The only reasonable alternative is low-temperature radiant heat, which we ruled out without really doing a fair comparison.  Our underlying assumptions were that it would be more expensive and offer no performance advantage over the fan-coil.  An advantage of the radiant system is that it would have avoided the overhead of powering a big fan.  We probably should have looked into further.

For a hot water system, we used a Polaris high efficiency, gas fired, sealed combustion power vented unit, that delivers up to 130 thousand BTU per hour, which based on our projections is much more than we would ever need for combined heat and hot water, even if its zero degrees Fahrenheit outside.   Our heat loss spreadsheet predicts a heat loss of about 10 thousand BTU per hour on a typical Seattle winter day (about 45 degrees outside), and about 30 thousand BTU at zero Fahrenheit.  While the Polaris is a nice unit, at $3000 installed its outrageously more expensive than a typical gas HW tank, which is maybe $500 installed. To be fair, the installation is more difficult because the Polaris needs a fresh air intake (2" pipe) and an exhaust (also a 2" pipe), and there is the added cost of a fan built into the tank.  The Polaris is also a condensing type system which results in about 12% greater fuel efficiency compared to a standard "mid-efficiency" tank.  In spite of all this, its still hard to understand why it should be six times as expensive, and it would seem as if there was a market opportunity for someone there.

Sitting next to the Polaris is a fan coil unit, which is essentially just a coil of copper pipe sitting in a duct with a fan built into it.  An external pump moves the hot water from the tank through the heating coil.  Finally, next to the fan-coil is a Venmar HEPA3000 heat recovery ventilation unit (HRV), which exchanges fresh outside air for stale inside air while also filtering some of the re-circulating air.  Cost wise, it doesn't appear that the fan coil unit is particularly cheaper than a standard forced air furnace, and HRV units seem unreasonably expensive to install.

Controller
The heating unit must turn on when the house feels cold and the ventilation system must provide fresh air when it is needed.  For heat, air temperature is generally a good enough measure (although relative humidity and radiant temperature strongly affect the temperature we perceive also).  For ventilation, no one sensor provides a good enough measure of when fresh air is needed: when there is too much carbon dioxide, too high a relative humidity or too many particulates or other airborne contaminants.  One option is to use just a carbon dioxide sensor and rely on manually operated exhaust fans to solve the other problems.  Another alternative is to run the ventilation system for a certain number of hours a day, or a certain percent of every hour.   While the later method tends to over ventilate, the use of an HRV keeps the energy penalty low, and over ventilation guarantees good indoor air quality and the simple timer system is more reliable than any currently available combination of sensors.

Before deciding on a controller, we need to figure out whether the fan-coil fan needs to run along with the HRV fan.  This is mainly determined by how much duct resistance we have, and whether the HRV will deliver enough air given this amount of resistance.  Some sources imply that both motors have to run at the same time, but this makes ventilation use much more energy since the fan-coil fan uses quite a bit of energy (exact number TBD, but guess its around 350 watts).   Alternatively, other sources indicate that it is common to run an HRV continuously (24 hours a day), completely independent of the heating fan, but that also is a huge energy use. 

We actually purchased a aircycler thermostat, which provides timed ventilation along with heating, by supplying ventilation during the heating cycle when possible, and its designed to run the main fan for ventilation.  For more information on this see www.aircycler.com.

Analysis

In re-examining the decision to use an HRV for ventilation, it is no longer clear that it makes sense to use an HRV in our climate, or at least not the one we were given (Venmar HEPA3000), because its an energy hog, probably due to providing HEPA filtration, which is something we don't really need.  In addition, it would seem like the aircycler isn't the best ventilation controller because its inconvenient to change how much ventilation it provides (you have to unsnap it from its base, and snap it back in holding down a key so that you enter "setup" mode).  This is important, since the amount of mechanical ventilation needed throughout the year varies significantly, because the natural infiltration also varies (the actual need is determined only by the number of occupants).  It would seem like any controller that took into account outside temperature would make a better guess than any fixed duty timer, although without resorting to a plethora of carbon dioxide and moisture sensors.

By simple calculations, it would seem that the HRV saves more energy than it uses, but its more complex because your trading electric for natural gas, and even if you ignore that fact, the best way to save energy is to only ventilate how much is necessary.  On cold days, when the HRV pays back the most, there is actually less mechanical ventilation necessary because there is greater natural ventilation.  On warm days, when its pays back the least, you need a lot of ventilation.

While I have yet to find full data, our HRV uses 232 watts on the high setting (110 CFM fresh air and 270CFM of recirculating air), and so I'd guess it uses at least 150W on low (70CFM fresh, 180CFM recirc).  If run 24hrs at 150W, that's 2.6KWh a day, or about 3 times what an energy efficient refigerator uses, so it would be the largest single electrical load in the house.  By comparison, the heat recovery of 70CFM of air on a typical cold Seattle day (avg 45F, 70F inside) is about 10KWh, but when its 60F, the heat recovery is only about 4KWh.

