachusetts or Wisconsin, or the average temperature of your incoming cold water is higher than 50°F, then your new PV system will be producing extra electricity that you can use for other purposes.

In fact, many families use significantly less than 64 gallons of hot water a day. A Canadian researcher, Martin Thomas, monitored hot water use in 30 Canadian homes in 2008; the average hot water use by the monitored families was only 44 gallons a day. If your family uses 44 gallons of hot water a day, you'll only need a 1.2-kW photovoltaic array (costing about $5,000) — or, in a sunny climate, an even smaller PV array — rather than the 1.7-kW array proposed for northern families using 64 gallons of hot water a day.

What if you use an electric-resistance water heater?

Let's do the math for those who prefer to use an electric-resistance water heater. If you invest $10,000 in a PV system, you'll get a 2.2-kW system (assuming a PV equipment cost of $4.54 per watt). The PV system will produce 2,736 kWh a year in Boston. Using electric resistance heat, it takes 0.171 kWh to raise a gallon of 50 degree water to 120 degrees, so you'll end up with 16,000 gallons of hot water per year, or about 44 gallons a day — about exactly the average water use by U.S. and Canadian families, according to two recent studies.

If the family with the hypothetical $10,000 solar thermal system uses 44 gallons a day, and the solar fraction is 63%, their solar thermal system heats about 28 gallons a day on average. The PV option produces 37% more hot water, even with an electric resistance heater — and with far less hassle.

To make 28 gallons a day — an amount equal to the average output of the solar thermal system — with an electric-resistance heater, all you would need in Boston is a 1.4-kW PV system costing about $6,300.

But — but — but —

Here’s the part of the blog where I admit that my chosen title — “Solar Thermal Is Dead” — was deliberately provocative and somewhat inaccurate.

So I’ll list a few exceptions to my new rule:

  • Solar thermal systems still make sense for off-grid homes.
  • If you can get a two-collector solar thermal system installed for $5,000 or less — an attainable price in areas of the country where people don't have to worry about freeze protection — it may make sense to install one.
  • In a sunny, warm climate, where a solar hot water system will have a higher solar fraction than 63%, an investment in a solar thermal system makes more sense than it does in Wisconsin or Massachusetts. (On the other hand, a PV system produces more electricity in a sunny climate, too.)
  • If you are skeptical about the longevity of heat-pump water heaters, you may prefer to wait a few years before buying one, and to stick with a solar thermal system in the meantime.
  • Before taking the advice given in this article, compare the costs and energy production figures of a solar thermal system and a PV system in your area, using location-specific energy production figures and local equipment costs and installation costs.

There are many factors to consider when choosing equipment to heat domestic hot water. One point is clear, however: if you plan to install a heat-pump water heater, you definitely don't want to also install a solar thermal system. The correct solar complement to a heat-pump water heater is a PV array.

Green builders have an emotional connection to solar hot water systems, because they represent a fairly simple technology that's been around for over 100 years. But it's time to admit that a PV array is cheaper and less troublesome than fluid-filled solar collectors on your roof.

Author's postscript: For an updated (December 2014) analysis that compares the cost of heating domestic hot water with a solar thermal system to the cost of heating domestic hot water with a PV system, see Solar Thermal Is Really, Really Dead.

Last week’s blog: “A Superinsulated House in Rural Minnesota.”

Tags: , , , , , , , ,

Image Credits:

  2. Metro Solar Atlanta

Dec 13, 2012 5:53 PM ET

Solar Hot Water Springs Back To Life
by Jeffrey Morrow

Solar thermal's value is a function of both cost and benefit. Heating domestic hot water with natural gas rather than solar may be cost effective. But if you're heating more than domestic hot water, the math changes quickly. Attached is an image of a Simple Drainback hot water tank plumbed in thermosiphon with a gas backup tank, with combi-system capability. This low cost system can heat floors, do snow melt, heat hot tubs, even be hooked to a wood stove. Solar thermal is alive, and healthy in circumstances where electricity or propane is displaced, and where a single solar hot water system serves several functions.

Feb 2, 2013 7:46 PM ET

Your Mileage May Vary
by Sherwood Botsford

Here in Canada EnMaxx will lease you a 3 kW system for $60/month for 15 years. At the end of that time you can buy out at $350. About $11,000.

In our climate one kW installalation will generate about 1100 kWh/year.

Our local power rate is about 15c/kWh

So the system will save about $500/year. Thus, the simple payback time is 22 years -- 7 years longer than the lease. Looking at it another way, it costs you 720/year to save 500/year.

This assumes that the system stays in perfect working order for the entire time. In practice PV systems degrade at about 10% per decade.

Feb 3, 2013 1:14 PM ET

Edited Feb 3, 2013 1:16 PM ET.

Response to Sherwood Botsford
by Martin Holladay

As far as I can tell, your comments refer to the payback period for a PV system, not a solar thermal system.

While PV payback is not the topic of this blog, your comments are interesting and welcome. PV payback periods vary widely from one region to another. The important variables are the number of hours of sunshine each year, the local cost of grid-supplied electricity, and the local availability of subsidies, incentives, and tax breaks.

For more information on PV payback, see PV Systems Have Gotten Dirt Cheap.

Feb 16, 2013 1:03 AM ET

Sherwood and MArtin : i was
by Jin Kazama

Sherwood and MArtin :

i was just about to post a similar comment !!
( after reading through "most" of the comments in 1 shot ... )

I don't believe that PV systems are appropriate for us here in Quebec either for now .
Their installed price would have to cut down almost in half , which will probably not happen anytime soon.

We are still at ~ 0.8$/KW/h of pretty "clean" electricity,
and production of PV are usually in the 1 for 1 region from what i heard around
( as in 1KW array = 1000 KW/h produced yearly )

Since ~60% of our electricity bills go toward heating
i'm considering researching advanced solar thermal gathering, but not for hot water ..
Storage during daytime and release during nighttime is what i believe to be the key here.

Though, we don't see any/much thermal solar systems here, it never really went ON so it cannot really die i guess :p

Mar 2, 2013 3:28 AM ET

PV Panels to Preheat Water
by Peter Crisp

I'm late to the table joining this conversation. I have an idea, which might be dumb or useful and I can't find any examples where it's been done. Could a small PV system and storage tank be installed to pre-heat water for either conventional hot water tanks or tankless units using heat trace and a gravity loop for warming? If it works, it should be small, easy to retrofit (no house circuit re-wiring and no double-wall heat exchangers to start with) with a fast payback if it works.. I'm an engineer but I won't profess to being too sharp on thermodynamics. I live up in Canada where the incoming water temperature can get pretty cold. I'd really like some feedback on if this could work, or why it wouldn't. If it could work, we have a couple of rental houses to test the idea out on. Thanks!

Mar 3, 2013 7:56 AM ET

Response to Peter Crisp
by Martin Holladay

If your house is grid-connected, and you want to install a small PV system, go ahead. All of the electricity produced by the PV system will contribute to lowering your electric bill.

If you want to install an electric-resistance water heater to pre-heat your domestic hot water -- or, for that matter, to bring your domestic hot water up to any temperature you want -- go ahead. You are free to do so.

The two systems -- the PV system and the electric-resistance water heater -- are independent. The decision to install either system does not depend on the decision to install the other system.

Apr 3, 2013 11:58 PM ET

vacuum tubes, other uses for hot water
by David Jones

"Green builders have an emotional connection to solar hot water systems, because they represent a fairly simple technology that's been around for over 100 years. But it's time to admit that a PV array is cheaper and less troublesome than fluid-filled solar collectors on your roof."

Well said. I admit I am reluctant to let go of solar thermal as the best value per unit of energy collected. However any zero energy home needs some electricity production and in most cases that is PV. Regardless of the exact economics having 2 systems (PV and solar thermal) is more complicated than having a single larger PV system.

I have some concerns about the how the deck was stacked here, but that probably doesn’t change the conclusion.

"Some solar-heated water goes to waste
"After all, if great quantities of hot water are produced on a day when it isn’t needed, you can’t really count the energy production in your annual tally. Solar thermal energy is inconsistent, and during the long sunny days of summer, most solar thermal systems make more hot water than the typical family can use."

A pool or hot tub can use all that extra thermal energy. The extra energy happens to come at the perfect time of year. Granted most green homes won’t have a pool/hot tub.

Does anyone ever have too much hot water? Wash your car or your dog with warm water. Heat the soil under your vegetable garden to extend the growing season.

Low temperature water gathered in colder months that is not usable for hot water or space heating could go to snow melting.

