Water Heaters: More than you wanted to know

David D. left a long comment to my post from August about fixing Paul & Debbie's water heater, which got me thinking... I've never done a post covering the basics of keeping a water heater running. So here's my post on the subject, as a service to my friends, and to the intarweb-at-large (if it's useful to anyone). I find it frustrating that gas water heaters are pretty much a throwaway item--they pretty much last as long as the warranty period (6-12 years), and then fail spectacularly, flooding your basement, possibly geysering all over the place. Also, I'm including some information about water heating technology better than the standard $200-300 tank, such as instantaneous water heaters, and boiler sidearm tanks.

Water Heater Longevity


General Disclaimer: If you're doing some of this work yourself, be sure to keep in mind that you're dealing with 40 to 100 gallons of water at 140 F, inside a gas-fired appliance. This can be dangerous (duh); please behave accordingly.

First of all, I should give credit where it is due: a lot of the knowledge that I have about water heaters comes from the website www.waterheaterrescue.com--they have a wealth of information, and they also sell retrofit kits to increase the lifespan of your heater. I've bought lots of items from them, which are probably installed in half a dozen water heaters around the country.

So you can check out their full story on water heater longevity... t'sokay, I'll wait. However, let me see if I can summarize the story for those of you in a rush. I'll cover two important topics: tanks rusting out, and sludge/scale building up at the bottom.

A gas water heater is conceptually simple--check out this cross section. There's a burner at the bottom of a tank of water (the burner's a bit like a turkey fryer burner; there is a hole through the middle of the tank of water, so the exhaust gases can rise up, heat the water, and escape out the flue.

The problem with a big tank of 140 F water is that it really wants to rust out--think of the acceleration of chemical reactions with temperature, the Arrhenius equation, all that. A "normal" tank deals with this in two ways. First, the inside of the tank is glass-lined; I believe they put on powdered glass and then bake it on. Second, there's something called a "sacrificial anode"--it's a bar of magnesium or aluminum, installed from a threaded fitting at the top of the tank. If there's a nick in the glass coating, rust preferentially starts on the anode (the more reactive metal)--similar concept to zinc anodes used to protect boat propellers from corrosion. The problem, of course, is that as they protect, the anode gets eaten away. When the anode is entirely eaten, the tank starts to rust--see below for an anode that ain't doing anything no more.


This shows a comparison between a brand new and dead anode (there's a steel wire running down the middle of the anode; that's the "skeleton" that you see there).


So, makes sense to replace anodes when they wear out (maybe every six years or so), right? Well, hellsyeah. Unfortunately, I've never met a plumber who does this. This would also require a buying public who realizes that this procedure is necessary. So instead, we just landfill our water heaters every few years.

It is also unfortunately annoying to replace anodes (requires a 1-1/16" hex socket and a 1/2" breaker bar. 1-1/16"?! WTF? Is this the manufacturers trying to make it a "no user serviceable parts inside" situation, to increase their sales?). But I've done it a bunch of times--with the right tools, it's not too bad, depending on overhead clearance space. 1/2" breaker bar shown below, with 3/8" socket, 1/4" socket, and dollar bill for scale.


The second thing: tank sediment and sludge. When you heat water, certain minerals (such as calcium) become less soluble--I know that is counterintuitive to our everyday experience (sugar dissolves more easily in hot tea than iced tea), but that's the chemistry. As a result, when you heat a tank of hot water, some minerals start to plate out (forming "scale")--solidified calcium carbonate or others. Why does this matter? Well, it all falls to the bottom of the tank (where the burner is), forming a great layer of insulation between, uh, your burner and the water (see that cross section from before). That's why when you drain a water heater (with that valve at the bottom), you often get "sand" or "pebbles" coming out of there. In addition, if it hasn't been drained for a while, you end up with sludge (reddish--iron/rust based, maybe?) at the bottom. Similar problem.


So, what to do? A curved dip (water input) tube, to stir up the sediment and get it out of the tank, and regular flushing of the tank. Flushing the tank maybe every six months is a good idea--shut the burner down to "vacation," leave the water on, and flush out water from that bottom drain (use a short length of hose and a bucket, or a floor drain).

Instantaneous Water Heaters

So, moving on to other technologies: tankless/instantaneous (or "on demand") water heaters--why should you care about these newfangled things?


First, I need to explain water heater efficiency; it is rated in something called energy factor (or EF). This measurement rates how much of the fuel’s energy comes out of the tap in hot water, under "average" conditions. It takes into account heat lost up the flue, heat lost from the water sitting in the tank (standby losses), and how much of the fuel’s energy gets captured in the first place.

The common gas tank hot water heaters have EFs in the region of 0.50 to 0.55. You can get gas storage tanks with much higher efficiencies (American Water Heater Polaris, 0.86 EF)--but they're up in the $2000-3000 purchase price range.

