Water heater not working, trying to track down fault (FIXED)

[edy]

I have a compact electrical water heater and it has stopped heating. It wasn't a gradual stop. Either the element burnt out or there is something wrong with the thermostat control. I am not sure what best practice is to start. I'm pretty sure the main junction box breaker works. So are there any techniques to debugging the issue and finding the cause? Any help would be greatly appreciated.

I was thinking of checking continuity of element. Also the reset fuse perhaps checking continuity across it. Could be thermostat limiter. I am hoping the controller is at fault and so tank doesn't need replacement as that would need a plumber.

[Ian.M]

Its a simple series circuit: L - double pole thermal trip (red button) - thermostat - element - back through the thermal trip - N.
Odds are the thermostat has failed, as it takes a beating repeatedly switching to maintain temperature.  With the feed isolated, check its continuity - should be shorted when cold, as it opens when hot.

An element failure without a short to Earth that would trip the breaker is *extremely* unlikely, but you can check the element resistance and calculate it from the wattage by R=V²/P to see if its in the right ballpark.  Also check resistance to earth with the element isolated.

[Fisher77]

Kill the breaker to the water heater if you have not already.

You will have to ohm the element out. Should be between 10 to 16 ohms.
A 3500 watt element should be around 16 ohms. A 4500 watt between 12 and 13 ohms. A 5500 watt between 10 and 11 ohms.

[DTJ]

If the element is open circuit and you are capable of DIY then swap it out yourself.
I replaced mine at a cost of $40. Having a plumber do it would have cost 5 times as much.
Just be safe and turn off the circuit breaker and your mains switch when working on it.

[SeanB]

Does it click when you open the water flow, the most common fault is the differential pressure switch, as that operates the most, and can shift in trip point. Then the thermostat, then the element in that order can fail.

Power off, turn on the water and check from terminal to terminal for continuity, going from point to point along the chain. When you get the open there is the fault. This applies to the instant type, and as a first test try pressing the red reset button, and see if it works then.

[Ian.M]

Does it click when you open the water flow, the most common fault is the differential pressure switch, as that operates the most, and can shift in trip point.

The wiring diagram is for an immersion heater not a 'instant' demand heater, so no flow switch.

[Paul Moir]

Also it's a 120v unit identified by the white & black coming in so the heater will be 1500W ~ 13 ohms.  If it's one of those little ones you can't substitute a 240v element because they're too long and won't fit.
If it is the element expect it to be in there super-tight.  Try to find a socket to fit since getting a wrench in there is pretty tricky.

[Ian.M]

North America has the cheap-skate habit of using a glass enamel lined mild steel tank for hot water heaters (unlike the copper or stainless tanks used in Europe).  If you have to pull the element, and you don't use a socket with a T-bar driver, there is a risk of the side forces flexing the tank round the element port, which can crack the enamel and lead to rapid tank failure.  If you have to pull the element you should change the sacrificial anode at the same time unless less than 1/3 of it has wasted away.

[tpowell1830]

All of the above are good checks and actions.  The other question that I would have is what type of water that you have.  If you have high concentrations of calcium in your water, the area at the bottom of your tank will build up with calcium granules which will eventually surround the heating element causing a problem. If you find this buildup in your tank, you may want to get a processing system in line with your water heater that will mitigate this buildup.  After the element is surrounded with these solids, the system has a difficult time regulating the water temperature due to the insulating effect of all the granules around the heating element.

PEACE===>T

[Cubdriver]

First off - be careful measuring line voltage.  I'm taking for granted that you have a meter or test light capable of doing so, and that you yourself are comfortable with making such measurements.

That being said -

The first thing I'd do is put an AC voltmeter across the input terminals (1 and 3 on the high limit control on the diagram) to make sure you have power to the unit itself.  If that's good, I'd then check for voltage across the element terminals.  If there's line voltage there and it's cold, it's likely that the element is toast.

If there is low or no voltage across the element, start measuring across the contacts of the high limit and t-stat to determine which is open.  On the high limit switch, measure across terminals 1 and 2 (left side), then across 3 and 4 (right side).  In this case, you DON'T want to see voltage - these should be closed contacts with little or no drop.  If the high limit passes muster, move on to the thermostat, and measure between terminals 1 and 2.  Again, if things are good you should have little or no voltage drop if the contacts are closed.  This series of tests should quickly isolate the fault.