"In a 2006 study, researchers from Steven Winter Associates monitored two residential solar thermal systems for a year, one in Wisconsin and one in Massachusetts. Each house had two solar collectors. The solar fractions of these two systems were 63% and 61%, respectively."

Were the solar installations SWA monitored flat plate or vacuum tube?

I am not trying to start the flat plate vs. vacuum tube debate. I know there are many who prefer the flat plate collector for good reasons. However; a vacuum tube system will produce higher temperature water, more days per year, and more hours per day. That can have a significant effect on the solar fraction, (as can adding a 2nd storage tank).

"Compared to a PV system, a solar thermal system has several disadvantages:
"Unlike a PV system, most solar thermal systems have moving parts (pumps and solenoid valves). "

True but circulators and solenoid valves are cheap compared to replacing an inverter.

"While a pole-mounted PV array can include a tracking mechanism to follow the sun's path across the sky, it's virtually impossible to install solar thermal collectors on a tracker."

A tracker is a lot of complicated and expensive moving parts. If Honda can’t build a minivan with a dependable automatically closing side door, I can’t imagine that a site assembled low production solar tracking assembly could be considered a maintenance free item.

A vacuum tube collector is always perpendicular to the sun with no moving parts!

"On average, PV systems probably last longer than solar thermal systems."

Inverters are reputed to have short life cycles. These are big ticket items compared with circulators and solenoid valves.

The PV panel seems to last a very long time, (but possibly suffer a minor loss of efficiency over time). Solar thermal panels should last a long time as well. No moving parts in the panels. In both PV and solar thermal the maintenance is likely to be in the mechanical room.

"(During the same time period, sales of solar thermal systems have also been hurt by a third factor: dropping natural gas prices. But that's a topic for another article.)"

I my experience, solar thermal sales are hurt far more by subsidies that are radically tilted toward solar PV rather than solar thermal. (I suspect the PV industry has more money to lobby for subsidies because PVs are bigger ticket items.) In many cases it less expensive for a homeowner to acquire a $30,000 PV system than a $6000 solar thermal system.

"Of course, if your family uses less than 64 gallons of hot water a day, or your heat-pump water heater has a higher average COP than 2.0, or you live in a state with more sunny days per year than Massachusetts or Wisconsin, or the average temperature of your incoming cold water is higher than 50°F, then your new PV system will be producing extra electricity that you can use for other purposes."

Of course if you live in a state with more sunny days than Massachusetts, your solar thermal system will produce more hot water too.

"If the family with the hypothetical $10,000 solar thermal system uses 44 gallons a day, and the solar fraction is 63%, their solar thermal system heats about 28 gallons a day on average. The PV option produces 37% more hot water, even with an electric resistance heater — and with far less hassle."

I suspect the solar fraction from the SWA study of 63% is for a flat plate collector system and that it would be significantly better for a vacuum tube collector system. However it will never be 100% without massive storage. The PV system does appear to come out on top.

The economics of this debate can be affected by:

1. New Construction vs. Existing
2. Available roof space. (I believe solar thermal can collect a great deal more energy in a given amount of space).
3. Size of system. A large portion of the solar thermal pricing in your comparison is installation. That figure is non-linear. If a house can utilize a larger solar thermal system for space heating/pool heating then the cost for the domestic hot water production comes down.
4. Heating system in the house. A hydronic radiant system is a perfect match to low temperature solar thermal. A radiant floor system can potentially utilize very low water temps which makes it possible to get more energy out of a given thermal collector by directing the energy toward space heating or domestic hot water production.

It’s generally accepted that people should focus on improving (or building) the best envelope possible before they add PV or solar thermal. If we assume a very well insulated building (near zero energy), then it seems reasonable to assume that essentially all of the energy being “pumped” by the indoor air source heat pump water heater is actually coming from the building heating system. In that case I don’t think it makes sense to use the heat pump water heater at all. There is probably a significant efficiency penalty in heating the house with a mini-split and then moving that energy again with an indoor heat pump water heater.

I do like the idea of the air source heat pump water heater during the cooling season however. I think the unit could be located in a closet, and the air from the closet exchanged with a small fan to provide some cooling and dehumidification to the house.

Apr 16, 2013 8:41 AM ET

by JP Jon Pierce

"But it's time to admit that a PV array is cheaper and less troublesome than fluid-filled solar collectors on your roof."

IF the water is on the roof: AGREED.
But why would a geothermal nut
(both direct fluid tubes without a heat pump / as well as with...)
look in 1980 at the Solstar(tm) system to heat Hot Water?
Perhaps it is STILL under 5000 bucks today for the nominal-normal-to-Sun 2 collector system to do 6800+ btuh (2kwh) water-heating WITH SPACE HEATING.

Air-Solar/ filtered closed-loop air circulation; simple 3way combo-blower-to-air-diverter ducted; all flexduct (that has withstood 30+years of non-brittle-izing); fin-tube for water-heated-from-air exchanger; 80w magnetic coupled impeller-pump; 1/2" short tubing to HW tank; simple snap disc and paralleled redundant safeties; one Temp-differential controller; one 4 pole double-throw relay (small ice cube style); small induction relay to fan motor; a "relay in box addition if needed; two great roofing-flashing-mounting labor-techs; plumbing is too simple.

one or two duct discharge to warm air space heating changeover when winter HW is 'FULL' !

Air collectors of 3-layer window screen 3/4 inch separations of an OEM baked 500f charcoal gray. box like that of the Gutter-Pipe SUCCESSFUL air solar guy uses (see youtube) of 1.1/2" poly-foam or equal to keep from roof surfaces overheating air solar collectors.
Clear PTFE layer above screens ~ 3/4"; and solar glazing (Filon (sp?)) was replaced in 12 to 14 years although yellowed early on. Professionally installed under 5000 and provides SPACE COMFORT.

Apr 16, 2013 9:01 AM ET

Edited Apr 16, 2013 9:02 AM ET.

Response to Jon Pierce
by Martin Holladay

Your writing style is hard to decipher, but it sounds as if you are a fan of using solar air collectors for space heating.

This is an old topic. Go ahead and build a few such collectors if you want. They are harmless. But the energy they provide isn't worth the cost of the equipment used to collect the energy.

The basic problem with solar air systems is that storing the energy (usually in a bin full of rocks) uses a lot of fan energy, and the rocks only stay warm for a few days. The other problem with any solar space heating system: when you really need the heat, the sun rarely shines; and when the sun shines, you don't need any space heat (because all you need are a few south-facing windows to keep your house warm if your house has a decent thermal envelope).

Apr 18, 2013 1:23 PM ET

Limited space
by Jeff Auxier

I believe a thermal collector is far more efficient use of available space than a number of PV panels. This is an important consideration in most areas.

Apr 18, 2013 1:39 PM ET

Response to Jeff Auxier
by Martin Holladay

As my article points out, the average solar thermal system with two 4'x8' collectors produces about 28 gallons of hot water a day in a northern climate (annual average production). To produce that much hot water with a PV system and a heat-pump water heater with a COP of 2.0 would require (in a northern climate) a PV array rated at 763 watts. Since typical PV modules produce 12 watts peak per square foot, that means that you would need a PV array measuring 64 square feet -- exactly the same size as two solar thermal collectors.

So the area of the rooftop solar panels is exactly the same, whether you prefer a solar thermal system or a PV system.

May 31, 2013 6:27 AM ET

Solar Thermal vs PV maintenance costs?
by Jonathan Mitchell

Having read back through previous comments I have seen attention drawn to the fact that solar thermal DHW systems are costly, especially so becasue of future maintenance issues. Unfortunately I havent seen anyone mention the fact that Inverters for PV systems don't last forever and will typically need replacing every 10 years or so! These a lot more costly than perhaps a small Grundfos pump that may need replacing or refilling the system with Glycol once every 5 years.

The fact is, Drainback solar thermal systems are easier to install, quicker to install and require almost zero maintenance and are lower cost than pressursied systems. The do not suffer from Stagnation issues and neither do they suffer from Freezing conditions.

As a UK Manufacturer of both PV & Solar thermal flat plate collectors we see benefits with both technologies. And by far the most suitable and reliable system for residential properties is flat plate collectors taking into account normal house roof space availability.

Another thing you fail to mention with PV panels is that for them to work effectively at maximum efficiency you have to ensure there is no shading. If shading encroches on one panel it can knock out an entire string rendering it completely powerless. How many residential roof tops can gurantee no shading at all - either from trees, chimneys, other roofs, other buildings, etc.... Solar thermal is not affected by shading in the same way.

