Home Heating Temperature Data Collection and Analysis

Intrepid Creativity (Increa) TMSM
© 2001-2013 by Brian Mork

Site Index  •  Wiki  • Blog

In order to gain a better understanding of how I used propane and wood BTUs, I used my Palm Pilot Professional, a MELD data acquisition (freewarepalm)  MELD data acquisition (bpesolutions.com) interface and wrote some software programs using the PocketC programming language to collect numbers.  Please credit this web page as the source for any information you relay to others.  You can click on any graph to see a full-size version.

Update February 2022 - The data acquisition system with the Palm Pilot was pretty innovative and wokred fine at the time.  Now, 21 years later, I find myself taking similar performance data with a modern Raspberry Pi and Windows 10 laptop and smart sensor.

Wood Stove Temperature GraphIn order to understand the legend on this graph, you need to understand how my home office is set up.  It has two stories, with an open stairway between them.  The wood stove (Jøtul Model 118) is downstairs.  I use a box fan to circulate the air from around the stove into the full volume of the room.  The second floor is heated only by convection up the stairwell, or conduction through the floor.

Although you can choose to build your own wood stove, the Jøtul Model 118 is excellent for log heating, and has a reflux air path that forces full combustion to the back of the chamber.   I push coals back as I put in new logs, and then toward the end of the day, I pull the coals forward for the last few hours of heat.  Or, if I want auto-light feature the next morning, I leave the coals stacked, and pull them forward the next morning, toss on a few logs, and it picks up where it left off the day prior.

I monitored the temperature on the ceiling of the downstairs, on the floor of the upstairs, and at sitting height upstairs.  You can see from the upper trace the six times I added wood or opened the damper.

The second graph shows the propane furnace heating my house.  First off, notice the disparity between the setpoint (red line) and the three measured temperatures.  I'm not sure the sensors were accurate, and it's best to consider relative movement rather than actual temperature numbers.

One temperature sensor was stuffed into the forced air vent, and two others monitored the room temperature on the floor, and about 2' above the floor.

I turned the thermostat up, hooked up the monitoring equipment, then turned the thermostat down for the night.  The furnace came on about 13 times during the six hours of uninterrupted night equilibrium.  Notice when I turned up the thermostat at about 6:45am (red line bumps up) the burn time of the furnace was much longer (width of blue hump).  By measuring the width of each blue peak to determine "on-time" of the furnace, I could calculate how many BTUs I was putting into the house, and back calculate the effective R value of the surface area of my house.  This calculation was actually done for the data coming up next.

Effect of 5 degree thermostatThis graph is similar to the previous one,  but shows a more dramatic effect of turning down the thermostat at night.  This time, I let the heating system reach equilibrium before lowering the thermostat.  The result of a 5 degree lowering of the setpoint was more than 4 hours of no furnace activity.  That's roughly half my night time heating cost, or 1/4 to 1/6 of the total heating cost for the day.  16 to 25% savings for 5 degrees overnight!  Think about this as you read the numbers from my page giving all the numerical comparisons of fuel sources.  You can make a HUGE dent in heating costs by simply changing how you use your present heating system.  Be sure to consider your assumptions of comfort and convenience while you pursue less expensive fuel.

I hoped to take the longer temperature decay time of this graph an use it to graphically determine the Whole-wall R-value of my house.  R is the inverse of U, which has units of BTU / (sqft * hr * degF).  In other words, U expresses how much energy passes through a square foot of the house in an hour for every degree Fahrenheit difference in temperature.  In order to estimate these values for my house, I can see how much energy is spent to keep a constant temperature in the house. 

After the large temperature decay, the furnace starts cycling on and off.  It comes on for about 4 minutes every 40 minutes or so.  Adding up the peaks indicates it spends 36 minutes of time kicking out 74,000 BTU/hr during 3hr:52min of monitored activity.  That's 44,400 BTU over the course of 3.9 hours.  Toss in the fact that my house has 4360 square feet of surface area (walls, ceiling, and floor), and the 29 degree temperature spread, and you get:

R = 4360 * 3.9 * 29 / 44400 = 11

Normally, what you get quoted in the store as you plunk down the big dollars for rolls of pink stuff is the Center-of-Cavity R-value.  What you experience in a real house is the Whole-wall R-value, which allows for door drafts, window sill conduction, and other non-ideal behavior.  I think R-11 isn't that bad, but I'm interested in results from others who repeat the same experiment.

This page is maintained by Brian Mork, owner & operator of IncreaTMSM // It was last modified February 2022. Suggestions for changes and comments are always welcome. The easiest way is to contact me via e-mail.