Here is our journey to upgrading our 40 gallon gas fired water heater to a 19 gallon electric one in July 2007.
We've put our 19 gallon water heater onto a timer (electric, 7-day programmable - 2W power draw) and set it to only turn on for 1 hour before the price rise in the morning. This provides mostly enough heat (only a bath requires it to be turned on otherwise). If necessary it can be boosted by turning it on for 30 minutes during a mid-peak time period.
August, 2021 - AO Smith 19 Gallon Water Heater
Standby losses seem to be 0.5kWh/day when set at 140F
August, 2021 Update - Tank Failure - rustthru
The water heater started leaking slowly and we considered electric on-demand models that can do about 2 gal/min flow but it seems a non-electrician can't buy 6ga wire needed for a 9kW model so we purchased an AO Smith19 gallon tank. No anode was accessable on the failed tank.
April, 2019 Update - Westek Digital Timer Failure
The timer failed again in the exact same place. First failure ~4 years after purchase and then 4 years after that.
This time I bent the relay pin down on the pad and put a 14 ga copper wire ring around the whole thing.
June, 2015 Update - Westek Digital Timer Failure
The Westek hardwired Digital Timer TMDW10CAN failed. Everything looked fine but the water heater was clearly not
being turned on and the red light on the on/off switch would not turn on and one could not hear the relay
click on or off.
Disassembling the timer was straight forward if a bit slow. The solder joint for the relay coil had clearly failed - a ring about 1mm from the relay pin was a clear sign of the solder joint failure. I was surprised to see this on the low voltage / power side of the relay. Re-soldering the joint fixed the problem.
March, 2013 Update
We are still happily using the system. When relatives visit we just leave the water heater on all of the time. Using a 240V timer and 240V heating element would allow the water heater to recover/heat up twice as fast but that's not a priority for us.
Somewhere in 2011 the original plug-in 15A timer failed. The socket into which the water heater power cord plugged started to get very hot, blacken the plastic case of the timer, and then failed. I replaced the timer with a wired in one from RONA - also rated at 15A - but now the connection from the timer to the heater element has much better electrical contact and we've had no issues. It's just not easy to measure the power draw now.
March 1, 2010 Water Heater Timer Energy Summary
Here is the summary of having our water heater on, and off, a timer. Note that a steady
2W power draw is from the timer.
Estimated daily losses (standby 0.6kWh, shower 0.2 and 0.6 kWh, dishes 0.15 kWh, laundry 0.3 kWh, bath 1.2 kWh).
30 days on a timer 1.19kWh/day average actual use, 1.29 kWh/day calculated average (-7.4%)
12 days w/o timer 1.45kWh/day average actual use, 1.16 kWh/day calculated average (+12.4%)
Summary: The timer saves us 20% of our energy use. Also - with the timer 30 to 50% of the losses were standby,
indicating a need for better insulation.
Feb 2010 Water Heater Electricity Usage (using a timer so it turns on 5:30 to 7am):
average: 1.2 kWh/day
0.65 to 3.5 kWH/day (0.65 is standby loss, 3.5kWh involves bath, 2 showers, dishes, laundry load)
Jan 2010 Using a 15A 120V electronic timer (2W power draw) we've put the water heater
onto a budget - it only turns on between 5:30 and 7am. As long as we don't use the bath
we have enough hot water for the day.
Estimated Loss delta T = 49-18=31C or 110-61=50F, water heater is aprox 20" dia and
28" high = 28 sq-ft of area. 1W = 3.41 BTU/h, loss (BTU/h) = delta T * area / insulation.
R = R8 "green foam" + R4 (fiberglas jacket) = R12
Estimated Heat loss = 117 BTU/h = 34W
Measured loss is 27W average but the temperature starts at 49C and drops over 23 hours so
one would expect the measure loss to be less than the expected.
Dec 2008 we changed the water heater thermostat from 49C to 52C as it seems to have gotten a bit cooler. Over the next 46 days (with relatives visiting for xmas and again) our average use was 1.96kWh/day or 83W avergae draw at 52C. 20W is the standby draw at 49C; for a total of 103W average draw.
Nov 16, 2009 - Water heater elec use was 9.31kWh over 6 days or 1.33kWh/day with typical use.
Nov 23, 2009 - Water heater elec use was 16.72kWh over 13 days or 1.26kWh/day with typical use.
Our original water heater was a 15 year old 1991 GSW 40 gallon natural gas model. It used $72/yr for the pilot light and an estimated $12/year heating water and the water inside was aprox 46C.
I have metered a 9 gallon electric GSW water heater and determined that it takes 25W ($23/yr) to keep the tank warm. The heater turns on 2% of the time to keep the tank at 50C. Draining water from the tank 2L at a time revealed that 40L of hot water could be removed from the 44L tank before the temperature started to fall.