Anyway, the fact is you can have both technologies side by side roof integrated if you want! Solar Thermal for your DHW needs and PV for your electricity needs or selling back to the grid if you benefit from a Feed InTariff.

Great to see this article get so much attention - it clearly demonstrates there is a lot of interest in the US for residential solar - whichever technology it may be.

May 31, 2013 7:00 AM ET

Response to Jonathan Mitchell
by Martin Holladay

It's very hard to gather objective, quantified data on maintenance costs, so most of us have to rely on anecdotes. Here are two anecdoctes from my family:

1. My first inverter (manufactured by Trace) lasted 18 years before it needed to be replaced.

2. My brother's solar thermal system had expensive maintenance problems with its pump, and the tank began to leak after 14 years.

Jun 1, 2013 3:55 PM ET

solar thermal professional weighs in
by Tom Scheel

Excellent article - it is a conclusion I have been forming myself, even though greatly reduces the value of my hard-won solar thermal expertise.

A few quibbles
* you seem to have moved from 63% SF of a 64 GPD government/SRCC figures to 63% of other studies that show the actual figure is 44GPD - but the SF would go up with lower usage - not by 50%, (the difference in water usage) but by much more than zero. I apologize if I have missed where you accounted for that.
(note you check my figures by going to SRCC and looking at a 64 square foot drainback system annual SF for New York, NY (I thought a fair compromise given all the locations discussed) - and further note that SRCC assumes 64 GPD)

That is a quibble for DHW - maybe PV would need to be 150% of solar thermal (worst case for your argument). But where solar thermal really shines (pun 100% intended) is in space heating - and there the math means the south facing roof becomes the limiting factor - as I have often told my customers - PV is great but you would need two rooftops to hold enough PV to heat your house (using electric resistance).

* solar thermal has a steeper curve related to low water temps (PV gets marginally better in colder weather; solar thermal gets markedly better as the delta T between ambient and inlet approaches zero). This means that as we leave the design conditions you specify (ie colder water temperatures in the north or warmer air temps in the south) - all of this is captured in SRCC data (the gold standard of objective 3rd party evaluation of solar thermal - but still a model) - and not captured in the studies based on northern climates. {this is a much smaller issue that the first one I raise}.

I whole heartedly agree in the need for solar thermal monitoring - I do it on all our major systems and we inevitably find ways to tweak better performance because we have the real time and aggregate data - and we can verify for our customers that we are hitting the savings upon which we sold the system. I think the solar thermal industry is really behind the curve in not monitoring every installation.

As for reliability - I find HUGE variability - I have serviced systems with 20 year old pumps, and I've replaced pumps on systems within a year (with both Grundfos and Taco pumps). I strongly suspect the inverter reliability will be resolved, and or dealt with via insurance (share the risk).

Jun 19, 2013 3:20 PM ET

Nice Article!
by Seth Maciejowski

Interesting article. I ran the same calculation recently when deciding on an approach for my renewable energy system. I am running a 4.6KW Allsun dual axis solar tracking system making around 7600Kwh/year and a GE Geospring Heatpump hot water heater for my main hot water loop. My house is in vermont and the temps hit -22F a few days last winter. The heat pump is in my basement and I must say that one of the unexpected benefits of the heat pump is the dehumidfying effect when it runs in heat pump mode. The temps in my basement are mid to high 50s in the summer with very low humidity which is very nice when its' 90 outside and 90% humidity. GE says that the heat pump can operate with a decent COP well into the low 40s. The Geospring can be run in straight heat pump, resistive mode or hybrid mode where it decides what to do in high demand situations. So far I am pretty pleased with this arrangement which has displaced my oil fired tankless hot water system (so direct negation of fossil fuel usage). The fan is MUCH quieter than the old oil fired system and is pretty much not noticeable upstairs. So far, I am very pleased with this approach and am considering an air source heat pump for shoulder season heating when the wood stove is hard to start as I am generating excess kwh off my solar system right now (no reimbursement under net metering rules).

Thanks for the informative article!

Oct 9, 2013 2:25 AM ET

by Samuel Hay

Inadequate description. Worthless is more like it. This truth found in this article is about as rare as a rifle rack in a Volvo. So many less than accurate statements. I could blast them all but in the interest of time, I will be brief. "Heat all your hot water year round from a heat pump". Like if you take the heat to heat your water in winter you will freeze your ass off. Reliability is, well, non existent even now with solar PV ancillary equipment. For what Ive paid to replace inverters I could have bought 3 thermal systems. And then there are the batteries. Storage systems for thermal systems dont eat expensive batteries. I have both systems. Wanna know what works best for me? My passive solar 'thermal' greenhouse. It grows citrus year round in North Georgia. Oranges, date palms, lemons, limes, kiwi, avocado, olives, pineapple(My personal favorite). And doubles as a sprouting base for my spring gardens. It has one small fan. Open windows in spring, close them in winter. Air is drawn into the greenhouse (connected to the house) after raising two operating windows on each side of the entry door about 2". Then heated air is brought back into the house via an open transom window above the previously described structures. The material is twinwall tuffex and in the middle of the swamp, it has never been damaged by flying debris or falling limbs and it is insular. I have over 3000 watts of PV. I have four thermal collectors. But the secret to it all is design and insulation. As for the article, forget what you read, chose what will work for you. I have seen my thermal collectors (Dry) in mid winter at 425 degrees F. Diffuse radiation from clouds and snow works wonders on the thermal panels. Making the most heat when you need it most! I dont have a sophisticated delivery system, only copper finned baseboards where the excess heat flows when the set point for the hot bathing water is met. It is totally bulletproof and WILL NOT BE DESTROYED by EMP from the sun or a nuclear blast. I can stay warm and take warm baths while you are sitting there looking at your fried panels and all the control systems that go with it.

Oct 9, 2013 5:41 AM ET

Edited Oct 9, 2013 5:54 AM ET.

Response to Seth Maciejowski
by Martin Holladay

I will address your points in the order you raised them.

1. You question the reliability of heat-pump water heaters. I addressed your concern in my article; in fact, I wrote, "If you are skeptical about the longevity of heat-pump water heaters, you may prefer to wait a few years before buying one, and to stick with a solar thermal system in the meantime." Moreover, I also performed calculations that show that in many areas of the country, an electric-resistance water heater (with or without a PV system) makes more sense than a solar thermal system.

2. You question the reliability of PV equipment, citing the need to replace inverters and batteries. First of all, very few PV systems have batteries; the vast majority of PV systems are grid-connected. The only homeowners with batteries are off-grid homeowners. It sounds like you live off-grid. If you do, you apparently missed an important point I made in my article: "Solar thermal systems still make sense for off-grid homes." I'm sorry to hear about your inverter problems. My first inverter lasted 18 years; I'm now on my second.

3. I'm glad you have a greenhouse attached to your home. I do, too. They're great -- especially if you like to grow things. No argument there.

4. If you are worried that electro-magnetic pulses will hurt your electronic equipment in some future solar-flare catastrophe, then it makes sense to move to an off-grid location and grow your own food. Good luck.

Oct 17, 2013 1:45 PM ET

Edited Oct 18, 2013 10:49 AM ET.

Parabolic Trough Vacuum Chamber Water Heater
by Andrew Gray

Martin, I wonder if I could get your comments on the idea of a parabolic trough water heater like George Plhak has started:
What say you about this system? I works in the winter because of the vacuum chamber, and it has heat dump capability in the summer because the whole array can be rotated towards the north (to "turn it off").
Andrew Ancel Gray

Oct 17, 2013 2:24 PM ET

Edited Oct 17, 2013 2:26 PM ET.

Response to Andrew Gray
by Martin Holladay

Parabolic troughs, 12-volt motors, gear drives, vacuum tubing, stainless-steel rods, computer-controlled angle adjustments, counterweights, hydraulic hose -- all cobbled together by a backyard tinkerer, in order to make sure his water heater reaches 150 degrees.

Not exactly simple or cheap -- and the net effect is the same as my simple 2-collector system (which has only 2% of his system's complexity).

I'm guessing that the payback period for this Rube Goldberg device is -- oh, I don't know -- about 200 years.

Oct 17, 2013 4:02 PM ET

Parabolic Troughs
by Andrew Gray

OK, Martin. Got it. Your perspective is indeed helpful, believe it or not. One more thing. Tell me how your simple 2-collector system works in the wintertime, for my info. Do you get 120F temperatures in your tank when it is say sunny but 35F outside? Thanks 1,000,000.
Andrew Ancel Gray

Oct 17, 2013 4:08 PM ET

Edited Oct 18, 2013 11:02 AM ET.

