I got the temperature dialed in close enough today and started my 24 hour baseline. Anecdotally it appears the interior temperature setting and ambient temperature are both sensitive factors for power consumption (not really surprising). I decided to track the variables on an hourly basis (except for overnight; no automatic logging facilities).
Well... best laid plans of mice and men, and all that. It took me longer than I wanted to get my refrigerator measurements to compare the standard analog controller and the SEC (digital) controller, and circumstances intervened and made my data collection not nearly as scientific as I envisioned.
The net result is that I observed the Isotherm SEC installed in my Isotherm Cruise Elegance 130 refrigerator to reduce battery energy consumption somewhere between 22% and 29% under testbed conditions. The reduction came while running on battery-only power levels in a 24-hour period with 8 hours charging power (supply voltage above 13.2 VDC) and 16 hours battery-only power (supply voltage below 13.2 VDC) [actual reduction experienced depends on how much inefficiency was present in the 120VAC to 12VDC power supply used during the test].
My rough measurements also suggest that power consumption during long-term operation on battery-only power may be slightly more with the SEC (digital controller) than the standard analog controller, but my testing did not have the necessary rigor to conclusively show this result.
The results are not rigorous because energy consumption with both controllers was very sensitive to the set point, ambient temperatures, and the thermal mass of the refrigerator contents. I maintained a consistent set point fairly well, maintained the same thermal mass (but with a slightly different distribution), but could not control the ambient temperature well between and during test runs. Unfortunately I was not able to include ambient temperature data during the collection of baseline data with the standard analog controller.
A higher thermal mass in the refrigerator would probably increase the battery energy savings as it would take longer for the internal temperature to rise during battery-only operation. A shorter period of "battery-only" operation (versus "charging" operation) would certainly increase the energy savings; it took 10 hours for the refrigerator to "coast" up to the 41°F set point, during which very little energy was consumed at all.
Method
I loaded my Isotherm Cruise Elegance 130 refrigerator (BD35 compressor) with twenty 16.9 fl. oz. water bottles and eight 12 fl.oz. soda cans (Coca-Cola if it matters). I measured temperatures in a glass of water that I located at the forward edge of the bottom wire shelf of two (above the vegetable draw top) at approximately the height that the SEC temperature sensor is installed. Only the factory insulation was used. The refrigerator was free-standing (not installed in a cabinet), but was protected from direct sunlight.
I measured electricity consumption using a 120VAC Kill-a-watt monitor at the input of a 12 VDC power supply that was in turn wired to the refrigerator power leads. My measurements are probably slightly high due to overhead/inefficiency in the 120VAC to 12VDC power conversion; I saw a fairly constant 5 watt load on the Kill-a-watt even when the compressor wasn't running. I'm don't know how much of that is the power supply versus parasitic load in the refrigerator.
Baseline with standard analog controller
I found that the standard analog control maintained approximately 41°F at a control setting of "2". I ran the refrigerator for a total of 62 hours 51 minutes after stabilizing the temperature, with a total of 1.26 kilowatt-hours recorded on the Kill-a-watt. That's 20.05 watts average usage, or about 40 amp-hours/day assuming 12 VDC. Unfortunately I did not manage to record the ambient temperatures, but they were fluctuating between 69 and 70-something degrees. If the 5 watt draw is all attributed to the power supply, then it it would be a 30.1 amp-hours/day at a nominal 12VDC.
SEC (digital) controller installation
After capturing this baseline with the standard analog temperature controller, I installed the Smart Energy Controller (digital). It was slightly harder than the many-year-old Isotherm video shows, partially because my refrigerator is a different model than the one used in the video, and partially because the video cuts out certain parts. I had to pull the capillary tube for the analog temperature sensor out from behind the refrigerator liner in the ceiling, not just pull it off the wall. It might be re-installable, but only with difficulty.
I briefly assumed that the SEC unit would have the DIP switches set for the "standard" refrigerator configuration as shown in the manual; this was not the case. All of the DIP switches were set to the "off"/0 position out of the box. Fortunately I checked them before installation, and I set them to:
- [Bat] 1 (V < 10.8),
- [VLS] 0 (13.2 VDC),
- [R/F] 0 (Refrigerator),
- [T/deltaT] 10 (Fridge mode "2", with T 41°F, and deltaT 50°F), and
- [Offset] 100 (offset "4", +8.1F).
Except for the [Bat] setting, this is the "standard" refrigerator configuration shown in my SEC manual. I wanted the [Bat] setting higher to protect my battery more (this resulted in a problem after installation - see the end of this for the reason why).
If someone installs the SEC without adjusting the all-zero DIP switch settings, they will probably get somewhat unsatisfactory operation and less energy savings as the refrigerator would try to maintain a 34.7°F temperature on battery-only, and possibly as low as 31.1°F when VDC > VLS ("cold storage" mode).
