Thermosyphon Collector: Efficiency, Flow Rate, Vent Size

This page covers some testing done on my thermosyphon solar air heating collector I use to heat my shop. 

The testing covers these areas:


Back to the shop heating collector main page...

Effect of Reducing Inlet and Outlet Vent Size on Collector Performance

I use this thermosyphon solar air heating collector to heat my shop.  Its a very nice, simple, cheap, easy to build, and efficient way to provide space heating.  A number of people have built variations on this collector with good results.

solar air heating collector

About the only downside to this collector is that it requires fairly large inlet and outlet vents in each collector bay in order to work efficiently.  These vents are time consuming to cut through the wall, and have a visual impact on the inside space.   This brings up the question:

How much does thermosyphon collector performance suffer as vent size is reduced?

If smaller vents could be used, it would also be possible to use a large hole saw (6 or 8 inch) to cut the vent holes, which would save a lot of time.

This test is an attempt to answer the question on how much you can cut back on vent size and still get good performance.   The bottom line answer appears to be not very much.


Test Setup

The collector has 5 identical 4 ft wide bays.  Two adjacent bays of the collector are used for the reduced vent area test.   The West bay is the reference and always has the standard two inlet vents at 4 by 18 inches each and two outlet vents at 4 by 18 inches each. 

The vents on the East bay were tested in these three configurations:

- Config 1: Both the inlet and outlet vents are masked down to a width of 7.5 inches so they are 4 by 7.5 inches instead of 4 by 18 inches.  This approximates the area of a 6 inch round duct.

- Config 2: Both the inlet and outlet vents are masked down to a width of 12 inches so they are 4 by 12 inches instead of 4 by 18 inches.  This approximates the area of an 8 inch round duct.

- Config 3: Same as Config 2 except that one inlet vent and one (opposite) outlet vent are completely masked off.  This approximates going down to just one vent per 4 ft bay.

The standard vent provides 1 sqft of inlet and 1 sqft of outlet vent for the 4 ft wide bay.

Config 1 provides 0.42 sqft of inlet and 0.42 sqft of outlet for the 4 ft wide bay.

Config 2 provides 0.67 sqft of inlet and 0.67 sqft of outlet for the 4 ft wide bay.

Config 3 provides 0.33 sqft of inlet and 0.33 sqft of outlet for the 4 ft wide bay.

In all cases, all the measurements needed to calculate heat output are taken on both the bay with the reduced size vents and on the reference bay -- this allows an accurate estimation of the heat output drop due to the reduced size vents. 


Thermosyphon Collector Design Rules
The "standard" vent size is based on the design ground rules for thermosyphon collectors that I believe were formulated by Steve Baer many years ago.   The rules are:

- The vent area should be at least half of the cross sectional area of the collector.  So, if the collector is (say) 6 inches deep and 48 inches wide, then the cross sectional area is 6*48 = 288 sq inches, and the vent area should be half that, or 144 sq inches.  My collector is exactly sized to the rule with two 4 by 18 (72 sqin) inlet vents and two 4 by 18 outlet vent per 4 ft wide bay.

- The depth of the collector should be at least 1/15th of the height of the collector.  So, a 96 inch high collector should be about 96/15 = 6.4 inches deep.  My collector is just short of meeting this ground rule.

- The absorber should be a flow through design with low air resistance but lots of surface area.  I use 2 layers of metal window (insect) screen, but others have used multiple layers of metal lath, furnace filter media, ... 

- The airflow path should be as smooth as possible to reduce air resistance.  I did worry much about smoothing the flow paths -- this would be an interesting area to explore.


From the inside showing the inlet and outlet vents.
The right two vents and stud bays are the West (reference) 4 ft
collector bay.  The next two stud bays are the East (test)
4 ft wide collector bay where the reduced sized vents are used.

One of the exit vents masked down to 7.5 inches wide.
This provides about the same area as a 6 inch round duct.



The low mass thermistor on the right logging outlet temperature.
The Kestrel wind turbine to the left measuring outlet velocity.

The Apogee pyranometer mounted on the same plane
as the collector.

A nice day.

A nice clear day.

This shows the snow cover in front of the collector.
Normally by this time of year, there would be better
snow cover with a corresponding increase in reflected
sunlight and increased heat output.

The three reduced size vent configurations tested:

Config 1: all vents masked off to 4 inches high by
7.5 inches wide.  About same area as a 6 inch round duct.

Config 2: all vents masked off to 4 inches high by
12 inches wide.  Same areas as an 8 inch round duct.