Response to Andrew Gray
by Martin Holladay

Clearly, I was being flippant, and was exaggerating to make a point. Parabolic trough collectors are ingenious and useful, especially for utility-scale solar thermal plants. I don't doubt that your backyard invention makes a lot more hot water in the winter than my two solar collectors. And I love backyard tinkerers -- they contribute to the strength of the U.S. economy.

I just don't think that this is a cost-effective way to make hot water.

Oct 17, 2013 6:16 PM ET

Parabolic Troughs
by Andrew Gray

Yeah, I hear you about cost effectiveness. I did manage to put a self-designed heat exchanger into an existing electric water heater (saves about $1200 for heat exchanger WH), and I have an EcoSmart Tankless backup so there is no 2nd-tank-wasted-energy-loss for a single person household (the first two showers are free instead of the 2nd and 3rd). Also, this parabolic system will mount on the North or East side of my home, LEAVING MORE ROOM FOR PV PANELS(!) (You will like that). So I am not displacing PV space. Finally, the materials are cheap for this system. It is the labor required that adds up. It would need an economy-of-scale to get it priced to your liking I am betting. Thanks again for your input.
Andrew Ancel Gray

Feb 25, 2014 7:44 AM ET

Nice post
by Michael Clarke

You gave very useful information. I will keep follow your post. Keep it up.

May 1, 2014 5:11 PM ET

Doing Solar Thermal incorrectly SHOULD be dead
by Kelly Livingston


I think you are misleading people about ST and what it is capable of. What this article should be telling people is that the ST industry is falling very short of what it could do rather than hyping up PV.

1) First, I want to comment just on conversion efficiencies, that is what is the most efficient way to make hot water from the sun. Like most things, the physics can go either way depending on conditions. ST can be anywhere from 80% to 0% thermally efficient at collecting solar energy depending on flat plate vs. evacuated tube and on conditions partial vs. full sun. PV maxs out at 20% for commercial nonconcentrating cells and can go as low as 5% for thin film. PV efficiency doesn't change with light intensity like in ST but shading can really hurt PV without proper inverters/bipass diodes. An important thing to note is that PV performs worse in the heat whereas ST performs better on a hot day (with a low delta T to the hot water especially with flate plate collectors). So yes, if you slap a flat plate collector in Massachusetts or Wisonsin where it will only be 15% efficient for much of the year and compare it to a PV array that is near its max efficiency of 20% due to the cold sunny days, then PV is better. However, if you look at testing of a heliodyne gobi vs a fairly efficient panel that Marc Rosenbaum cited, it's clear to see that ST is roughly 2.35 more efficient than PV which is why you need the HPWH to make up that difference. It's already been challenged that the HPWH wouldn't even come close to that efficiency if it was in a cold basement or sucking the heat out of an already cold house just as you've challenged the gobi wouldn't be able to perform to its tested outputs. I think both claims are absolutely true and there is no real data to say which one would perform better given real world conditions. My vote is on properly designing a ST system.

2) As people have already mentioned, PV grid-tied has free storage so that problem is solved (until utilities start charging storage fees which I'm sure they will at some point). With ST, storage is a critical factor that must properly be designed. This is what people get wrong. There are 2 ways and thus 2 design parameters for thermal storage: temperature and volume. The common 64 sqft collection with 80 gallon storage is a primary reason ST isn't efficient and gets a bad reputation, because this ratio favors temperature storage more than it should which renders collectors more inefficient when solar resources are available and doesn't have enough volume/total btu storage given the load to last more than 1 day. So why do people choose these ratios? Because they are mass produced for other purposes and therefore cheap to contractors. 2 40 gallon hot water heaters and 2 4x8 (common dimension of plywood and other pallet transported items) are typically used in ST applications. This gives a 1.25 gallons/sqft ratio which isn't nearly enough. A ratio of 2:1 would be best and since storage isnt the dominate cost for many of these installations, I'd go for 3-4:1. There is a reason dedicated solar tank companies like have a 119 and 211 gallon model. Just look at it from a common sense point of view, you want 64 gallons of hot water a day from a 80 gallon tank (75% of the volume and at minimum 46% of the btu storage assuming 160 degree max tank temp,120 degree outlet temp, 50 degree inlet temp = 37184 btu/79680btu) Unless the sun shines consistently everyday in your area, of course you'll only get 63% fraction. Its obvious to me why the post from California said he didn't have a problem with his system. The sun shines there more consistently. The BEST thing you can do for a ST system like this is spend the extra $450 on a 3rd hot water tank and you'll get 20-30% increase in system efficiency for 5% more capital cost. Better yet, get a tank that's actually designed for ST and you might get good performance that can easily beat any PV+DHW of your choice.

3) People usually get the tilt wrong. The best way to destroy the performance of a ST system is to lay it flat on your roof, and that's what so many people do. Why is the tilt so critical? Because you need hot water all year round! During the summer when it's hot outside and the days are long, a ST collector doesn't even break a sweat and usually the pumps shut down, the system stagnates, steam is generated, and it lowers the longevity of the panels. In the winter, when its freezing cold and there are a shorter number of sun hours available to capture energy, the collector struggles to get water even luke warm. So what can we do as designers to help our collectors more reliably heat water *every day of the year*? The answer is tilt it with a bias towards winter production! shows my point exactly if you look at "Tilt Fixed at Winter Angle" table. Notice that even tilted at 60 degrees, a collector in Vermont would receive 30% more insolation in the summer than in the winter. Although the intensity of the insolation will be much less which doesn't favor ST production, the summer heat usually makes up for it. You make an incredibly valid point that a SRCC year long production prediciton is almost meaningless. The goal isn't to produce the most hot water in June or through the year. It's to produce enough hot water when you need it all year round.

4) I'm absolutely in support of data collection and system monitoring. I'd go even further and say that a good ST system should be simulated with TMY data if the installer is to do the job right. There is no other way for anyone to learn what you're doing right or wrong unless you get your report card back. Unfortunately, the studies you cite really only proves my point that people don't do ST correctly, and it doesn't show that ST can't collect all the btus that a SRCC test claims a collector should because they were both poorly designed studies. This is evident from the hard data and the sparse experimental descriptions given from your studies. In the Wisconsin/Massachusetts study they got the storage wrong. As you can see from my 2nd point, they underprovisioned storage using the typical 80/64 ratio. They appear to actually attempt to adjust the tilt to the proper amount although it doesn't appear to be at the proper tilt in the ballpark of 60-65 degree, but they used flat plate collectors (I assume single pane but no description is given) in a northern climate. This is a mistake! Twinwall polycrabonate/double paned flat plate, or evacuated tube collectors would have been the proper choice for that climate. Not surprisingly, they had 87% and 93% fraction in the highest month (bet that was summer) and 28% and 19% in the lowest month (bet that was winter where poor tilt and cold temperature got them). The data is vague on energy accounting for pumping as I don't believe the PV pump was included, but the COP for the Wisconsin study is 390 (390 kwh for ever 1 kwh of electricity used). This is not unreasonable for an indirect loop system. The Colorado study made 2 other very important mistakes. While they had the proper storage ratio of almost exactly 2:1, they 1) oversized the system for the load and 2) poorly tilted the collector. For the area of Colorado they experimented, the ballpark year round optimal tilt would be 33.5 degrees. The optimal summer tilt would have been 16 and the optimal winter tilt would be 54 based on the website I shared. They laid the collector flat on the roof at 23 degrees! They didn't even tilt it high enough for year round maximum insolation! So during the hot long sunny hour of summer the panels sat there mostly unused (2.8% energy collected) and again in the winter they did poorly because they were off tilt flat plates either covered in snow or too cold to even circulate. Yet they saw better utilization (7.8%) because they didn't even meet with demand. You can see (not easily) in the energy consumption data that a huge amount of natural gas was burnt in December when the lowest sun angle would have rendered the collector nearly useless. And lastly, they oversized the system 3 fold (20 gallons per day consumption vs 64 gallons per day design parameter) so when you cite 5.7 kwh/sqft/day it is meaningless to the larger point. It should be made clear the system provided 64% of the hot water needs including getting rid of parasitic loses and that's with an incorrect collection angle. Why do you even bring up that study? NREL should be embarrased to even publish it.