Heat and ventilation in the apartment
We wanted to avoid sharing a ventilation system with the apartment to avoid cross contamination of cooking odors and so the apartment has a separate heating and ventilation system.  It heat is provided by two wall mounted fan coil units by Turbonics (feedback on their noise level when they work).  We would have liked to put a second HRV in the apartment, but we couldn't find anyone who would put one in for less than two thousand dollars, which was beyond our budget.  Rather than use the standard window slot for ventilation, whose ventilation varies with weather, we ran a duct to the attic from near the back of the refrigerator and capped it off with a damper that only allows air in.  Because the duct runs straight up, the theory is that the stack effect will prevent the wind from going down the duct, except when the exhaust fan is running, which creates a negative pressure which opens the damper and draws fresh air down the  duct (and hopefully cooling the refrigerator coils in the process, making it run more efficiently.  The colder it is outside the greater the pressure of warm air in the duct and so the more wind gusts it will keep out. As the outside temperature nears inside temperature, any small breeze can open the damper.

Spot Ventilation
We used Panasonic super quiet high efficiency fans in every bathroom and in the laundry room as well (and the garage).  While its somewhat frustrating to have yet another system of ducts (and more holes in our building envelope due to them), there would seem to be no better answer, since a whole house ventilation system performs a very different function than a spot ventilator.  Each kitchen stove also needs a range hood: in the apartment we used a over the range microwave exhaust fan combination, while in the main unit we used a ceiling mounted Broan lo-sone 300.  Although range exhaust fans are generally only effective when mounted no more than 30" from the burners, we don't cook (or eat) meat, and in fact fry very rarely, and cook most things covered.  If we have problems we will clearly have to move the fan closer to the stove top, but because we wanted a strong connection between the kitchen and the dining space, we are hoping we don't have to do this.

Including two dryer vents, four bath fans, one laundry fan (note: the apartment has is washer/dryer in the bathroom and so has a combined bath/laundry fan, although it is our experience that laundry fans are rarely if every used) and two kitchen fans we have nine extra holes in our building envelope.  Combined they represent approximately 162 square inches of voids in our envelope, or a little over one square foot.  Each ducts is also a possible air leak.  From the point of view of heat loss, it would clearly be better to combine some of them into one duct, but the ducts all have to be sized so that every fan can run at the same time, even though in reality that is very unlikely.

Utility room (far left), isn't as insulated as the rest of the house.  The ADU is heated with a wall mounted fan-coil unit (center).  Spot ventilation is all with Panasonic ultra quiet fans. The dull gray area in the bend is duct mastic sealing the joint.

Ducts
One significant drawback of our heated utility room with the unheated basement space was that many of the ducts had to be run in the basement ceiling to feed first floor rooms, and even with 12" TGIs to run the ducts between, each duct displaced a lot of insulation.  They were all installed directly next to the subfloor so that every duct is totally within the heated space.  When we had to run ducts perpendicular to the floor framing, we had to build insulated soffits to contain them, and currently these soffits are pretty under insulated: in some case having only an inch of so of insulation between them and the unheated basement.  We will probably cover those soffits with 2" of foam board later on to remedy the situation.  

Each duct run was designed to have an adjustable damper on it so that the 800CFM of air coming out of the fan coil unit is distributed evenly to all to rooms.  When space constraints made this impossible, we agreed to use registers with dampers on them,  but it remains to be seen whether this system will work correctly.

In spite of attempting to provide places for ductwork, getting them all in was a tight squeeze, and in particular the return ducts where a problem.  Our initial design called for only one big return duct to get around this, and we attempted to find a way to get one return per floor, but once you reduce the number of returns, they each have to be quite big, and getting a duct that big  was just impossible in a 2x4 wall (total return size is 200 sq in of cross section, so the minimum size to make it work it would be about 75 sq in).  We ended up putting all 200 sq in in the living room, near the bottom of the stairway, with the idea that we would put a 4" x 48" grill in the bottom of a bookcase, but its unclear whether this will work well or not.

Solar HW system
We added a Thermomax evacuated tube solar collector to our system, installed as pre-heat tank.  The Thermomax system is quite expensive, but in the cloudy northwest only an evacuated tube system will work well in any season other than summer.   We installed the system on the sunniest part of the roof we had, and run copper pipe from the collector down to the storage tank, which along with the pumps and controller is in the basement utility room.  The total system cost for a twenty tube system installed is around five thousand dollars, from Puget Sound Solar, www.pugetsoundsolar.com.

Finding space for ducts is often difficult--in this case the 2x4 walls were too small to extend the 12" round return duct to the second floor, so the only return is two 12" rounds, in the floor near the stairs (right).

Miscellaneous Issues
In order to keep dust out of the ducts, each joint is thoroughly sealed with duct mastic, not duct tape which fails after a few years.

Although our utility room is insulated, its only with a 2x6 wall (ie 40% less than the rest of the walls), and the floor of isn't insulated at all (although the slab does have perimeter insulation).  Our intent is to put pipe insulation on all the exposed H/W pipes, and put a blanket on the Polaris Unit and the Solaraide storage tank.

The return air duct appears to be sitting on the concrete floor, which is the only uninsulated part of our utility room.

We guess that the cost of the overall system (heat, hot water and HRV) is at least 30% more than a more standard system, and could be as much a twice the price of a simpler low end system.

To find out what we thought one year later, click here.