If the element was at some partial voltage rather than at full line voltage, you should find the 'missing' voltage across one of the sets of contacts being measured here (series circuit).  The set of supposedly closed contacts showing appreciable voltage is the faulty part.

While a bit more dangerous due to line voltage testing, doing it this way looks at the circuit under actual operating conditions and will show poor/partial contacts that fail under load that straight resistance checking might not indicate.

Just to reiterate, be cautious checking line voltage.

-Pat

[Ian.M]

First off - be careful measuring line voltage.  I'm taking for granted that you have a meter or test light capable of doing so, and that you yourself are comfortable with making such measurements.
...
While a bit more dangerous due to line voltage testing, doing it this way looks at the circuit under actual operating conditions and will show poor/partial contacts that fail under load that straight resistance checking might not indicate.

I'm not surprised no one else encouraged the O.P. to test it live.   Unless he's got a CAT III or better rated meter (not Chinese fake 'Cat III'), with fused test leads with screw-on clip ends, + the experience with electrical installations to use it safely, its not a good idea.  One slip and it could be 'game over'.  If he had the required experience, then why the heck would he be asking how to diagnose a simple series circuit like this?

However if its just a matter of confirming that power is reaching the heater, and a suitable meter isn't available, a low wattage mains incandescent light bulb in an insulated holder could be patched in (with the supply isolated while connecting it) across the mains in terminals (1,3) of the cutout ('High Limit Control').   Everything else can be tested with the circuit isolated, except for the possibility some part is going open circuit when hot - for which a clamp-on current meter on the incoming black wire + a test lamp across the element will be helpful.

[edy]

Thanks for the suggestions. I've had a go at testing continuity at least along the control module pictured (see attachment with labelled terminals). I've also tried the element and I'm getting wonky results which I don't understand.... I will explain below.

By the way, the control module is completely disconnected when I am measuring it. It is not in circuit. Same goes for the heater element (which is still screwed into the base of the tank).

Here are the results:

1. From the picture I posted, I get continuity along the left side from #1 --> #2 ---> #5 ---> #6.

#1 is the black input from the mains. It is continuous to #2 which indicates the RESET switch is working ok and not tripped. There is a metal contact physically connecting #2 to #5 so obviously they are really the same node and continuous. #5 is input to the thermostat control/limit, and #6 is the output. I have continuity between #5 and #6 as well, indicating the thermostat limiter is closed switch (since tank is cold) and has not failed either. So #1 --> #6 is all working.... and #6 connects up to the heater element.

2. From the picture posted, I get continuity along the right side from #3 --> #4.

#3 is the white input from the mains. It is continuous to #4 which also confirms the RESET switch is working. If the RESET was tripped, it wouldn't connect #1 to #2, and #3 to #4. Then #4 is connected to the other side of the heater element.

So far... all ok. The mains BLACK at #1 is getting through to #6 and to one side of the heater element, and the mains WHITE at #3 is getting through to #4 which is connected to the other side of the heater element. There must be electricity getting to the heater. I could connect a light bulb socket to #6 and #4 and it should light up. If I use say a 30W bulb, at 120 V, that will put 0.25 A across the RESET and thermostat (P=VA=120V x 0.25A=30W). I know the cold bulb will draw higher current, but it shouldn't trip the RESET and I would imagine quickly settle to 30W and give me a glow indicating I am getting power from the mains to the tank and through the control box. The only variable then would be the heater element.

Do you think that light bulb test is a good one to run next? I don't want to touch live mains. I can hook it all up and stand by the fuse box, trip the mains fuse and see if my bulb lights up from a safe distance away.

3. The heater element/coil - I'm a bit confused.

I hooked up my DMM on ohms across the terminals of the element. At first it was showing resistance in the few K, then started to rise, then it went high to many M, then it started falling again, never quite settling. As soon as I take off the DMM and try it again, it starts at a different number. It never gets to some stable number. At first I thought maybe there is some "inductive" effect happening or even capacitive, but I can't rationalize it because if it is just a simple element/coil it should reach a value fairly quickly.

Anyways, for now the heater element seems to be giving me very strange readings measuring ohms on my DMM. So I'm a bit confused. I am more confident that the RESET and thermostat limit part is working correctly, and that the coil is buggered up but that still doesn't explain how the ohms changes all over the place and creeps up and down seemingly randomly. I must be doing something wrong.