5) Your claim about parasitic pump losses is going to be a thing of the past. Pumps for ST systems, like many industries, are either sized for worst case scenario or even oversized. This is horrible for efficiency in most applications but especially so in ST. Since pumps weren't variable speed, you consumed the same power no matter what level of insolation you are receiving. This is even worse for drainback systems since pumps must have a high head requirement (even though it's only needed for 5 minutes until the drainback syphon kicks in) There are single speed pump solutions such as installing a booster pump with check valve in parallel to a low power circulator but I speculate most ST contractors don't do that due to the increased points of failure or ignorance of its importance. So yes, if you use a bad single speed pump, you can expect a thermal COP of 1-20. But the entire HVAC and ST industry is about to change due to variable speed pumps using VFD and ECM motors. These pumps are going to be great not just for ST but all water pumping with variable loads. Condensing boilers that don't operate at their peak efficiency (one commenter mentioned this) will now have circulators that can slow down flow to ensure a lower temperature return and ensure condensing actually happens. Zone valves will last longer since the pressure of the system can be lowered dynamically. Here is a great explanation of that.

So ST has a promising future of very low parasitic pump losses and thermal COP from 20-400. This is the future.

6) I think the economic analysis you have about ST vs PV that is implied by your title and the article itself is the opposite of what is true. First of all, there is NO cheaper way to generate a BTU at less than 200 degrees F than a piece of polycarbonate. We both agree the attached greenhouses are amazing and it's a no brainer! Of course, there is some cost in collecting and storing that BTU in your hot water tank, but the industry has made this stuff overly expensive and for no reason. I don't doubt that it cost 8k for a ST system to be installed professionally, but does it really need to be that expensive? Gary from builditsolar ( built a system for $2k using plywood, pond liner, polycarbonate, water pipe, aluminum soffits, and insulation. That's it. Oh wait I forgot the silicone caulk and the pump. He has a 99% solar fraction and on a good day the COP of his SINGLE SPEED pump is 63 in February in Montana (these are all data driven numbers). His labor is free so I'm not expecting systems to cost $2k but given that he basically put together a system with popcicle sticks, I'm sure an industry with variable speed pumps and years of experience could provide a reliable system for $4k that provides 99% fraction and used minimal electricity. That is if the costumer knew it was possible and demanded this kind of product and performance.

7) PV, on the otherhand, doesn't have the promise of %50 reduction in cost. Thin film relies on many rare earth metals that are hard to mine in massive quantities and are highly toxic. Silicon-based solar cells have dropped in price mostly to speculation ( and there is a limit to the price decrease due purely to physics. The base material of silicon like many metals requires trememndous amounts of energy to produce. Nearly 1/3 of the cost of a module is silicon and almost all of the cost of silicon is energy (sand is pretty bountiful). Just as you'll see copper and aluminum prices track with energy, so will solar cells.

Now clearly since you wrote this article silicon cell prices have gone down even further to .73$/watt yet the installed costs are still in the $4.5 range ($3-$7 depending on region). I think I'd still rather do ST right instead of buy some Chinese made PV.

I look forward to your response, and I've really enjoyed the article and comments even though I disagree with you.

May 2, 2014 8:31 AM ET

Response to Kelly Livingston
by Martin Holladay

I agree with you that some solar thermal system designers and installers do a better job than others. But even when we look at systems designed by top-notch designers and installers, my conclusions hold.

After all, PV systems are governed by the same constraints: some are well designed, and others are not. Yet, on average, PV systems are outperforming solar thermal systems, in terms of dollars saved per dollar invested.

Efficiency alone (BTUs collected per square foot of collector) doesn't tell the whole story. As I noted in my article, not all of the heat that is collected by a solar thermal system can be used.

Changes in net-metering contracts have the potential to change payback periods for PV systems, of course. But PV system costs are continuing to plummet ... unlike solar thermal system costs.

Coincidentally, last week my brother's solar thermal system developed a leak in the expansion tank -- the solar thermal contractor quoted him $900 for repairs -- while my solar thermal system required all of the circulating antifreeze solution to be replaced (due to the fact that the solar panels that provide power to the DC pump were covered with snow on a day that the solar thermal collectors were snow-free -- this situation leading to the fluid in the collectors to overheat). My system is now repaired -- I did the work myself -- but my brother is still mulling whether the $900 investment is worth it.

May 30, 2014 4:35 PM ET

Solar Thermal is Dead article
by Gary Steps

I haven't seen any responses lately. A lot of things have changed in two years. PV costs continue to drop. Grid electrical costs have continued to rise. But from my standpoint as an energy consultant and PH consultant, it is more important that buildings continue to get a lot better.

A client moved into the first Passive House in Missouri last summer. The house is extremely tight and with super insulation compared to modern building codes. For a number of reasons, all of them good, it has no basement, is all electric, and has a maximum heat load of ~9000Btu/h. We laugh about heating with a hair dryer, but we could do that.

Because of the above, the portion of the total heating load associated with domestic hot water is a much bigger fraction of the total than in a "normal" home. If we were to install a HPWH, it would chill the air in the conditioned space. Of course, with a tiny ground source heat pump with a very high COP under normal conditions, it would cost very little energy to fill in for the heat lost to the HPWH. However, a desuperheater in the GSHP would deliver the DHW at a fraction of the energy cost of the HPWH, without either the cost of the HPWH hardware or ant comfort issues with stealing heat from the air in winter.

So - with the addition of a small PV array and a smaller ST array, the building produces 100+% of the total electrical power on an annual basis, and ~80% of the DHW. If there is a need to fill in hot water, an electric coil delivers it at a very low cost of electricity. And the ST unit delivers high quality heat in the middle of winter, when the PV array is sucking wind when it is not covered by snow. .

So - I believe that there are two answers to this discussion - "That Depends", and as we say a lot about the weather here in St Louis, "Just Wait, things will change fast"

May 30, 2014 5:14 PM ET

Response to Gary Steps
by Martin Holladay

As PV prices continue to fall, the cost of solar thermal systems is, as far as I know, remaining constant or rising. So in the months since this article was written, the argument in favor of PV over solar thermal continues to strengthen.

Of course it's possible to meet the domestic hot water needs of a client with a solar thermal system, or even with a combination of a solar thermal system plus PV. But the PV-only route will be cheaper.

I agree that in some climates, for some homes, a heat-pump water heater doesn't make much sense. So in those cases, install an electric-resistance water heater. Once you do that, it's still cheaper to meet your domestic hot water needs with PV than with solar thermal.

Jun 12, 2014 12:51 PM ET

Edited Jun 12, 2014 1:00 PM ET.

All the information on Solar Thermal this article left out
by Gabriel Stinson

I know this article was written a couple of years ago, but I still could not disagree with it more. Where to start?
Under Comparing solar thermal and PV systems, lets go point by point:
-Yes solar thermal has moving parts, no argument there. They consist of one inexpensive circulator for the glycol loop, and a standard 3-way mixing valve.
-Freezing climates is a non issue when you have a pressurized propylene glycol system which prevents any freezing. Drain backs are rarely used in modern systems.
-Solar thermal will require maintenance and so will PV. But propylene glycol for a reputable system (think German, like Viessmann) is rated to last 8-10 years. So over the course of a systems 25-30 lifespan, we're talking 2 single day service calls including cost of glycol, maybe a few hundred dollars each. Versus replacing your inverter after 10 years for easily $1000-1500+
-This article seems to ignore the heightened storage capability of solar thermal tanks. ST tanks must be rated to hold water up to just below boiling, meaning, we can keep pumping solar BTUs into the tank generally loading up to 180°. Despite the tanks super insulation, there are standby losses, but you're still storing and saving potential hot water, even if efficiency drops the hotter you heat the tank. PV may have net metering back to the grid, but no one ever tells you that you're selling back kWh for half the price you buy, because your utility owns the delivery lines, so you won't get paid the delivery portion.
-Pole mounted tracking systems are nearly always prohibitively expensive. I've yet to see the numbers justifying the added expense once.
-This last one is great, "On average, PV systems probably last longer than solar thermal systems." While PV may not have moving parts, you're grossly washing over two pertinent points. Your inverter will fail after 10-12 years, resulting in a costly replacement every time. Most importantly, the silicon wafers used in PV panels degrade over time. Industry standard warranty is that they'll be at 80% in 25 years. Of course, this is only based on lab projections, because none of these systems have been in operation more than 10 years, so hopefully it will be 80%. Conversely, solar thermal collectors have zero degradation, and have been around since the 70s. I've personally seen still functioning 30 year old collectors.