Config 3: diagonal 4 by 12 inch vents. 
Only 1 inlet and 1 outlet vent.



The table below summarizes the results.  It provides heat outputs for both test bays for the three reduced size vents that were tried.  Temperature rise from inlet to outlet, and outlet vent velocities are also provided.

  Temperature Rise (F) Velocity (fpm) Heat Output per bay (BTU/hr) "Efficiency" (%)
Vent Configuration Small Vent Standard Vent Small Vent Standard Vent Small Vent Std Vent Small Vent Std Vent
1: Simulated 6 inch ducts 80.1 68 177 127 4826 7055 44.4 64.9
2- Simulated 8 inch ducts 67.6 69.7 148 118 5449 6719 50.3 62.1
3 - Diagonal 8 inch ducts 75.3 65.9 184 123.5 3773 6649 35.2 62.1

It appears from the testing that any significant reduction in vent area does decrease the heat output and efficiency of the collector.   The best small vent option is the simulated 8 inch diameter ducts for vents, but this still results in a nearly 20% drop in heat output from the standard vents.  Going down to the equivalent of 6 inch diameter ducts cuts the heat output by about 32% from the standard vents.

Using only one inlet and one outlet vent per collector bay with an area equivalent to an 8 inch round duct cuts the heat output by 43% from the stand vents.

As the vents sizes are reduced, the extra air resistance of the smaller vents reduces the collector flow volume.  Since there is less air flowing through the collector, the air heats up to a higher temperature, which increases the thermosyphon effect, and increases the air velocity through the vents -- this helps to make up for the smaller vent area, but, as the heat output numbers show, its not enough to compensate for the smaller vent area.

Another way to look at this is that the lower collector flows and higher collector temperatures that the small vents cause result in a hotter collector that loses more heat through the glazing to the outside.

The chart above is an attempt (possibly a stretch) to generalize the effect of reducing vent area on heat output.  The plot shows the drop in relative heat output as the vent area is decreased compared to a standard vent area.  The standard vent area follows the design ground rule stated in the table above.  For example, cutting the vent area in half reduces heat output to about 75% of what it is with full size vents.

Based on the results above, I would expect that increasing the vent size from the "standard" vents would yield some further increase in heat output.

This is logger plot for temperatures and sun intensity -- the table above is based on values from this plot and from hand recorded vent velocities.

Explanation of plot:

Standard vent outlet temperature (F) -- solid black

Small vent outlet temperature (F) -- dashed red

Inlet temperature (F)  -- long dash green

Solar intensity (watts/sm) --  solid blue

Time Line:

11:20 am to 11:44 am -- Small vent bay has both inlet and outlet vents masked down to 4 inches by 7.5 inches -- about the same area as a 6 inch dia duct.

11:48 am to 11:54 am -- Small vent bay has both inlet and outlet vents masked down to 4 inches by 12 inches -- about the same area as a 8 inch dia duct

12:02 pm to 12:10 pm -- Small vent bay has one inlet and one outlet at 4 by 12 inches with other inlet and outlet blocked -- about the same as having one 8 inch duct per bay.

Note that the major dip in temperature on the red line at 12:03 pm was caused by the temperature sensor slipping out of position for a few minutes. 

Ambient temperature started at 43F at 11:30 am and went up to 44F by 12:45 pm.

An air density of 0.061 lb/cf is used -- this is based on our altitude of 5000 ft and an average temperature in the collector of about 85F.  This calculator is used for the density estimate...

The velocity measurements are taken in the center of the vent area using a new Kestrel turbine anemometer.   To account for the drop off in velocity toward the edges, I have factored down the center velocity by 7% -- the 7% is based on measuring velocities toward the edge of the vent with the same Kestrel meter.

Solar intensity measurements are taken using the logged values of a nearly new, calibrated  Apogee pyranometer. 

Heat output for the bay in BTU/hr is calculated as:

Heat Out = (Vent Area)*(Vent Velocity)*(Kvp)*(Air Density)*(Tout - Tin)*(60 minutes/hr)

Vent Area is the total outlet vent area in sf

Vent Velocity is the center of vent velocity in ft/min

Kvp corrects for the velocity profile of the air leaving the vent -- I am using 0.93 based on measurements around the edges.

