Do New Baseboard Thermostats Save Energy?

Higher quality electronic line voltage thermostats (ELVTs) have been available for several years. Most models indicate temperature in degrees, and they're touted as energy savers. But how well do they work? Do savings warrant replacing old thermostats? In Oregon, the Portland Energy Office (PEO) asked these questions. The result was a test of two no-frills ELVTs--Cadet TDLV-240H and Honeywell T4800--as retrofit baseboard thermostats.

A good thermostat should keep room temperature within a degree of the setting. Bimetallic central heating thermostats can do this at 24 volts, but they carry only enough current to operate a relay. Older baseboard thermostats also use a bimetallic element. It senses temperature and switches heater current--two conflicting jobs. The bimetal strip must respond rapidly to temperature changes, and be robust enough to switch currents that sometimes exceed 20 amps without overheating. Mishandled, this compromise can cause excessive hysteresis--temperature swings above and below the setpoint. Droop--lowered setpoint as the heater nears full-time operation--can also be a problem.

ELVT development was sponsored by the Electric Power Research Institute (EPRI). ELVTs avoid the compromise inherent in bimetal thermostats by separating sensing and switching. A lightweight thermistor senses temperature; semiconductors control heater current. Laboratory tests show low hysteresis and droop.

In theory, ELVTs save energy by maintaining comfort at a lower average temperature. Figure 1 shows how two thermostats maintain the same minimum temperature with different setpoints. The higher hysteresis of the bimetal thermostat forces a higher setpoint to get the same minimum temperature. The assumption is that people set thermostats to keep temperatures above their minimum comfort level. If this is true, ELVTs will lower average temperatures by half the difference between the two thermostats' hysteresis.

However, unless people really do set their thermostats to maintain a minimum temperature, savings may not occur. Also, Figure 1 oversimplifies thermostat use. Real-world effects include night setback, heat turn-down during absences, and many other behavioral factors.

Previous field tests have assumed constant thermostat settings for evaluating ELVTs. But PEO needed to be convinced that savings were a realistic expectation, without assuming "set and forget" operation. Through winter-long testing, we compared ELVTs to original-equipment bimetal thermostats in 27 two-bedroom apartments.

We controlled for random effects of occupant behavior and weather. Both old and new thermostats were used by the same occupants, each for two months of the winter. Each family was its own control group, to ensure that measured differences were thermostat related. To avoid self-fulfilling expectations, occupants weren't told that the ELVTs were energy savers.

Figure 1. Temperature variations (hysteresis, set points, and minimum temperature) with bimetallic and electronic line voltage thermostats, assuming a 6°F temperature swing with the bimetals. We tested thermostat usage to determine if, as shown here, residents would set the bimetal thermostats up to 2.5°F higher than ELVTs to achieve the same minimum temperature.

At the beginning of the experiment, about half of the bimetal thermostats were replaced. Two months later, the remaining old thermostats were replaced with ELVTs, and the apartments with ELVTs were changed back to their original thermostats. A constant half-and-half mix of thermostat types prevented weather bias.

Data loggers monitored whole-apartment power, baseboard zonal power, and zonal temperatures near thermostats. We also made unique measurements, crucial to the test, of zonal setpoint and hysteresis. The Data-Trap loggers we used can record event-specific or condition-specific data. We programmed them to record instantaneous temperatures when the baseboards turned on and off.

Results were a mixed bag. ELVTs provided excellent temperature control accuracy compared to the bimetals. After the test, several families asked for reinstallation of ELVTs that had been replaced by original thermostats. However, the original thermostats weren't uniformly poor.

ELVT manufacturers' literature suggests that typical bimetals had hysteresis of about 6°F, as shown in Figure 1. But hysteresis actually averaged only 2.5°F, reducing savings expectations by 65%. One of the three apartment complexes--19 of 27 units--had reasonably good bimetal thermostats, averaging 1.4°F hysteresis. Our test protocol dictated that they should be retrofitted; otherwise we wouldn't have bothered.

ELVT hysteresis averaged 0.6°F, with minor variations. Most occupants favored ELVTs. As theory predicted, setpoints were lower with the ELVTs, by about 0.9°F. If this had been the only change due to ELVTs, annual savings would have been about 220 kWh per apartment-year ($14 per year at 6.5¢/kWh).

But the energy performance results were initially baffling. To our surprise, average inside temperature (and therefore energy use) didn't decrease with ELVTs. How, if setpoints decreased and everything else stayed the same, could average temperature not decrease?

Our data showed that most occupants used some night setback. Daily highs and lows differed, on average, by about 2.2°F, ranging from 0.4°F to 6.1°F. These occupants just weren't "set and forget" types! Further checking showed that the use of night setback changed when thermostats were switched. With ELVTs, occupants didn't turn temperatures down as much overnight as they did with bimetal thermostats. The result: no statistically demonstrable energy savings. But this wasn't the whole story.

Instead of energy savings with ELVTs, we found a significant reduction in peak demand. This was partly because the heaters didn't run as long to achieve morning warm-up, due to the reduced setback with ELVTs.

To determine how setback affected savings, we split our sample of bimetal users into a low-setback half and a high-setback half. The high-setback group showed greater demand reductions, but no energy savings, with ELVTs. The low-setback group showed smaller demand reductions, and some evidence of energy savings, with ELVTs.

Unanswered behavioral questions remain. For example, some occupants turned the heat off--not down--at night with bimetals. Faced with obscure choices ("Do I turn it down to the c or the m in 'comfort zone'?"), did occupants simply turn off the heat? With ELVTs, and real temperature settings to choose for setback, did they choose logical, but higher, setback temperatures? Or were bimetals turned off to avoid overwarm sleeping temperatures?

For people who are considering retrofitting baseboard thermostats, here's what we recommend:

• For improved comfort, install ELVTs; they control far better than the bimetals.
• For energy savings, first check the bimetal thermostat. A good-quality bimetal will perform nearly as well as an ELVT. Replacement may not be warranted. During mild weather, if room temperatures fluctuate only a degree or two during operation, or if the heat runs for only 10 to 20 minutes before turning off, the bimetals aren't bad. Individual bimetal models performed consistently; spot checks may suffice.
• Retrofit energy savings are fairly certain with poor bimetals when occupants are unaccustomed to setback.
• If occupants already set back their thermostat, there are two options. For energy savings, occupants must use as much setback with ELVTs as they used previously. (They may need coaching.) For occupants with unpredictable habits, expect reduced peak morning demand, but don't promise energy savings.

Home Energy can be reached at: contact@homeenergy.org
Home Energy magazine -- Please read our Copyright Notice