A couple more thoughts:
-The efficiency of PV panels is grossly inferior to solar thermal. We're talking the best that money can currently buy (and most expensive) SunPower E19 modules are 19% efficient. Usually affordable Chinese collectors are more 12-14%. Compare that to high quality ST collectors at 70-80%. Even if you're running the PV in connection with an efficient heat pump at a COP 3.0, which isn't the worst idea, you're netting only 42-57% overall efficiency.
-Not sure why statistics here keep quoting at 61% solar fraction for ST? A residential boiler backup Viessmann system has an 80% solar fraction as rated by the SRCC.

Lastly Cost:
This is the end all point, if you don't have a reasonable ROI then all the efficiency in the world is irrelevant. Depending on your state, there may be incentive money up front in addition to tax credits. At least at the federal level you'll get a flat 30% no cap credit on any solar install, PV or ST. Speaking from experience specifically in New York State, even with NYSERDA incentive money, the 30% federal credit, and our 25% state solar tax credit, the PV payback remains at ~10 years, and that's before spending another $3000 on a heat pump. Solar thermal also has incentive money available from NYSERDA, plus the two credits, which brings the net cost of a system down to $2700 with an ROI of 4-5 years if you're on anything other than natural gas. Even without the $4000 from NYSERDA, with just tax credits that's a net cost of $4500, which still presents a reasonable ROI.
-Lastly I will admit that neither PV or ST makes much sense at current pricing without some combination of tax credit and/or state incentive program, so take advantage of them while they're available.

I invite anyone to dispute my points.

Jun 12, 2014 3:03 PM ET

Response to Gabriel Stinson
by Martin Holladay

Some of the information you provide is accurate, but some of it is flat-out wrong.

You wrote, "PV may have net metering back to the grid, but no one ever tells you that you're selling back kWh for half the price you buy." Net metering contracts vary from utility to utility, but the vast majority of net metering contracts are exactly as described -- the customer pays for the net electricity used, after PV production is subtracted from total electricity used. In other words, the customer is credited for the full retail price of the electricity. Some utilities don't like net metering contracts, of course, but such contracts are common.

I'm not a fan of tracking mounts for PV arrays, and I didn't advocate their use in this article.

You wrote, "None of these [PV] systems have been in operation more than 10 years." My system has been in continuous use for 34 years, and the PV modules are working fine. For more information, see Testing a Thirty-Year-Old Photovoltaic Module.

You wrote, "The silicon wafers used in PV panels degrade over time." This is not true in the case of my own PV module which I tested at age 30; see the link above.

You wrote, "The efficiency of PV panels is grossly inferior to solar thermal." But efficiency (output per square foot) is irrelevant; what matters is cost-effectiveness, and that's where solar thermal falls down. Moreover, not all of the energy collected by a solar thermal system is usable, while 100% of the energy collected by a PV system is usable.

You wrote, "Not sure why statistics here keep quoting at 61% solar fraction for ST." Here is the source: “Cost, Design and Performance of Solar Hot Water In Cold-Climate Homes” (a paper by researchers from Steven Winters Associates).

Concerning the question of whether investments in PV are a good investment, your information is out of date, and the situation is fast-changing in favor of PV. In many areas of the country, PV investments are cash-flow positive from Day One, and the payback situation is only improving.

Jun 14, 2014 1:03 PM ET

Edited Jun 14, 2014 1:28 PM ET.

Response to Martin Holladay
by Gabriel Stinson

Let me state before my responses as someone that previously worked in the PV solar industry, I 100% support PV, I just feel its grossly inferior specifically for the purpose of heating water. Solar thermal is great, but as super efficient at heating water as it is, its mostly a use it or lose it scenario, and obviously its not making you any electricity if that’s the use you need.

You are correct on the net metering part, I guess I misunderstood your statement in the article. Net metering essentially uses the grid as we'll say your battery storage. As you produce more than currently using, your bi-directional meter spins backward as its pushed into the grid. However, if you somehow manage to make more than you use at the end of the year, I am correct (at least in NYS) that your end of the year payout from the utility supplier for excess kWh production will only pay you for supply and Not delivery, so effectively half the cost. That is a unique situation, when PV system design is limited to 110% your annual historical use (again in NYS) that almost never occurs. I guess my response stems from frustration of uniformed people thinking they will be selling their kWh back to grid and making money from it.

Glad to hear you're not a fan of tracking systems; only brought it up because you did in the article.

I honestly wasn't familiar with any residential applications that are 30+ years old, and that’s my oversight. That is fantastic that your panels are still operating 34 years later, but I would say you are one of the few and far between with such a unique system. To that extent, there are also few properly installed ST systems over 30 years old, I just think there are more examples to draw from in the later. Since you've volunteered this info about your 34 year system, I'm genuinely curious at how many times you've replaced your inverter up to now? Approximately how much was a new inverter, and for what size kW system?

Hate to burst your bubble Martin, but PV collector degradation is very real. Check out this study ( conducted over 40 years across 2000 different rates, which clearly shows 0.5%/year. At that rate your collectors should have likely decreased about 17% since their installation, now operating at 83% of their likely 10-12% maximum potential if they're that old.

On efficiency or output per square foot, it most certainly is relevant when you have limited real estate on your southern facing roof. When 2-3 solar thermal panels can match the same kWh output as 7-8 PV modules, there's the best proof of efficiency right there. Your vague statement about how "...not all of the energy collected by a solar thermal system is usable, while 100% of the energy collected by a PV system is usable," I'm guessing is referring to the tank storage capacity. Lets be clear, all the potential energy is being collected and utilized, but efficiency drops as the tank gets hotter. Generally a ST systems is set to a delta of 12°, so the collectors need to be 12° hotter than the tank to get the circulator to turn on. Its relatively easy to heat to 120°, but getting to 150°+ is much harder. All the potential energy is being collected, and you're losing >1° an hour due to normal standby loses. Guess that will drop the efficiency slightly. I'll still take 70+% efficient over 14% any day for heating water. It certainly isn't cheaper buying 3 times more modules to produce the same kWh output.

I'm going to point out that that article about 61% solar fraction is now 13 years. So lets move on. In the meantime, I've attached current SRCC system ratings for boiler backed up Viessmann systems with solar fraction ratings of 96 and 86. Granted that will drop if you're say backing up with an electric element, but if you're going for efficiency, you can make it happen.

I don't think my information is out of date, at most by maybe two years since I left the PV industry, but it certainly is more relevant than that 13 year old article you linked to. I had mentioned the Sunpower E19 module that was 19% efficient, but yes in the last two years they've pumped it up to 21.5%. The added expense of those modules should set you up for a 15+ year payback. Yes PV systems are cash positive from day one, and so are solar thermal systems. The difference is most people don't want to wait 8-10 years for a PV system payback. Once again I'll add that I fully support PV (although the prices still need to come down for greater adoption), but I think its simply a terrible idea to use it for water heating.

Viessmann SRCC system.pdf 787.32 KB
Viessmann SRCC system 2.pdf 787.04 KB

Jun 15, 2014 6:57 AM ET

Response to Gabriel Stinson
by Martin Holladay

Q. "I'm genuinely curious at how many times you've replaced your inverter up to now?"

A. I'm on my second inverter. My most recent inverter cost $1,100; I installed it myself. Total inverter cost over 34 years has been about $50/year.

Solar thermal systems work fine; I have a solar thermal system at my house. If you want such a system, feel free to install one.

My main warning to inexperienced homeowners -- the one on which I base the arguments in my article -- is that the same investment in PV equipment will yield more useful energy than a similar investment in solar thermal. I present the math in my article.

Your costs may vary, but all signs are pointing in the opposite direction than the one you claim: PV is getting cheaper, while solar thermal equipment is not.

So do your own math. You have been warned. If you still want a solar thermal system, it's your choice.

Jul 18, 2014 2:07 AM ET

Looks more and more like PV is the way to go
by Jason D

2 years later and your analysis still holds. PV is even cheaper now. Only way ST makes sense is if you build it yourself or can't get net metering.

One thing I noticed reading the comments is that people think a heat pump only moves heat. This isn't true - the electricity used to run the HP is (mostly) converted to heat as well. So a HP both generates heat like resistance heating and "moves" heat. Several calculations throughout the comments don't account for this. As an example, a HPWH with a COP of 2 that heats water 10000 BTU's is only "moving" roughly 5000 BTU's from the air with the other 5000 BTU's coming from the electricity running the HPWH.