Given what I knew of our water use (we never used more than about 20L of hot water at a time) we decided that a sub 20 gallon water heater can meet our needs. In the end we choose a 19L electric hot water heater because it was just big enough to provide hot water for taking a bath. If we didn't care about taking a bath the 9 gallon unit would have easily met our needs (except that they are less well insulated by default).
Metering the pilot on our water heater over a few minutes. The draw was 18cf/day ($0.40/m^3 or $1.10/ccf) $72/yr
Estimate water heating 3cf/day (based upon an average over 3 summer months), 6.25 ccf/mo or $12/yr
Water softener is manually regenerated every 8 weeks and it uses (mfg spec) 51 gallons.
Here is our water use and estimated hot water use (assume 50% of water used is hot unless stated otherwise):
1.2m^3/month toilet (4L / flush * 10 flushes/day of cold water)
0.4m^3/month showers (6L daily, 20L one every 3 days)
0.3m^3/month washing dishes (est 10L/day)
0.3m^3/month hand washing (1/2L * 5 washes/day/person of cold water)
1.0m^3/month clothes washing (110L water (20L hot water), 2 loads/week)
0.25m^3/month baths (50L / bath, once per week)
0.15m^3/month cooking (5L / day cold water)
That totals to 3.9m^3/mo while we use 4m^3/mo.
It equates to around 1.1 m^3 or 10 US gal of warm water per day. That is about 7 US gal/day of "hot" water from the heater. Note that a hot shower is about 39C, while a cool one is 35C. With the water heater at 49C and intake water around 15C that means that we typically mix 70% hot water with 30% cold to get 39C. A typical American family uses 40 US gal of hot water per day.
Electric tanks I can easily install. They work well with time day elec. pricing. Different elements 120V, 240V are readily available to allow for metering and putting the heater on a timer.
Gas fired tanks require a licensed gas fitter. They can provide lots of hot water; but there are no small tanks (smallest is 25 gal). Most of the heat losses are up the flue so there is no ability to insulate them better.
Heat Traps These are ways in which to stop hot water circulating within the pipe (convection currents) from sucking heat away from the tank. One way is via an s-bend in the pipe and another is via small balls which shut off the water flow unless there is demand. The ball method, however, makes a clicking sound when you start and stop drawing water and when I talked to GSW in 2006 they commented that they were removing these due to customer complaints about noise.
The Building Code in our area now requires a blending valve to ensure that the hot water temperature can not exceed 49C. However, in the mouse print of such devices you'll find that the inlet temperature to such a device must be 10C higher than the blended water temperature.
As we have for 15 years maintained our water heater at around 46C - having to raise the tank to 60C in order for the anti-scald valve to work is innane. This would increase standby losses by 25% (25% increase in temperature above ambient)!
I noted that the Building Code did not specificy that a blending valve was necessary; just that the max. water temperature must be 49C. In a new building here on campus (built late 2006) a water heater was installed without a blending valve (the water heater was set to 49C) and so we adopted the same "solution" for our home.
Estimated and actual time to heat a 65L tank: 1.5h at 1400W for a 30C rise (19C -> 49C; so who needs a 3kW element?)
Heating costs (electricity is 1.5x as expensive as gas)
During summer 2007 natural gas is $0.413 cu-meter or $1.17/ccf and heater is 65% efficient. Elec is $0.09/kWh and
elec. heaters are 93% efficient (Energy Factor) and 1 BTU = 252 cal, 1W = 3.4 BTU/h, gas holds
1,000 BTU/cu-ft, water rise is +40C (14C ground temp to 54C in the tank). Summer city water is
about 19C, estimated more like 14C in the winter.
So the calculated costs to heat the 150L in a standard 40 gallon tank by 40C is:
Natural Gas 24cu-ft @ 65% eff = 37 cu-ft = $0.43
Electricity 7.0kW x 1 hour = $0.70 or 1.6x as expensive
Reality check a kettle heats 7c (1.75L) of water in 6.2min +80C rise takes 0.14kWh. So 150L at 40C rise would be 6 kWh or $0.60 in aprox 4 hours.
I installed the basic tank and used aprox 3" of copper fittings on the tank; to which PEX connectors and pipe were connected. The copper and first few feet of PEX were covered with foam, slip-on, insulators.
Adding about 8 layers of packing foam (total thickness about 2") raised the surface temperature of the tank about 3C. Since the water inside is 30C above average this leads me to believe that an insulating jacket would reduce standby losses by 10%.