If the coil was broken, it would be infinite resistance, open circuit, not vary with time? 

[tpowell1830]

If the coil was broken, it would be infinite resistance, open circuit, not vary with time? 

Your heater element is broken.  If there is still water in the tank and possibly some crud, you will get strange readings.

PEACE===>T

[Johnny10]

Sounds like element is bad.
Should only read between 10-20 ohms.
No other reading.

I fixed mine two months ago.
They make a special tool to remove element. Only 4-6 dollars.

From The web.

Turning on the faucets for a relaxing shower only to find tepid or icy water makes for a frustrating experience. When an electric hot water heater’s elements short out or burn through, cool water is the result. Typically, the lower element goes first, but that’s not always the case. Fortunately, a few quick electrical tests reveal which element you must replace in order to restore hot water to your home.
1

Turn off the power to the electric hot water heater. Some units are wired to plug into a wall socket, and power is shut off simply by unplugging them. Most units are hard wired directly into the home’s circuitry, so turn the breaker for the hot water heater to “OFF” at the home’s main electrical panel.
2

Remove the two wires connected to the water heater element and push them aside. Unscrew the mounting hardware with a screwdriver, slip the wires out and bend them out of your way.
3

Set the multitester to measure ohms, or resistance.
4

Touch a probe on the multitester to each screw on the element. If you get no reading, or a maximum reading, the element is bad. Elements do have some resistance, so a reading of 10-16 ohms is normal, with higher ohm readings for 3,500 watt elements and lower readings for 5,500 watt elements. The wattage of your element is printed on the plastic block between the two screws where the wires were attached.
5

Touch one probe to a screw on the element and the other probe to bare metal on the water heater. Any ohm reading or slight movement of the multitester’s needle indicates an element that has shorted out. Check each screw on both elements.
6

Touch one probe to the element screw and the other to the metal frame of the element (but not to the other screw). Any needle movement or reading indicates an element that has shorted out. Check each screw on both elements in the same manner.

[salbayeng]

Hi ,
Your meter indications indicate the element is faulty , it has  a split in it , this allows water into the internal part of the element, the resistance element then corrodes and fails .

Water in the element gives you flaky DMM readings, typically the resistance will be low initially then rise to megohms, if you reverse the leads, it will go low(or negative) then rise to megohms.

The element can fail due to a calcium buildup in the bottom of the heater, if it is high enough to cover the element, then the element will overheat as it can't circulate the water, and melt the casing at one spot.

You need to flush out the calcium gunk when you go to change the element, there are YouTube videos on how to do this with an American tank.  The drain port is connected so it will siphon most of the gunk out. If water stops coming out of the drain port it might be clogged  (or air locked )( not actually empty) poke a wire down the drain and/or open the relief valve to avoid a surprise when you remove the element.

Getting the element out can be tricky, I think the US ones are mostly screw in, so you will need the correct sized socket, you also need to apply the torque in such a way as you don't damage the bottom of the tank, i.e. you need to use a T handle wrench.

You will also need to change/replace the sacrificial anode, (see youtube) somewhere on top of your heater there will be a hexagonal plug looking like it's plugging an unused port, it actually has a rod of zinc attached.

(edit).. Also the blown-open element will als0 generate copious amounts of gas (mainly hydrogen) , so the last time you used the heater you would have noticed a massive amount of "air" coming out the tap (faucet?)

[salbayeng]

North America has the cheap-skate habit of using a glass enamel lined mild steel tank for hot water heaters (unlike the copper or stainless tanks used in Europe).

We have glass lined steel tanks in Australia too (solar HWS get stainless though).
If you don't religiously change the anode every 5 years they will die around 7-10 yrs of age.
Usually indicated by a trickle of rusty water from around the thermostat area, if ignored turns into a fountain.

We get a lot of scale too and it gunks up the pressure regulator/check/relief valves.
(The old place had tank water, and the water heater was going strong there after 25yrs (and is probably still there! anode was ~ 2/3 long when checked at 25yrs)

If you forget to suck the gunk out of the bottom of the heater with anode changes, then it will blow out the heater element (words of experience there!)

[Paul Moir]

Home depot carried those little elements last time I needed one a few years ago.  Shoot, I was just there.

[edy]

I am tyring to drain the tank completely so I can inspect it better. It is a "water saver" tank, one of those tiny half-size ones, mounted up near the ceiling in the office.