Jul 23, 2014 4:20 PM ET

by Doug Nichols

I know this may be a no-no as I haven't read every comment and may be repeating information. I agree with Martin on this and would also add these benefits:
1. In new super insulated construction you can forego the gas service all together. Where I live it's 10$ a month just to be connected to gas service. Not to mention the initial savings of not installing gas lines. I know many people prefer to cook on gas but new electric stoves are quite nice and indoor air quality improves when we don't burn fuels in our living space.
2. PV can do everything solar thermal can do (efficiency aside) but not vise versa. PV also does so much more-- it can even charge your car at this point. Anyone looking to off set energy use will almost certainly want PV and it's uses are much more versatile and adaptable to meet their needs. Solar thermal panels do one thing, provide hot water. And like noted, especially in hot Utah where I live, they sit idle because they heat the storage tank water in the first few hours of the day in summer.
3. As we hopefully move to a more renewable energy future solar PV dovetails while gas distribution does not. Meaning that there are lots of ways to produce and deliver renewable power over the the electrical grid; I know of no utility scale projects to bring renewable gas to your home. Therefore, it makes more long term sense to invest into the electrical grid and its sequential parts.
4. One professional is better than two. One post commented he only had to replace one 5$ part on his SHWS over the years most people will have to pay a professional to do this-- $60 minimum to have someone come and fix stuff. If you have to have two professionals fix two $5 parts... you get the idea. I think Martin and I are both talking about the mass market here-- not us nerds who would fix it ourselves.
5. The collateral damage potential of a liquid filled system is much higher than PV system.

Aug 17, 2014 5:49 PM ET

ST is dead? Yes, here the proof

2 Years ago I built a electronic PCB that control PV production and home consumption in real time mode. This PCB unit, drives three power relays that activate three 1200W resistors inserted into 200L iron tank ( I studied all details: Hydraulic , mechanical, electric power and electronic, safety certification etc. ) the system is completed with a feedback temperature and command for a recirculation pump connected with boiler.
The first my installation was in May 2013 on customer with 3 trackers ( 5980 Wp) and so far all work very good.
I confirm what Martin says:
a) the trackers works very good
b) Insertion and deactivation of 3 resistors depend by sun irradiation and home loads. All is working in real time and when water temperature is up ( 60 - 65°C) and recirculation pump have worked to equalize temperature of boiler tank and resistors tank, the system enter in standby mode.
c) All energy for heating is produced from PV and the resistors working are modulated in stepping mode.
d) The customer say ( he use GPL for boiler) that more 40% cut the cost of GPL in 1 year!
e) All surplus energy is sended to distributor after water is hot. Therefore the owner
can sell the residual electricity.
f) All domestic machines (wash machine dishwasher etc..) do not suffer of presence of the pv-tank. They have the priority.
g) The European Union and Fraunhofer Insitute are undertake to study a smart system to use energy for heating in more houses with a nearest homes equiped with PV plant.

h) Very easy to install

i) I am grateful for this discussion because 2 years ago I was worried about the project and I did without feedbak on results of this new technology

All my customer are satisfied and the ROI is planned in 3-4 years
thanks a lot

Sep 24, 2014 3:09 PM ET

SHW costs
by j moore

This article should be rewritten or at least revised to reflect recent incentives which bring down net costs for solar hot water systems. Like PV, tax incentives and rebates are critical components in cost calculations. In MA, the net cost of a SHW system for an average home (3 collectors; 80 gallon storage tank) is now below $2,500! The state of MA is encouraging SHW technology for good reasons- it is an accessible, inexpensive, efficient and proven renewable technology. Check out:

Oct 7, 2014 8:52 PM ET

Using solar to heat hot water in radiators?
by Robert Berkman

My wife and I just moved to Rochester NY in a 1929 home with an old oil tank, burner and circulating hot water (not steam) radiators. We are exploring all sorts of options for our upgrade, ranging from just putting in a new gas boiler, mitsubishi hyper heat pumps and/or solar.

One option--would it ever be possible to generate enough heat from photoelectric for circulating hot water in the radiators? (house is 1400 sq feet; lower floor is 1000 sq ft. and we could keep the upstairs turned way down). We also are going to be making our house much tighter after getting an energy audit--thanks! Bob

Oct 8, 2014 4:36 AM ET

Edited Oct 8, 2014 6:25 AM ET.

Response to Robert Berkman
by Martin Holladay

Q. "Would it ever be possible to generate enough heat from photoelectric for circulating hot water in the radiators?"

A. The electricity needed to operate the circulator(s) that are part of your hydonic heating system isn't much. But you are probably also thinking of finding a way to use electricity to heat the water used for your hydonic heating system.

The most efficient way to do that with electricity would be with an air-to-water heat pump like the Daikin Altherma. That would be possible, but a Daikin Altherma system is expensive -- usually in the $20,000 to $30,000 range.

Here's the bottom line: if you decide to install a photovoltaic (PV) system, the limiting factors are usually your budget or the area of your roof. If you can afford an 8-kWh (for example) PV system, and you have a good place to put it, it's possible that you can make enough electricity to meet all your needs. If you install a smaller system, you will probably make just some of your electricity. But there is no way to determine what the PV power is used for. It just reduces the amount of electricity you need to buy. A PV system lowers your electric bill, but the PV system doesn't care what the electricity is used for.

If you want to use electricity for space heating, the most economical way to proceed is to buy one or more ductless (or ducted) minisplit heat pumps.

Nov 24, 2014 8:26 AM ET

Evacuated tube solar collectors
by Eric McKinney

[Editor's note: Eric McKinney is employed by a company that sells evacuated tube solar collectors.]

Given that solar PV and Thermal provide different types of energies the only way to compare these two technologies is by comparing applications that each technology can provide and it’s correlating ROI. While Thermal collectors do not produce electricity, the heat generated by evacuated thermal tube collectors when used in a properly design system far and away beats the return on investment into any PV panel/system on the market. Thermal energy also reduces the reliance on electricity by not requiring the need of electricity to fire up a system.

One solar evacuated thermal tube collector capable of producing 300,000+ BTU's per day (or 1,095,000,000 BTU's per year) would suffice quite nicely for a properly designed residential system for the application of both space heat and domestic hot water and would work as such: First priority: space heat via forced air (electric, gas, geo-thermal) with dissipation heat (left over heat from the system) used for domestic hot water.

This system would be geared toward both applications to maximize the thermal heat generated by this collector. The ROI for this system would be less than 3 years, while a PV installation targeting only hot water would be, at best, 6-8 years and perhaps closer to 10 years.

Other advantages of Evacuated Thermal (Tube) Collectors over PV technology:
1) Solar thermal energy will extend the life of mechanical equipment, which PV solar technology absolutely can not do. This is something proponents of PV conveniently overlook when comparing PV technology to thermal.
2) Evacuated Thermal (Tube) Collectors work from the sun's UV rays not direct sunlight, as do thermal flat panels and PV technology. An evacuated thermal collector capable of producing 300,000+ a day during peak seasons will produce 200,000 BTU's per day, even during the cloudy winter months.
3) Thermal heat applications provide for industrial/commercial and residential space heat, space cooling and domestic hot water with the possibilities limited only by the imagination and creativeness of engineers.
4) Ease of repairs: ETC consist of tubes and there are no moving parts to the collector other than the transfer fluid running through the manifold. If a tube becomes damaged, the collector will still perform at a very high level while the damaged tube is easily replaced (within 5-10 minutes). In the heat of the day an actively UV collecting tube from an ETC would feel, to the touch of a bare hand, like an empty Coke bottle sitting out in the sun (a staple of heat collecting efficiency). PV panels, on the other hand, require substantially more time to disassemble/repair with concern for “hot to the touch” surfaces (a staple of heat loss inefficiencies for all solar panels).
5) Dissipating heat in a solar thermal collecting system is required. Dissipation protects and enhances the performance of the system regardless of how the dissipation occurs – either redirected to another application, via the dissipater of the system or both.

Here is a nice comparison of PV and Thermal: The Indianapolis International Airport has an installation of 44,128 solar PV panels with an “expectation” of producing 16,500,00 kWh of electrical energy annually. See the link for a few more bits of info. The figures for the cost for this installation “range” from $30 - $40 million dollars. While BTU’s and watts are different types of energy, when converting these 16,500,000 Kwh to BTU’s it would require only 685 evacuated thermal collectors capable of producing 300,000+ BTU’s per collector/per day at peak (with sizing for this raw comparison at 225,000 BTU’s per collector/per day) to produce this same amount of energy at a cost of less than 1/12 of the 44/128 PV panels. How about that for an ROI comparison?