August 20 - a 1.5" (nominal R4.5) thick metal backed fiberglass insulating jacket was added to the heater. Internal temperature rose to 31C (ambiend 22.0C, 30.9C inside jacket, 49C water temperature) - indicating that it's adding 50% of the insulation of the 1" foam jacket (nominal R8) on the water heater. The standby losses reduced from 30W to 22W. (0.520 kWh/day or 21.7W average over 3 days 18 hours) It cost $15 and will save $7.5/yr.
Given: ambient temperature of 21.5C (July/August 2007), tank set to 49C.
From nighttime, no water being used, measurements the heater ran 7:36, 7:18, 7:54, 15:08, 13:44,
13:42, 14:14, 13:24, 7:08.
That is an average of 7:15 (0.16kWh) each time it turned on.
Over 2 days 21 hours while no water was ??????????
being used it ran 2:53:08 and used 3.93kWh. That's an average draw of 1.362kWh when it's running.
It uses 1447W max when turning on. 1.36kWh average means that it cycled on 8 times; every 8h 38min
on average. It takes 90 minutes to heat the tank when filled with cold water so running for 7:15
is 7% average on-time every 8h 38min. Ie it looses 7% of it's total heat aprox. every 8.5 hours. If the heater
were only turned on at night the heat loss during the day would only be 16% or 5C; quite acceptable.
NOTE: Given the above standby loss measurement 0.17kWh every 8h 38min is 0.48kWh/day or 20W average.
Over 2 days 21 hours 2.07kWh were used to keep the tank hot. To about 10% accuracy; this is
30W average or 0.72kWh/day or 4.24 power cycles per day or it turning on for 7:15 min every 5.7 hours.
NOTE: This doesn't agree with the above figure of 20W; but the measurements are done 2 different ways.
It, roughly, runs for about 7 minutes every 8.5 hours; so accurately measuring the standby loss is hard. If I hooked a buzzer to it so that it buzzed every time it turned on and I timed its heating cycles I'd be more accurate.
Turning the heater on only ever few days is fine for about 1/2 of the tank size. So for this 65L tank I can draw about 30L of hot/warm water over 3 days before having to heat the tank again (ie 3 showers and dish washing 3 times).
Average Elec Use: July 23 to Aug 5: 13 days 4h 20.21kWh is 1.53kWh/day; but 2 days of that was
vacation so correct that is 11days 4h 19.21kWh = 1.72kWh/day.
July 23 to Aug 9: 17 days 15h 25.2kWh is 1.43kWh/day
The heater is a GSW Spacesaver 12 US gal, 1500W element, side taps, model SS12SEB, fiberglas 1" insulation (R3.5), set to 120F or 49C, housed in a closed kitchen cabinet. The room ambient temperature is 23.1C and the cabinet housing the water heater is 33.1C!
Standby and Loss During Use:
11:30am Dec 13, power switched over, power meter at 0kWh
1:15pm Dec 13, 3:30h at 11.23A 0.07kWh (est 1kWh/day)
2:50pm Dec 13, 6:54 0.14kWh (est 1.06kWh/day)
8:52am Dec 14, 37:04h at 11.38A max, 0.80kWh est 0.90 kWh/day
10:15am Dec 14, 37:04h 0.80kWh
11:45am Dec 14, 44:08h 0.95kWh, est 0.88kWh/day
2:55pm Dec 14, 1:05:48h 1.42kWh
9:00am Dec 15, 1:30:48h 1.96kWh, (mostly) overnight use was 0.76kWh/day
2:52pm Dec 15, 1:48:16h 2.34kWh
4:17pm Dec 15, 1:48:16h 2.34kWh
9:02am Dec 18, 3:05:12h 4.01kWh (0.62kWh/day, 26W average draw, weekend - not used?)
1:57pm Dec 18, 3:16:08h 4.24kWh
Bailing water out of the tank 2L at a time and measuring temperature revealed
that the peak temperature was 47C and that 40L of water was extracted above 44C.
After 42L were removed water temperature droped quickly and was 32C at 60L.
Supply water was at 19C (warm compared to our home) and the element was on during
the 10 minutes of the test.
Here is a web page to calculate GHGs for electricity and it works out to 0.94kg/kWh in Ontario. Gasoline is roughly 2.4kg/L CO2 and natural gas 1.9 kg/cubic-meter CO2
Car We drive aprox 8,000km/year at 8L/100km or 640L of fuel (aprox $640 at 2007 prices). That is 1,540 kg/year!
Electric Water Heater 0.5kWh/day to keep tank warm, aprox 1kWh/day heating water. That is 536kWh/yr or 504 kg/yr or 5.1 kBtu/day (kWh * 3.412 = kBtu)
Nat. Gas Water Heater $72/year (18cf/day to keep tank warm, $12/year (3cf/day) heating water.
