Heating Upstairs Versus Downstairs – Part III

This is the third and last blog post in the series on how and why most of the heating in a two-story building (with open stairs) happens downstairs.

The magnitude of the phenomenon has been so fascinating to me, with so much of the heat coming from the first floor, in buildings kept at the same temperature upstairs and downstairs, I felt I needed one more solid set of measurements.  So, I went to a FIFTH house, heated with electric baseboard heaters with wall-mounted thermostats, perfect for measuring heat output by room and by floor.  The result:  With open doors, 73% of the heat comes from downstairs.  With closed doors, 58% of the heat comes from downstairs.  And with the heat turned off upstairs, in other words 100% of the heat coming from downstairs, and the doors closed, the upstairs air temperature in a bedroom dropped by 4.6 degrees, from 70.1 F to 65.4 F, on average.

To try to understand this phenomenon, let’s go back to the first scenario, with both stories being heated separately, and all doors open. If both stories are being heated, why would there be any warm air to rise from the first floor, and why would there be any cold air to fall from the second floor?  The answer is that within the vertical height of each floor, there is a temperature variation:  The temperature increases from floor to ceiling.  See the bar graph above. In a limited number of measurements, I found a few patterns:

  • The floor-to-ceiling temperature differences are far greater on the first story. Where the floor-to-ceiling variation upstairs is typically in the 1.5-2.5 degree range, downstairs it’s a much more pronounced 4-6 degrees, typically.
  • Heating with higher-temperature heat sources (like a wood stove) causes higher temperature variations, just as we might expect.
  • Ceiling temperatures at the first floor are in the 73-80 degree range, and floor temperatures at the second floor are in the 68-72 degree range. As these air masses meet on the open stairway, the warm first floor air at ceiling level rises, and the cold second floor air at floor level falls.  This is exactly as I found with smoke testing, in a prior post.
  • Interestingly, while there does not appear to be a very cold surface layer right at the floor (let’s say within a couple inches of the floor), or conversely a warm layer right at the ceiling, there does appear to be a more pronounced temperature drop below about 24” above the floor, and a slightly more pronounced temperature rise above 70” above the floor. Air temperatures between 24” and 70” are slightly more uniform.  This is evident on the bar graph, above.

In my first blog post in this series, I eyeballed the magnitude of air rising and falling at the stairs as being in the hundreds of cubic feet per minute (CFM).  To help orient ourselves, recall that a bathroom exhaust fan draws about 50 CFM, a kitchen exhaust fan maybe 100 CFM, and a central forced air furnace circulates about 1000 CFM.  I took a very rough measurement by examining the velocity of smoke being drawn downstairs in a home heated with hot water baseboard radiators.  What did I find?  350 CFM.  That is a LOT of airflow, falling downstairs.  And we must assume it is supplemented by air falling in other locations, too: Through plumbing chases, through floor-to-floor holes for piping and wiring, and more.  And it is complemented by warm air rising, simultaneously.

Let’s say that 400 CFM of cold air falls, and 400 CFM of warm air rises, and the average difference in temperature between them is conservatively 5 degrees. That makes about 2200 Btu/hr in “cold” (heat load) being transported from upstairs to downstairs, and vice versa 2200 Btu/hr of heat rises from the first floor to the second floor.  The key is that this happens 24 hours a day, not just when the heating system is firing.  In fact, it is probably higher when the heating system is firing and so causing higher temperature variations.