First I turned off all the hot water supplies to all sinks/faucets... especially the "mixing type" single-handle type, since I notice pressure from the cold side will push water to the hot side. The only hot tap I left open was to a sink that has separate knobs for hot and cold water.

That way, I can close the cold water knob, open up the hot water fully to allow air in to the system so that it bleeds the negative pressure... otherwise it takes forever for the water to drain out of the tank.

I also noticed that if I open up the pressure relief valve at the side of the tank it will also let air bleed in and drain the tank faster. My only concern is if it is safe to use that for this purpose, or will it cause any damage by being open for so long while the tank drains?

Question 2 - I don't see any anode on the tank. Only a heater element. There is nothing else sticking itself into the tank. Am I missing something here or do some mini-tanks not have these? It os a GSW Water Heating, Model SS12SEB15 Cat No. A1229. 1500 W lower element. Capacity 43 L.

[Bud]

Water heater troubleshooting?  You need a 1GHz Keysight scope.

Sorry could not resist. Many of you know what I am talking about

[tautech]

Edy, you need confirm there is an incoming mains supply to terminals 1 and 3.

Don't know how Canada controls their HW heating but here it's sometimes a control circuit from your supplier so they can remotely control their network loadings. This will likely control a relay and we all know they can sometimes go bad.

OR

A Ripple switch control configuration which is wired in series with the live phase and is activated by your network supplier by means of overlaying a specific ripple on the supply to open or close said switch.
These can go bad too, sometimes they get out of phase figuratively, meaning when everybody else has HW heating with their ripple switch in the on mode yours might be in the off state. Seen it happen here but it's not common.

Either method and their mode/s can be diagnosed easily with just a DMM, some thought and a little know how. 

[DTJ]

Question 2 - I don't see any anode on the tank. Only a heater element. There is nothing else sticking itself into the tank. Am I missing something here or do some mini-tanks not have these?

My HWS is solar with a stainless steel tank. It does not have a sacrificial anode. I understand that stainless tanks don't have the anodes.

Was your tank stainless or the vitreous coated steel?

[edy]

Here are the relevant documents:

http://johnwoodwaterheaters.com/files/literatures/en/Space%20Saver%20English%20SE106N%20Historic.pdf

http://www.gsw-wh.com/assets/documents/current/spec-sheets/GSW%20Spacer%20Saver%20English%20GSE306N_0116.pdf

So I have model # GSW Water Heating, Model SS12SEB15 Cat No. A1229. According to the spec sheets linked above, it is a SIDE ENTRY SINGLE ELEMENT MODEL (6 year tank and 1 year parts warranty).  Capacity 9 Imp. Gal. = 12 US Gal. = 43 L. 1500 W/120 V. It claims to have a removable anode, and is glass-lined tank, galvanized bottom pan.

I found the manual here:

http://www.johnwoodwaterheaters.com/files/products/en/Space%20Saver%20Electric%206-19G%20Models%20Manual%20English%2061515%20Rev%20P.pdf

Looks like the anode is located at #3... How the heck am I going to remove it? They mounted it up near the ceiling, there is barely a foot of clearance above it. Those bastards! 

Seriously, I will have to somehow get up there and drill a hole through the drywall. I don't remember, I hope it is a tiled drop-ceiling like the rest of the office so I can pop up the tile and access and remove the anode rod if needed. And so the plot thickens.

As far as ohm rating, we have P=VI and V=IR. Substituting I=V/R in the P=VI we get P=V²/R, and then if we multiply both sides by R/P we get R=V²/P = (120 x 120)/1500 =  14400/1500 = 9.6 ohm. And then I=12.5 Amp. So it draws 12.5 Amp when 120 V is put across a 9.6 ohm resistance. I should be measuring around this amount on the element when it is out of the tank, and a new one should be that as well I presume.

[Paul Moir]

I once drilled a hole through the floor to get one out.  Fortunately carpeted so patching the hole was easy 
The blowoff is designed to be opened like that, but doing often causes it to leak especially if old.  It's a case of letting sleeping dogs lie.
Good luck finding a replacement anode!  Not many suppliers carry them where I live.  You'll probably have to get a long one and cut it down with a hacksaw.  Also GSW puts these in tighter than even the elements, I think to prevent you from trying to replace them.  2' long breaker bar at the minimum.
Don't know where I went wrong with my math:  you're right 9.6 ohms.  Also I was picturing one of those tiny 7.5l water heaters: you may be able to use one of those bigger elements which are much easier to find.
 