The suggested [Offset] was far too high for my refrigerator/SEC combination. I had to run the unit for two days, making incremental adjustments to the offset, until I achieved the approximately correct set point at [Offset] 001 (offset "1", 0°F). At this offset, my temperature measurements in "cold storage" mode were 33.5°F (only slight off from the nominal 33.8°F), and in "normal" mode (battery-only operation) about 41.3°F. At two times while making adjustments the refrigerator didn't start up properly after shutting it down, changing the [Offset], and starting it up again. Both times simply shutting the refrigerator off again (via the control dial), waiting some 10s of seconds, then starting it back up again cleared it up.
Measurement for SEC (digital) controller
My load/measurement setup was the same as for the standard analog control. I varied the my power supply voltage from 12.02 VDC (simulated battery-only operation) to 13.42 VDC (simulated "charging" operation).
I started my data collection with the fridge in "cold storage" mode and temperature stablized at 33.5°F. In hindsight, I should have started it at the battery-only mode set point of 41.3°F in order to capture the energy usage while reducing the temperature to the "cold storage" set point. I ran the refrigerator for an 8 hour period as if it were in charging mode (13.42 VDC supply), then I changed the voltage to battery-only mode (12.02 VDC supply) for a 16 hour period. These were the results:
- First eight hours (VDC > VLS): this period average 33.7 watts (33.5*F temperature maintained)
- next ten hours (VDC < VLS): this period average 11.52 watts (temperature rose to 41.6*F)
- at 24 hour mark (VDC < VLS): overall average 15.64 watts (temperature was maintained at 41.2-41.4°F for the last 6 hours)
Results
Based on this 8-hour charge/16-hour battery duty cycle, then SEC reduced consumption during battery-only operation from 320.8 watt-hours (standard analog control) to 250.24 watt-hours (SEC control), a 22% reduction. It did this at an overall 25.38 watts (inclusive of charging period plus battery-only period) rate, which is higher than the 20.05 watt average for the standard analog control.
If the 120VAC to 12VDC power supply's inefficiency is responsible for the entire observed 5 watt consumption when the refrigerator was at idle, then the energy savings are more impressive, as the SEC would be reducing consumption during battery-only operation from 240.8 watt-hours to 170.24 watt-hours, a 29% reduction.
I then ran the refrigerator in battery-only mode (VDC < VLS) for a total of 52 hours (28 hours past the end of the first 24-hour period). I found that the average watts used after the initial warm-up to 41.6°F was 24.13 watts while maintaining the temperature at about 41°F (compared to only 20.05 watts average with the standard analog controller). This suggests that long-term operation on battery-only with the SEC may be less efficient than long-term operation on battery-only with the standard analog controller. However, poor control over the ambient temperature, differences in the way the refrigerator was reloaded (same content, but possibly arrange differently) may explain that difference - my data suggests that the energy usage is sensitive to ambient temperature, and may be somewhat sensitive to the loading.
Other findings
I installed the refrigerator in my van after completing all of my testing, the night before leaving for a 2-week trip. I was disappointed to find the refrigerator not working correctly. It would start up, run briefly, then shut down. After a brief pause, it would repeat the behavior. It did this continuously.
Troubleshooting revealed several issues:
- My branch circuit for the refrigerator is 12 AWG wire over an estimated 18-20 foot run (this circuit was not originally intended to run the refrigerator). The Blue Sea Systems circuit wizard suggests that a load of 2 amps (average consumption by the refrigerator) would have < 3% loss over 40 feet (round trip) with AWG 12 wire, but the refrigerator manual suggests that the maximum run for 13 AWG wire is 8 feet, and for 11 AWG cable is 13 feet (one way). Why?
- I observed the peak load when the compressor starts up to be at least 80 watts (6.6667 amps at 12 VDC), and the actual peak is probably higher (load observed on Kill-a-watt meter during refrigerator testing may not see actual peak).
- The Blue Sea Systems circuit wizard suggests that a load of 7 amps needs 10 AWG wire to have < 3% loss over 40 feet (round trip).
- The refrigerator's low-voltage cutoff ([Bat] setting) was set to 10.8 VDC, an increase from the "standard" setting of 9.6 VDC
Measuring circuit voltage while the refrigerator attempted to start up in the van was revealing. Before the refrigerator tried to start, it was at 13.28 VDC at the refrigerator's power leads. At peak load while starting, the voltage sagged below the 10.6 VDC low-voltage cutoff (more than 20% voltage drop). As soon as this happened, the refrigerator cut off the compressor, even though it was just in its startup phase. The fan ran briefly after the compressor, then stopped. The voltage returned to about 13.2 volts, and the cycle repeated.
I changed the [Bat] setting on the SEC to the standard 9.6 VDC, and hooked up an external battery charger to the house battery. The refrigerator started normally and ran correctly. It continued to operate correctly on battery power for my 2-week trip without the battery charger, just normal charging by the van's alternator. However, based on the observed peak draw during startup, I'm going to run my dedicated refrigerator circuit using 10 AWG wire to reduce the possibility that a low house battery SOC will render the refrigerator inoperative due to voltage sag at peak startup current draw. I'm also going to investigate an alternative wire routing that might reduce the total circuit length.