That is 77 ccf/yr or 211 m^3/yr or 400kg CO2/yr or 21.3 kBtbu/day (ccf * 101.3 = kBtu)
NOTE: CO2 emissions are done assuming that there are no losses in
natural gas transmission from Alberta to Ontario - in reality they are quite
significent and so the CO2 emissions are definately worse for the gas fired
water heater.
The anode should be checked every 3 years. After 15 years it was junk on our old tank.
Apparently the 6/9/12 year warranty on gas tanks is just a warranty - there is no difference between the tanks.
For our family of four.
| Pollution Source | Consumption | CO2 (kg/yr) | |
|---|---|---|---|
| Pre 2008 | Post 2008 | ||
| Diet | Whole-Food Vegetarian/Vegan | ||
| Car (1997 Ford Escort gasoline automatic) | 6,000mi/yr at 34mpg | 1,900 kg/yr | |
| Car (2005 Toyota Echo std) | 6,000mi/yr at 44mpg | 1,500 kg/yr | |
| Home heating (low eff. nat gas) | 875 m^3/yr | 1,660 kg/yr | |
| Home heating (high eff. nat gas) | 590 m^3/yr | 1120 kg/yr | |
| Electricity (BullFrog Green) | 2200kWh/yr | 1,700 kg/yr | |
| Water heating (40 gal. nat gas) | 211 m^3/yr | 420 kg/yr | |
| Water heating (19 gal Elec) | 550kWh/yr | 500 kg/yr | |
| Total | 5680 kg/yr | 4820 kg/yr | |
The Passive House standard sets a limit on heating energy as well as total energy use. An investigation reveals that they expect a very large electrical/cooking energy use to basically heat the house. This prevents the need for a furnace.
Great things about the Passive House are attention to details (thermal bridging) and earth tubes to pre-heat fresh air as well as triple glazed (Heat Mirror) windows and passive solar orientation. They are lacking thermal mass to really take advantage of passive solar heating though.
Comparing our 1991 home to the Passive House standard:
Passive House allows 120kWh/m^2/yr total energy use Our home is 150m^2 Electric water heater is 100W average draw or 875 kWh/yr We used to have a gas fired water heater which was 211 cubic-meters of gas or about 2,300 kWh/yr Our home uses 875 m^3 of gas (high efficiency furnace) for heating which is: 875 m^3/yr / 2.74 = ccf/yr, * 101.3 = kBTU/yr, / 3.412 = 9,500 kWh/yr So our total home energy use is 12MWh/yr while the Passive House standard allows 18MWh/yr and we have 1/2 the insulation (walls, windows, ceiling) of a Passive House and no slab insulation at all.
To put that into perspective our existing gas furnace is 40,000 BTU/hr and runs 50% of the time on a windy -30C day - which is about 5.5kW continuous draw or about 4 hair driers. For the typical winter season leaving 3 hair driers running would allow our home to "heat itself without a furnace" but heating costs would be about 2x higher than they currently are ($950 vs $400).
To grid intertie a PV system requires a $3,000 inverter; not the regular $500 one.
Standard Offer contract requires >2kW system.
Wind turbine would require a $1,000 tower which we don't have space to install.
The only PV option would be to partition our house and run part of it on a PV/battery
system which can be charged from the hydro grid as necessary (automatically?)
Our water heating is 2kWh/day or 2.6GJ/yr
Our home heating peaks at 6GJ/mo (typical, not extream).
By windows (passive solar). Assume 6h light/day, 1kw/m^2 @ 30% eff
Aprox 57kWh/day or 32m^2 of windows
Using Thermo tube system we need aprox 500 tubes or 50m^2 of collectors.
Very roughly estimated at $50k
An 80 US gal (60 Imperal) water tank heated to 100C stores aprox
70C * 300L * 1000 cc/L * 1J/C/cc = 0.021 GJ
So a large water tank holds enough heat for aprox 1/10 of a day!!
That seems wierd but it fits with what I've seen:
http://me.queensu.ca/people/harrison/research/solar/Images/Cruickshank_Harrison_CanadianSolarBuildingsNetwork_Sesci_2006.pdf
Now we know why homes use thermal mass and not hot water tanks. Mind you
I've heard of open 1,000 gal tanks for heat storage!
The Ottawa group uses two 120 US Gal tanks. One tank was not
enough for their system. I'll get a copy of the CMHC report.
Ballparking the window size - a south-west facing home would be aprox
15m x 5m = 75m^2 so the window area required is 40% of the wall space
and that's on the large size from what I've read; about 2x too much.
Now coming at it from another angle.
50m^2 of space used instead for a PV array is roughly
10% efficiency, 1kW/m^2 => 5kW and aprox $50,000
So the solar hot water and solar PV costs per unit area are roughly similar.
The solar water heating is overkill most of the year while PV is useful year
round.
That's why you should exploit passive solar instead!!
NOTE: ground source heatpump would be about $25k for our home.