[boffin]

This is what you need

http://www.rona.ca/en/water-heater-element-wrench
http://www.rona.ca/en/water-heater-element

I don't remember them even being that much last time I did it (but that was a couple of houses back).  A decent plumbing supply store will have the wrench, even the local Shenzen Tire probably stocks the stuff.

[Ian.M]

Before butchering the drywall, it would be worth trying snaking an inspection camera in through the element port and trying to image the anode to see how badly wasted it is, and look at the interior condition of the tank.  Tank replacement may be the best/cheapest option.  If the tank's OK, you can get flexible anodes that consist of segments of Magnesium or Aluminium alloy on a wire that can be fitted with less top access clearance.  The old anode can be removed piecemeal - pull it out as far as possible, clamp with molegrips across the port, hacksaw off the length you've pulled, and continue till its all out.

[edy]

Thanks for the suggestions and help. I think I'll try getting the parts from a local store. Based on the RONA prices I see a wrench and a 1500 W 120 V element together is around $30-40 + tax. Then if I change that, forget the anode for now and maybe try to inspect it through the hole it might be easiest.

What happens if I don't change the anode? Will it just shorten the life of the element, or cause only faster corrosion of the tank itself, or both? If the anode is impossible to get out, or will cause more damage, I may opt to leave it alone and just have a plumber change the tank in a few years.

This tank is about $320, but the labor involved in changing it I would guest is more expensive.

https://www.homehardware.ca/en/rec/index.htm/Plumbing-Electrical/Plumbing/Waterworks/Water-Heaters/Electric/10-Gallon-120-Volt-1500-Watt-Electric-Water-Heater/_/N-ntjb3/R-I3284305

If the anode isn't too long and I can access it, I may be able to swap it out with this one:

https://www.homehardware.ca/en/rec/index.htm/Plumbing-Electrical/Plumbing/Waterworks/Water-Heaters/Parts/Gas-Electric-Water-Heater-Magnesium-Anode-Plug/_/N-2pqfZ67l/Ne-67n/Ntk-All_EN/R-I3285082?Ntt=water+heater+anode

Looks like they also have the wrench and element:

https://www.homehardware.ca/en/rec/index.htm/Plumbing-Electrical/Plumbing/Repair-Hookup/Water-Heaters/Elements/120-Volt-1500-Watt-Screw-In-Water-Heater-Element/_/N-2pqfZ67l/Ne-67n/Ntk-All_EN/R-I3285251?Ntt=water+heater+element

https://www.homehardware.ca/en/cat/search/_/N-2pqfZ67l/Ne-67n/Ntk-All_EN?Ntt=water+heater+wrench

Home Depot seems to have some options as well:

Element: https://www.homedepot.ca/en/home/p.water-htr-element-1500w120v.1000120198.html
Wrench: https://www.homedepot.ca/en/home/p.element-wrench.1000120184.html
Anode: https://www.homedepot.ca/en/home/p.42-inch---aluminum-anode-rod.1000150837.html

LOWES has them too, although a tad more expensive:

Wrench: https://www.lowes.ca/water-heaters/water-heater-socket-box-wrench-for-screw-in-element_g1453530.html?searchTerm=water-heater-wrench
Element: https://www.lowes.ca/water-heaters/water-heater-screw-element-1500w-120v_g1453598.html?searchTerm=water-heater-element
Anode: https://www.lowes.ca/water-heaters/whirlpool-9001829005-32-in-magnesium-anode-kit_g1453533.html?searchTerm=water-heater-anode

Is there any advantage to Magnesium vs. Aluminum anode? Is it an issue to cut them down to size if I can only find a longer one, or will it cause some problem? I assume I can just cut the anode to length once I see what my other anode length is if I manage to remove it.

[Ian.M]

Soluable Aluminium salts are bad for your health.
see https://www.atsdr.cdc.gov/phs/phs.asp?id=1076&tid=34
Always cook with cold water direct from a mains tap, preheating it in a kettle if you need hot water, never with water from a hot water system.  Do not drink water from a hot water system.