It might be helpful to have a term for this phenomenon of warm air rising and cold air falling.  “Stack effect” does not quite do it justice, because we have so come to associate stack effect with warm air rising (like combustion products in a chimney stack), and with associated induced infiltration and exfiltration.  We must capture the falling cold air portion of the phenomenon, it is so critically important, and it is something we have missed for too long.  I have heard the phenomenon referred to as convection, but convection can encompass both natural convection (which is what we are talking about) and forced convection (fan-induced).  Convection also overlaps into the field of convective heat transfer, which is not what we are talking about.  I myself like the term thermosiphon, which we have historically applied to passive solar hot water heaters, where flow is driven vertically up by warmth, and driven down by the weight of colder water entering the system.  And there is a relationship between the upward warm flow and the downward cold flow.  And there is no fan or pump involved.  So, I suggest we call this phenomenon of rising warm air and falling cold air, interior to buildings, “building interior thermosiphon air flow”.  Rolls right off the tongue, doesn’t it?  Let’s use that as a placeholder unless a better term comes along, or if somebody knows of a better term already in use – please let me know.

By the way, a colleague at work reported anecdotally that in his house, heated with radiant floor heat, he does not see effects of thermosiphon air flow. It makes sense that the effects would be significantly less, or even none, with radiant floor heat.  The hot water temperatures are far lower with radiant floor systems. Also, since most (or all) of the floor is a radiator, almost all falling cold air comes into contact with the heat source, the floor. And as soon as it touches this heat source, it becomes a load on this heat source.  And so falling air on the second floor, as soon as it touches the floor, remains a load on the second story heating system. It never reaches the first story, and most importantly the heating system on the first story, as cold air.  In addition, it is heated by the floor, and so is given the buoyancy it needs to rise again.

So, that’s it for this blog series on heating upstairs versus downstairs.  In summary, over 70% (and in most cases over 80%) of the heat of two story buildings, with open stairways, and separate heating controls upstairs and downstairs, ends up being provided by the downstairs heating distribution system. If doors are shut, such as upstairs bedroom doors, the percentage of heat provided by the downstairs is slightly lower, but is still way over 50%.  And if the upstairs heat is turned off, even with doors closed, enough heat rises and reaches the upstairs rooms to maintain temperatures within 4-8 degrees of downstairs, in general. If there are not separate controls upstairs and downstairs, the upstairs tends to overheat, and the downstairs is at risk of being too cold.

I hope we sometime get to more controlled tests.  My tests have all been rough, and complicated by real-world building dynamics.  Although, with five buildings examined, I am now confident of the general magnitude and impacts of this thermosiphon air flow.   A friend even suggested that good CFD analysis (computational fluid dynamics) would give us additional insights into this phenomenon, and I agree.  Wouldn’t THAT be fun?  We could really get into parametric analysis of variables, one by one, to look at the impacts of different kinds of heating systems, heating system temperatures, outdoor air temperatures, construction details, and much more.

Once I gather all my data and thoughts, full results of this study will be published later this year in Home Energy online, and then in a subsequent hard copy issue of Home Energy Magazine.   Home Energy is one of the finest and most authoritative sources of information on residential energy, including both single-family buildings and multifamily buildings.  If you don’t already subscribe, believe me, it’s worth it.  In a shaky new world of made-up facts, half-facts, and alternative facts, Home Energy is full of just plain old real facts.


2 thoughts on “Heating Upstairs Versus Downstairs – Part III

  1. Not sure this is that new. I know a number of people that use this principle to effectively heat the upper stories using an appliance on a lower floor. we heat our two story house with a wood insert located in the living room. A couple of days ago when the outside temperature dropped to 10 degrees our second floor thermostat was reading 62 degrees. Our master bedroom is located directly above the living room. We tend to sleep with the door closed. Yet thanks to conduction our bedroom is relatively comfortable. By the way, we also use electric heaters on timers in our bathrooms.

    By the way in physics the phenomena is called thermally induced air currents.


    1. Very interesting. I like the data in your description. I also agree that it’s not new, but I don’t think the magnitude of falling cold air has been adequately recognized, we have been so preoccupied with rising warm air and with infiltration aspects of the stack effect. Note also that you say “due to conduction our bedroom is relatively comfortable”. Our finding is specifically that this is NOT primarily conduction, that even with doors closed there is significant airflow, both warm air up and, as significantly, cold air moving downward. Thanks again for reading and commenting.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s