Air Density is the air density in lbs/cf -- sea level is 0.075 lb/cf, but for our 5000 ft altitude and average collector temperature it drops to 0.061 lb/cf

Tout is the exit vent temperature in degrees F

Tin is the inlet vent temperature in degrees F

60 min/hr converts the BTU/min to BTU/hr


Measuring Collector Efficiency

In addition to estimating the effect of reduced size vents, another aim was to get as good an estimate for the collector efficiency as possible with the instruments I have.   I took the following steps to try to get good efficiency estimates:

- Temperatures were measured with Onset Computer thermistors having a specified accuracy of 0.25F over the range of interest.

- The outlet vent thermistors are low mass, faster reaction versions to reduce lag problems.

- Sun intensity measured with a new, calibrated Apogee pyranometer mounted in the plane of the glazing half way up the collector.

- Very consistent and steady sun on this day, and measurements taken in stable midday conditions.

- Several trials for each measurement point -- these showed good consistent results.

- Air density corrected for altitude and temperature.

- Air velocity measurements taken with new kestrel turbine anemometer (each anemometer is factory checked for accuracy).

- Velocity profile over the vent area is accounted for using a correction factor.

One factor I have not taken into account is any dependency of the Kestrel velocity reading on air density.  Our air density at 5000 ft elevation is down about 17% from the sea level density.  Since the Kestrel is a turbine driven by aerodynamic lift and drag forces, one would think that the velocity reading might be effected (reduced) by our lower air density.  Kestrel responded to an inquiry on this saying that they do not believe the velocity readings at altitude are effected, but I'm still uncertain about this.  If you have knowledge or experience in this area, please let me know what you think.    If the velocity readings are lower at altitude, then the effect would be to increase the heat output and the efficiency from the numbers in the table above. 
Note: Kestrel answered this question, and have convinced me that because the turbine does not extract any energy from the flow (unlike a wind turbine generating power), and the bearings are very good, and the turbine weight is small --that the drop in air density with altitude does not effect the rotational speed of the Kestrel turbine. 
Thanks to Nick Pine and Ben Neilsen at Kestrel for getting this question answered.

Of the sources of error, the average vent velocity is probably the most significant.  Measuring air velocity accurately on any solar air heating collector is difficult, and the relatively low velocities over large vent areas make the thermosyphon collector measurements even more difficult.

In the end, I think that the efficiencies in the 62 to 65% area under the moderate winter conditions are reasonably accurate.  I also think that this kind of efficiency in this very simple, very cheap collector is not much short of simply amazing, and are quite competitive with high quality commercial collectors.

One thing to keep in mind is that efficiency varies with with solar intensity and with the temperature difference between the collector and the ambient air (and a few other things), so the 62 to 65% efficiency measured in this test is just one point on an efficiency curve -- a point that is typical of sunny day moderate winter conditions.   On warmer days the efficiency would be higher and on colder or less sunny days the efficiency would be lower.  

If you have any suggestions, comments, corrections ... I'd like to hear them.

Measuring Flow Rates

Its interesting to note that the flow rates achieved by this collector are quite respectable even by fan forced collector standards.  The flow rate for one 4 ft wide bay of the the standard vent size is:

Flow Rate = (Vent Area)*(Vent Velocity) = (1 sf)*(127 ft/min) = 127 cfm

This gives a flow rate per sqft of collector of (127 cfm) / (32 sf) = 3.9 cfm/sf    -- this is higher than most fan forced collectors achieve.

Using The Back of Collector as Large Radiant Heating Panel

Since the back wall of the collector IS the south wall of the shop, and since the back wall warms up when the sun is shining on it, there is some heat transfer from the collector directly into the shop that is independent of the main heat transfer through the vents.

The IR picture just below shows the south wall of the shop with the collector operating.  The wall is heated to about 85F at the top, 78F in the middle and 75F near the bottom.  The non-heated shop wall temperature at this time was about 55F.

So, there is some heat transfer to the shop from the back wall of the collector by air convection over the heated wall and also by radiation into the shop space.

I could not quickly find a good method to calculate the heat output of the big wall radiator, but a quick look indicates that the air convection alone is more than 2000 BTU/hr.

Note that when I did the efficiency measurements shown in the section above, I did insulate the back of the collector so that all of the heat output would be via the vents.


IR picture -- back of solar collector

The downside of leaving the wall behind the collector uninsulated is that the heat loss is greater when the sun is not shining on the collector.  The R value of collector glazing plus the wall itself might be about (R1 glazing + R1 wall air films + R1 wood) = R3 -- about the same as a low-e, double glazed window. 

So, in most circumstances it probably makes sense to insulate the wall behind the collector.  The exceptions might be if you don't care how much the room cools when the sun is off the collector, or if you can work out a movable insulation scheme so that the wall can be insulated when not collecting.

January 3, 2012