Why are tank water heaters so inefficient? First, they're holding a tank of water at setpoint 24/7/365. Second, even though they have insulation on the outside, check out that cutaway again--the inside of the tank is an uninsulated chimney, so any air that is rising through that flue pulls heat out of the water (you can't insulate it, otherwise, the heat wouldn't get into the water when you fire the tank).

So instead, instantaneous hot water heaters have a burner big enough that it can heat the water in a single pass, without any storage. In order to do this, the burner needs to modulate (ramp up and down), to match the burn rate/input to the temperature and flow of the incoming water. With this change in technology, you get EFs up in the 0.80 to 0.86 range; they cost about $400-1000, depending on size and manufacturer. They have been a pretty typical technology in Europe and Asia; they are gradually catching on here in the US. The big water heater manufacturers are starting to take notice, and resell instantaneous water heaters under their own brand.

The savings might be a function of your hot water usage patterns--but considering that many of my friends are couples that are not at home for the entire day, that schedule would make sense.

Also, instantaneous water heaters have better combustion safety than a typical ("naturally aspirated") tank--they are "sealed combustion" (draw air for burning gas from outside)--next to no chance of carbon monoxide backdrafting into the house.

So what's the downside? As a California report (PDF) on this technology notes:

Interviewing contractors, manufacturers, and building officials identified installation challenges and market barriers. High cost was identified as the predominant market barrier, since installed tankless costs are typically two to three times the cost of tank-type heaters. Increased installation costs result from higher equipment costs, larger gas line requirements, more costly venting, and the need for an electrical outlet required for the ignition controls and combustion air blower. Incremental costs are much lower for new construction because gas line upsizing and venting issues are more easily handled when a home or apartment is being built.

Also, there are a few quirky behavior issues--if you have really a low flow (slow trickle) of hot water, the modulation might not be able to keep up--so it might run cold instead. The heater takes a few moments to catch up with demand--it has to sense, "Ok, water's flowing... let me see, the flue pipe is clear... now we can fire things up." One result is that if two people take showers back to back, with a pause in between, the initial water out of the pipe will be hot (the residual water in the pipe)... then you get a slug of cold water (the unit firing up), and then it's hot.

I've been recommending that folks try out this technology for a while, but frustratingly, I have not had any takers yet. However, Paul and Debbie are considering getting one as their replacement water heater. They have run into a bunch of obstacles from the plumbing community:

The first guy we spoke to did everything he could to try to talk us out of it. He brought up the cost, the calcium buildup, etc etc. We ignored him and looked around some more. Then another guy approached us and we asked him what he thought of the tankless water heater. He was much more knowledgeable and encouraged us to consider it and schedule a time for the installer to come in and give us an estimate. So we went to the service desk, but they'd never heard of anybody installing a tankless water heater. Ugh. It was already after 5 by then, and whoever they tried to reach was gone for the day.

It was very interesting that everyone who was against the tankless was old (over 55 at least), and the one guy who was energy conscious and recommended it was our age.

If any of my readers have one of these, please let me know how your experience has been. My former coworker who is building his own house had one in his previous house, and had no complaints whatsoever.

Hot Water and Boilers

New England is the home of boilers and radiators (either hot water or steam), so I would be amiss if I didn't mention how that option works out--after all, that's what's in my basement (as well as A & Guy, U5 & Rebecca, Air Force Guy & wife, tEp, my office...). This system is called an "indirect water heater"--you have a hot water storage tank, and a loop of pipes connected to a boiler (typically used for heating the house as well). The controls are set up so that the hot water tank is another "zone"--when the water is cool enough, the boiler fires up, and the pump circulates the boiler water through a heat exchanger in the tank.



One advantage is that a typical indirect tank (HT Products Superstor Ultra) is stainless steel--lots more longevity than a regular tank. Also, the heated surface is that coil (not the bottom of the tank)--so when scale builds up there, it eventually cracks off, and falls to the bottom, where it is mostly harmless.

So what's the efficiency? Well, despite the fact that I work in this field, I've never found an actual EF rating number. One problem is that it depends on how efficient your boiler is. Second, the amount of piping also has an effect. As a ballpark, they should probably be a bit more efficient than typical gas tank water heaters--maybe up in the 0.60 to 0.70 EF range, when using a typical 75-80% efficient boiler. It might be much better with high efficiency condensing boilers... but that's a topic for another time.

Note that an instantaneous hot water loop through a steam boiler is a pretty bad idea--I did some work on an apartment building that uses this system. You basically have to keep the contents of the boiler hot enough year round (including the summer) that it will heat up your hot water in a single pass. They were looking at $80-100 worth of oil per month in the summer, just for hot water (in comparison, a sidearm tank runs us about $50-60 in Arlington). The hot water is provided by that pair of pipes in the black plate on the side of the boiler.


Finally, as a reward for any of you who have read all the way down to the bottom--be sure to check out The Mythbusters video clip on overheating a water heater and making it explode, making it into a rocket. It had me giggling uncontrollably for repeated viewings.