Aluminium anodes are safe and satisfactory in hard water areas, and last longer than magnesium anodes, but if you've got soft water or your water is acid, you should probably use a magnesium anode if the water is to be used for final rinsing of dishes and cooking utensils.     Magnesium anodes provide more protection to the tank and element but erode far more quickly and in some cases may require annual replacement.  If the water has very low dissolved solids, an aluminium anode may be ineffective.

Most anodes can be cut to length - as long as possible without hitting anything is good.   Personally, I'd pull the element and inspect the tank before buying parts.  Swapping the tank with one of the same make with ports in the same places isn't rocket science provided you've got physical access without ripping out drywall or support platforms etc. and isn't much harder than changing an element and an anode and getting everything you've disturbed watertight again.

[edy]

According to this sheet, we have hard water (123 mg/L of calcium carbonate or 8.63 grains/gallon):

https://www.vaughan.ca/services/residential/water/water_quality/General%20Documents/Water%20Facts111.pdf

So I guess aluminium should be ok.

[SeanB]

Standard wrench and standard element. Be sure to have a new seal for the element on hand, and be aware that the cheap wrenches can and do split on some older elements.  Hard water area probably the anode is still usable, they go fast in soft water, the hard water areas provide a nice film over the piping and element, which will have a nice brown crud film on it and will be missing a large chunk where it blew apart.

When you have it out, drain off the water out of the bottom, using a hosepipe as a syphon, and if not possible and you have a large wet and dry vacuum use it as well to get the crud ( which will look like tea) out of the tank.  When you have gotten the old element out, place a finger inside and feel for splits in the glass coat, and for any rust near there. Any splits, or big rust areas, time to buy a new tank ( buy the same identical model, much easier to put it in again as the mountings, piping and fittings will be in the same location and thus no plumbing) and swap it out.

I have done that before, got the same model tank, and just had to undo the old and put the new in. Then 2 years later I had to drain it to fix a leaking drain tap, by replacing it with a new one. Tank was fine on the inside, but the bottom was deep with sediment, despite having a 100 micron water filter feeding it. Swapped with a 10 micron one before filling it again, not going to have tea now in it.

The 40 year old gravity tank elsewhere also sprung a leak, so took it out, took to the manufacturer ( still in business, and local to me) and they refurbished it, new inner copper tank, new element and thermostat and new filling valve, and I hung it back a week later on the same bolts, connected the pipes to the tank with some new olives and nuts, connected power to it and filled it up again. Half the price of a new one, and a quarter the price of installing a pressure geyser. which likely would have popped all the old sweated joints as well.

[edy]

Success! I went down to Home Depot and picked up a water heater wrench, water heater element, and a few other bits. Total was under $30. I removed the old element, had a peek inside (see attached photos) and noticed my ANODE was covered in some crystalline material like lime scale (similar to the element).

I had a try at removing the anode as well, but no luck. I didn't have all the socket wrench tools and if I try to hammer on it up there, with no help or support, I'll probably break the pipes since this little water heater is not bolted down anywhere. Also I didn't want to risk damaging the head of the anode, so for now I'll wait until I have the proper tooling and help. I can always change the anode even with the tank mostly full.

As far as the element goes, I swapped the elements but noticed the new one didn't quite screw in easily. The threads looked the same, it went in a bit and got stuck. I cleaned the threads on the tank the best I could, lubed it up with some WD-40, tried again, gave it a bit more torque, it threaded some more, and worked it back and forth many times until I presume it cut it's way in to the point where the new element was flush as possible. I noticed quite a bit of dark greyish black stuff on my fingers as the WD-40 was I presume dissolving the crud and the metal components (tank and new element) were rubbing against each other.

For now I have hot water again. I hooked up the thermostat/regulator, flipped on the mains at the breaker panel and 1 hour later everything is fine. So for now I think I'll leave it. Will keep an eye on things though to make sure nothing is leaking.    Thanks guys for all the help!

[edy]

One more thing....

My water is REALLY REALLY HOT. I turned down the temperature as well, to see what happens. For safety, I cut the power to the tank over-night so I don't come back tomorrow morning to an exploded tank.

QUESTIONS:  I know the element works. Could the controller fail shorted? I do hear a "click" sound when I move the temperature knob around, so I figure it is the relay switching the element on and off as I pass the temperature with my set knob.

To test it, I will have to disengage the power to the temperature control and see whether I get continuity as I move the knob around. That will tell me if this thing ever shuts off properly at the higher temperature.