My suggestion to all – understand the expectations for your targeted applications to truly weigh the ROI when comparing Thermal to PV. And most importantly, do your research. An evacuated thermal collector capable of producing 300,000+ BTU’s per day does exist and is available.

Yours in sustainable energy ~ EM

Jan 13, 2015 4:25 PM ET

Solar Thermal isn't dead
by Mike Mazzeo

You stated "installing a solar hot water system doesn't make any sense" in the beginning of your blog, I wish you had an asterisk that linked to the end of your article where all of your "buts" were listed.

PV may be a cheaper install, but a limiting factor for PV systems on many homes is the lack of roof space. Most homes in my area of NJ do not have enough roof space to accommodate a PV system that can produce 100% of their existing electric usage let alone switch from gas/oil/propane to an electric HPWH.

If an OG-300 STHW system is used, with an COP of up to .90 versus .63, the economics for STWH become much better and require roughly 80 square feet of roof space versus the up to 170-200 square feet needed in your 1.7-2.2KW PV system size requirements. When roof space is limited, a properly designed and installed STWH will help save a homeowner more on their energy bills. Your math even makes this case but you didn't broach that. And only PV systems installed with ideal tilt and orientation produce at the 1.2-1.3 level you use in your examples. In NJ, few homes have roofs with ideal tilt and orientation. Many contractors new to the industry will install panels on any roof that sees the sun part of the year, further reducing the output per installed kilowatt annually. But most STHW systems, because of the small footprint required, can be installed to ideal tilt and orientation on most homes with full sun exposure.

I question the two cases systems with 63% and 61% solar fractions. What system designs were used? And how can you tout an argument on such a low quantity of samples? We use a separate pre-heat tank for the our STWH designs to provide more solar energy and not compete with the supplemental fuel heat. Most homes have an existing, and functional water heater, so adding the separate solar water storage tank makes the system more efficient. And the cost for a second, or pre-heat tank system is within your estimated install costs in our area. I have average winter tank temps of 80 degrees F plus, reaching 120 degrees with an outside high temp of 20 in February. And I have 2 x 32 sf collectors on my roof that are 30+ years old. PV is guaranteed to produce 80% of what it did when new at that age, and it's starting at less than 16% sun to electric conversion when new.

Summer over-performance and winter lack of performance can be adjusted with panel tilt, favoring the time of year the occupants of the site use the most hot water. Drain back and glycol based systems help negate overheating and freeze damage.

STHW system maintenance is not much when the system is properly designed and installed. Most of the systems that we service having the greatest service problems are those improperly designed and attempting to heat everything (pool, water, space heat). The most reliable are those simply heating the domestic water with anti-freeze and overheat protection planned in the design. We service systems with original collectors, pumps, control often 25+ years old. When we need to replace components, we use only those proven to last 25-30 years in the field, not from a marketing brochure or lab tests.

PV systems have a myriad of problems in their early existence. Inverter failures, rodents nesting under systems and chewing on wires causing shorts and possible fire hazards, first responder and local politicians fear of roofs covered with panels and electric hazards. Also, depending on what percent of the roof is covered with panels and where the damage exists, repair labor costs can become astronomical. Central inverters versus Micro or central with Power Optimizers will have an impact on troubleshooting and loss of savings during equipment failure/maintenance. I believe you are premature comparing 30 year old solar thermal system maintenance histories to several year old PV system maintenance. Repair costs will always be affected by the quality of the initial design and installation and how much the homeowner/system owner pays attention to the system like any appliance.

When do these HPWH achieve COP of 2 or greater? My experience is, like gas mileage on car stickers, they never live up to the hype in the field. The quality of the design and installation contractor will have a major impact on the performance of the equipment. Like any government based program, including STHW in the late 70s, every underemployed contractor with no experience goes where the money is flowing. For years after 1985, our company made a living fixing the horrible installations by inexperienced contractors, and much of that maintenance cost is negatively associated with ST than the quality of the contractor.

We will see the fallout of the quality, or lack there of, from the work of PV contractors soon. I expect you will have to update your low maintenance estimate of PV, not only for equipment failures, but for property damaged by poor installations from inexperienced contractors taking advantage of the government bubble created in the PV market.

And seeing this original article is from 2009, would you have any updates to your maintenance costs estimates for PV? I'd be surprised if they haven't significantly exceeded what you anticipated when this was originally written. I know because we are experiencing a significant increase in PV system service work requests. An overwhelming majority of those systems were installed by contractors who have gone out of business. History is about to repeat itself in the solar field.

Jan 26, 2015 4:26 PM ET

Edited Jan 29, 2015 12:35 AM ET.

Simplified DHW question
by John Schlosser

After travelling in China (incl northern areas) and seeing so very many residential rooftop solar hot water systems, am having trouble concluding that solar thermal for domestic hot water is impractical or relatively inefficient.
Perhaps the key is installing a simplified system, rather than one optimized for everything all the time.
Use the STHW system as pre-heat only; limit use to the 8-9 months without freeze risk; and use simplified overheat protection. Perhaps avoid anti-freeze and pumps altogether, routing DHW supply via the collector only with there's something to be gained thereby.

Can you please comment on availability of such a system off-the-shelf and possible ROI for the Pacific Northwest climate?

PS. Regarding freeze risk, a thermostatically-controlled drain valve is what I had in mind.

Jan 26, 2015 5:17 PM ET

Response to John Schlosser
by Martin Holladay

Whether or not a solar thermal system is cost-effective depends on the local cost of energy as well as the local cost of solar thermal equipment. The analyses in this article were made for North America; energy costs and equipment costs in China are, of course, different.

Solar thermal equipment can be simpler in climates that don't experience freezing temperatures. However, I don't recommend that cold-climate homeowners install systems without freeze protection, with the idea that they will remember to drain their systems for the three coldest months of the year. All it takes is one cold night to freeze plumbing; if the homeowner forgets to drain the system in time, the equipment will need costly repairs.

Mar 4, 2016 1:56 PM ET

Solar Thermal is dead is incorrect:
by Paul Soucy

Solar Thermal is not dead:
The SRCC OG 300 is a computer computation based on the design of the system and a properly designed system can offset 5000 kWh per year.
Over the years I seem many poor solar hot water installations, the design of the solar hot water system and installation is everything.
I always use a drain back system because I found over the last 30 years they are more reliable, I have systems still working over 25 years with no issues.
If I am heating domestic hot water I use separate tank for the solar a 150 gallon polypropylene tank and a thankless gas back up or booster so if we got 3 days of cloud weather the water would first go through the solar tank and then get heated to the desired temperature.
In the summer we have a bypass value and use only the water from the solar tank.
I grew up in a family of 8 my father installed our first solar system in 1980 it used 2 - 4 x 8 solar collectors with a 150 gallon polypropylene tank that last a life time we had pumps fail over 30 years but that is it. The system provided all the hot water for our family. And many of my customers systems are still working to this day and are extremely happy.
A properly designed system the quality of components used and installation has everything to do with the efficiency and payback. A properly designed solar hot water system will have next to no maintenance. Over 30 years we had a 1 pump fail the solar tank is still as good as the day we installed it. We needed to install new collectors when we did the roof over but that was not because they failed. We replaced them because we needed a new roof and the collectors where old.
Whoever wrote the article is way off base, solar thermal is definitely not dead and a properly designed system with good equipment and installation can collect much more solar energy per square foot than P.V. system.
On the high end our drain back solar hot water system installed runs about $8000 on the high end for 2 - 4 x 8 high quality collectors, 150 gallon solar tank and a tank-less booster or backup water heater.

Mar 4, 2016 2:10 PM ET

Reponse to Paul Soucy
by Martin Holladay

Thanks for your comments. Your comments confirm rather than undermine my argument.

Your reported cost for a solar thermal system ($8,000) is in line with my estimated range of $8,000 to $10,000.

Of course solar thermal systems work; I never claimed they didn't. I'm glad that you have had few maintenance problems with your solar thermal system.

Concerning re-roofing: I also had to replace the roofing on my house. While you found it necessary (or appropriate) to replace your solar thermal collectors when your were re-roofing -- an expensive proposition -- I felt that there was no need to replace my PV modules when I re-roofed. I simply dismantled the PV arrays and put the old PV modules right back on the roof when the job was done.

Register for a free account and join the conversation

Get a free account and join the conversation!
Become a GBA PRO!