[Piwoslaw]

I've finished this test and summed up the results from 2011 and 2012, a total of 12 weeks.

Date	Hours	Wh used	Timer	Wh/24h	Fridge temp min/max °C	Kitchen temp min/max °C
24.02-03.03.2011	166.8	3758	off	540.7	?/?	?/?
03.03-10.03.2011	168.0	3700	on	528.6	?/?	?/?
10.03-17.03.2011	167.0	3864	off	555.3	3.9/5.9	16.7/18.7
17.03-24.03.2011	170.2	4204	on	592.8	3.5/7.1	16.5/19.8
02.02-16.02.2012	334.2	8097	on	581.5	2.6/6.7	17.2/20.9
16.02-01.03.2012	333.5	8429	off	606.6	2.3/4.4	16.1/19.4
01.03-15.03.2012	337.9	8024	on	569.9	2.3/6.4	16.0/20.4
15.03-29.03.2012	331.1	8285	off	600.5	2.1/6.5	16.0/18.8

Comments:

1. The last week (17.03-24.03) of the 2011 test (timer on) saw lots of cooking, so the kitchen was slightly warmer and the refrigerator had more use. This noticeably increased energy consumption.
2. The last two weeks (15.03-29.03) of the 2012 test (timer off) saw lower average kitchen temperatures. The weather was warm enough to keep the central heating from turning on more often than once every 1-2 days, so the house's average temperature was about 0.5°C lower than during the previous weeks. Notice how the kitchen's minimum temps get lower toward the end of the test.
3. The timer itself uses a continuous 1W. It was removed during the 2011 "timer-off" tests, but left on during all 2012 tests.
4. The three situations above skew the results in favor of the "timer-off" tests.
5. The fridge's temps most likely maxed out when the door was open, not towards the end of the 6 hour off time when the timer was on. The thermometer's sensor was in the back, between water bottles, but when the door was open for a longer time, then warmer air had time to get deeper in.
6. During the "timer-on" tests in 2012 I noticed that around 6:30am the refrigerator's compressor was not yet on, even though the timer turned the power back on around 6:00. The temperature inside the fridge was ~5.8°C. On one hand, this would suggest that at night it takes the fridge more than 6 hours to gain enough heat to turn on. On the other, during the day, or during the "timer-off" tests, I never noticed an inside temp as high as 5.8°C - the compressor usually turns on just below 5°C. This leads me to believe that our refrigerator has some sort of safety feature that keeps it from turning on within 30-60 minutes after power is restored.
7. I noticed that the refrigerator's min/max were ~1.5°C lower in 2012 than in 2011. Maybe someone changed its settings?
8. Comparing results between 2011 and 2012 showed an increase in daily power usage: It went from 572.0 to 603.6 Wh (+5.5%) with the timer off, and from 560.9 to 575.7 Wh (+2.6%) with the timer on. I was expecting a decrease from the extra insulation I added to back of the fridge, but I guess this was offset by the lowered temperature.

Summing up all of the raw results:

• Timer on (fridge off from midnight to 6am): Total of 24025 Wh consumed over 1010.3 hours gives 570.7 Wh/24h.
• Timer off (fridge on 24h per day): 24336 Wh consumed over 998.4 hours gives 585.0 Wh/24h (+2.5% more than timer on).

If the 2011 "timer-off" results are increased by the amount of power that the timer would have used (see comment #3), then we get:

• Timer on: 24025 Wh, 1010.3 h, 570.7 Wh/24.
• Timer off: 24670 Wh, 998.4 h, 593.0 Wh/24h (+3.9%).

If all of the tests which used the timer are reduced by the amount of power that it would have used, then we get:

• Timer on: 23015 Wh, 1010.3 h, 546.7 Wh/24.
• Timer off: 23672 Wh, 998.4 h, 569.0 Wh/24h (+4.1%).