[Ian.M]

Its possible for the thermostat to fail shorted (by its contacts welding themselves together).  Depending on its mechanism, it may still 'click' as it passes through the setpoint. You need the tank hot at a temperature within the thermostat's range, then check it closes when you increase the setting to above the tank temperature and opens when you decrease it to below it.  As a final check, wire a bulb across the element and check it actually cycles as a thermostat may fail  'sticky' so it only switches if physically disturbed.

[helius]

Magnesium is more galvanically active than aluminum and so will sacrifice itself faster. There are also other reasons to prefer magnesium anodes to aluminum ones (actually they are usually a blend of zinc and aluminum).
http://www.waterheaterrescue.com/Longevity/water-heater-anodes.html

The picture you posted looks like an aluminum rod, as it has developed an alum coating (potassium aluminum sulfate hydrate). That makes it more difficult to remove.

[edy]

I wonder if the thermostat is faulty, could the element failure be linked? For example, say the thermostat failed first, stayed on until element overheated and burned out?

Or the other way, element shorts first maybe through some crud deposit, causes high amp across temp control switch fusing contacts short as well, then element short melts down secondarily causing open circuit across element again.

[SeanB]

Well, you tend to replace element and thermostat as a unit, as typically the one dies with the other, the element from thermal cycling every time it operates going from water temperature to 500C internally ( the element wire in the insulation gets red hot in operation), while the thermostat has contact erosion and wear on the pivot points of the expansion element.

New thermostat will fix that, and set the new one to 60C, makes water hot enough to kill the bacteria living in it, and saves a lot on electricity otherwise wasted heating it up.

Tip as well is to take that hot water outlet pipe and put some insulation on it, at least till it goes into the roof. You will see a noticeable drop in temperature in the cabinet there, and a drop in standby power usage. A little on the safety pipeline as well, just do not obstruct the test lever mechanism, will also reduce the heat loss. I use standard neoprene foam AC piping insulation, just buy 2m at the home improvement store, even the cheap stuff works, and a roll of gaffer tape to fix it on after slitting it.

[BurningTantalum]

The insulation available in my locality is really poor- very thin and not pre-slit and hugely expensive (compared to the UK/Europe)
I find that 'pool pasta' is at least twice as thick, rigid, can be mitred for corners and is about 1/4 the price at 'El Cheapo' shops. Plus it is pretty colours too.
And the magpies haven't ripped it up it in 9 years whilst the black stuff went in a few months.
BT

[DTJ]

The insulation available in my locality is really poor- very thin and not pre-slit and hugely expensive (compared to the UK/Europe)
I find that 'pool pasta' is at least twice as thick, rigid, can be mitred for corners and is about 1/4 the price at 'El Cheapo' shops. Plus it is pretty colours too.
And the magpies haven't ripped it up it in 9 years whilst the black stuff went in a few months.
BT

The pool noodles worked well on mine for many years. The UV killed it in the end. I should have painted it. Much cheaper and probably better than the real stuff.

[stj]

2 things to say.

1: it can be handy to connect a neon lamp (with inline resistor) to the element in order to see when it's on.

2: NEVER use WD40 for anything involving rubber seals.
infact never use WD40 for anything except rusty bolts on cars - just throw it away.
it's good at ruining things.
rubber gaskets, chip sockets, fiberglass pcb's - loads of things!! 

[edy]

Thanks for the advice.

Ok maybe I need to get a new thermostat then, just in case. I did turn it on again today, heated up the water and then went to check it.

I put the ohm-meter across the thermostat and moved the temperature set point up and down, heard a click and the meter went from showing 0 ohm (short) when I was set ABOVE the tank temperature, and then to infinite ohm (open) when I set the knob really low.

So it looks to be turning off.

The only problem is, when I set it to 60 C the water definitely is hotter than that. When I set it down to 45 C (the lowest marking) I will have to check. Perhaps it is just out of calibration. It still seems to be working but how can I trust it now?

I've turned off power to the tank, set the thermostat low so it is showing infinite resistance now (since the water is over the temperature set point) and will wait a few hours. If all goes ok, the tank water temperature should fall and the thermostat should flip to CLOSED mode since it will be below the set temperature.

[stj]

if the thermostat housing has limescale on it, it will act as a thermal insulator.

[edy]

The thermostat is clean, the tank is clean and it is sitting exactly where it was before. However, I don't trust the calibration on it. It is cycling on and off, but even at the lowest setting I am getting scorching hot water.

There is a disk that is supposed to pop in and out based on temperature... that toggles the power to the element. The thermostat setting knob adjusts a screw that essentially bends the disk more or less to make it "easier" for it to pop in/out. The lower the temperature setting, the more it pre-bends the disk so it pops open easier at a lower temperature.

If the housing of the thermostat, the metal backing or anything is off with it, it may change the forces required and the temperature calibration will be off. I'm not sure if I want to risk it. So better just to buy another one for now and hope that it solves the problem. I'll keep you posted!

[ADDED:]

I just checked the tank again, I touch the part of the tank where the thermostat is sitting. It does not feel hot. The pipe coming out supplying hot water is warm to touch, and I am getting scorching hot water from the tap. Why is the tank not hot at all? Maybe that is why the thermostat is not shutting off or seems to be "calibrated" off. It is never warming up enough to shut down the element. The tank is not heating up, yet the water is coming out hot and the pipes are hot, and the side of the tank isn't. 

[Paul Moir]

Just pop the cover on and let it sit a while until everything equalizes.  The hot water floats up to the top of the tank, which is where the hot water pipe draws from.  I'll bet the stratification is really bad on a 1500W unit since there won't be much convection going on.
Note that the threshold of pain for most people is right around 45 so it's hard to get a good idea of temperature there by feeling.  It's a very fine line between "That's nice and toasty hot" to "I feel like I'm being burnt.".

[edy]

I happen to have extra sensitive skin when it comes to heat. My wife always jokes about it... while she can hold on to a hot cup of coffee, I need to double-cup and add a sleeve. Perhaps I'm the one out of calibration. 

Anyways, I bought a new thermostat just to be sure, and hooked it up. I feel safer knowing I replaced it, just in case the other one was out of whack. I've set the new thermostat to the lowest setting and I want to see if I am maintaining a nice comfortable level of heat without it getting too hot over the next few hours. Then I know it is not over-shooting.

I let the water go for a few minutes and poured it on my hand continuously, it was not scalding and I was comfortable with full hot water only. I'll check in about 1-2 hours and see if it gets too hot. I should be able to pour a stream on my hand continuously without pain once the hot water runs for about 1 minute and stabilizes the temperature in the pipes.

If I am satisfied that it stays warm over a prolonged period and doesn't appear to getting hotter and hotter, I can rest peacefully leaving it to go overnight. Then I can check in the morning and slowly increase the heat over time. Otherwise I'll have to turn off the main breaker if I think it is running the element non-stop (which I think the other thermostat was doing because it wasn't tripping the set point properly or calibration was off as I thought I had it at the lowest setting but it was still hotter than I ever remember it).

[SeanB]

Yes those thermostats do drift with time as they wear the pivot points for the bimetal sensor and the contact points.

[edy]

So far so good.... I've had the power on for 2-3 hours now, testing out the water periodically.... It is maintaining a comfortably hot level. With the old thermostat it would have been scalding me already after 2 hours being on the lowest setting!

Now to leave it on for a few days and see how things go. I'm more confident now that the tank won't overheat.

I thought thermostats would fail by creating open circuit, not short out. Kind of a bad way to fail, no? If it fails open circuit no big deal.... cold water. But if you fail with a short it could explode your tank, pipes, etc... The reset is still there as a backup but what temperature does that pop out at to cut the mains off?

[EDIT: another few hours later, no problem. Looks like the old thermostat was also out of calibration. Now I wonder, did element fail independently or did a bad thermostat cause the element to burn out, and if so, why didn't the tank blow up from overheating before the element died, if that's what happened?]

[stj]

i had a stat weld shut, and the tank boiled!!!
it didnt damage anything.

[Paul Moir]

I thought thermostats would fail by creating open circuit, not short out. Kind of a bad way to fail, no? If it fails open circuit no big deal.... cold water. But if you fail with a short it could explode your tank, pipes, etc... The reset is still there as a backup but what temperature does that pop out at to cut the mains off?

That's why there's two redundancies:  the high limit cutout, which is the bimetallic thermodisk thing with the red switch, and the T+P valve.  That's why you never hear about explosions on unmodified tanks despite hundreds of millions of them being in the field.
High limit goes off somewhere around 80-90C.  I had to operate a tank at 95C once.