Notes on cleaning sediment from a gas water heater

[pasadena_commut]

Our water heater was made in 1994 by American Water Heaters (model G52-403T4-3N,
40 gallons, gas powered, installed inside the house in a utility closet with the
furnace). It had been bumping for a while, so recently I replaced the anode
(for the 3rd time), after a long vinegar soak, and discovered while trying to
drain it that it was packed up with sediment. This is the tale of the various
things that were tried to clean it out, with some notes on what worked and what
didn't. Not sure how long a post can be here, so this will be broken up into
several parts. All parts now posted, please post comments.

 

[pasadena_commut]

Pictures:

1. Water heater diagram (wh_diagram.png).
Roughly to scale, color coded. The heater sits directly in a drain pan
(not shown) which in turn is directly on a wooden floor. The closet
door is omitted.

2. Brass ball valve drain (brass_drain.jpg)
Replaced the useless nylon drain. The handle is not as close to the
shield as it appears in the photo.

3. Tools overview (wh_tools1.jpg)
Shows the brake line based sprayer, the PEX based sprayer, the brass
Y hose splitter and hose cap, and the two largest objects which came out
during the cleaning.

4. Tools closeup (wh_tools2.jpg)
Close up of the adapters on the ends of the sprayers

5. Object closeup (biggest_chunks.jpg)
A block of calcium carbonate and an RFID tag were the largest objects
to come out of the heater during cleaning. I have no idea how
the RFID tag got in there. A garden hose cap is shown for scale. The
carbonate chunk is so large that it only fits through the brass drain
when properly aligned with it.

 

[pasadena_commut]

Origins of all the hardware components:

Source abbreviations:
AZ = Autozone
HD = HomeDepot
OR = O'Reilly's
?? = do not recall

PEX based sprayer:
1/2" PEX-A 10 ft (cut in half) HD 309830184 EPPW1012 $4.94
1/2" PEX x 3/4" MIP adapter HD 202270503 U116LFA $8.37
3/4" FIP x 3/4" FH adapter HD 300096153 LFA-673 $9.67

Brake line based sprayer:
3/8" steel brake line x 51", flared with male fittings.
AZ AGS PA651 $10.19
3/8" flare nut female HD 207176731 A-160 $5.35 (2)
3/8" flare x 1/2 FIP adapter HD 207176428 LFA-177 $4.25
1/2" MIP x 3/4" FH adapter HD 207176404 LFA-677 $6.27

Brass water heater drain:
3/4" MIP x 3/4" MIP 4" nipple HD 207176458 $15.67
3/4" FIP x 3/4" FIP full port ball valve
HD 205888734 3/4 LFFBV-3C $25.87
(this is what HD has now. I actually used Everbilt 04YE11,
no longer in store at HD, but looks like HD 205812059)
3/4" MIP x 3/4" MH adapter HD 300095988 $6.25

Garden hose related:
3/4" garden hose brass Y splitter, 1 FHT x (2) MHT,
ball valves in each branch
Heavy Duty Garden Water Hose Connector Manifold 2 Way Swivel Y Splitter Shut off | eBay $9.87
3/4" hose cap HD 300096769 LFA-685 $2.85
3/4" FH (swivel) x 3/4" FH (swivel) adapter
HD 206480253 $2.98
3/4" garden hose x 100' x 2 ??

Miscellaneous:
3/4"-14 NPT tap HD 304632101 $14.04
13/16" 12 point 1/2" drive socket
OR GM8293 $6.99
Teflon Tape ??
Pipe Dope ??
Painters tape ??
1 lb. Rubber Mallet HF 69050 $3.99
14" Pipe wrench HF 61349 $6.99
11 Qt oblong bucket ??
Klein Tools boroscope HD 313832948 ET16 $39.97
Ridgid 9 gallon shop vac HD (older model) ~$90
Ridgid power tool adapter HD 100206293 $15.97

 

[pasadena_commut]

First pass - brass drain valve and hoses only.

Initially while replacing the anode it was determined that the heater was packed
with sediment. The tank was drained by vacuuming out the water through the
anode hole using the 10 ft PEX tube taped into the Ridgid power tool adapter
connected to the Ridgid shop vacuum hose. The existing nylon drain (which
jammed quickly no matter what tricks were applied to it) was removed,
revealing a wall of sediment. The brass drain valve (components listed on a
previous post) was installed. One 100' garden hose was run from the brass valve
to the curb. A second 100' garden hose was run to the same location. The tank
was unjammed by connecting the two with the FH to FH swivel adapter and back
flushing, then unhooking that adapter. This was a PITA since the adapter had to
go on and off for each pass. After a while the tank was clear enough that the
cold water valve at the tank could be used. A lot of sediment came out.
Eventually that petered off to nothing. Good enough? The anode was
reinstalled, the tank refilled, and the gas turned back on.

It still bumped.

Hypothesis - the original donut shaped sand bank thought to cover the bottom of
the tank had been only partially removed, mostly around the drain and between
the drain and the cold water inlet. A better method of flushing the tank
was required.

 

[pasadena_commut]

Second pass - brass drain valve, hoses, brass hose Y splitter with valves.

Some days later another method was tried. The brass Y splitter with both valves
closed was attached at its female end (C) to the hose from the water heater's
brass drain valve. It was also attached at one male end (A), valve closed, via
the FH to FH adapter to the hose running to the outside spigot, which was
turned on. The other male connector (B) was aimed into the gutter. By changing
the settings on the valves it was possible to force a back flush (A open, B
closed) or to drain (A closed, B open).

Some sediment came out, but overall this worked poorly. I believe the reasons
it was not effective were three. First of all, the valves are not "full port",
the diameter of the path through the valve is only about 3/8", less than half of
that through the 3/4" brass drain valve. Second of all, the drain hose had a
slight dip in it, of about 4", where it passed through a low point. This may
have allowed some sediment to settle there and further reduce flow. Lastly,
the low gradient (hose was essentially flat in most locations) combined with its
length induced some drag, which lowered the flow rate, which in turn reduced the
amount of sediment which it carried.

On the plus side, it was very easy to change the flow back and forth, and water
didn't spray all over the place on the transitions. This method might work with
a full port garden hose Y splitter (if anybody makes such a thing), especially
if the drop along the hose is greater than it was here.

 

[pasadena_commut]

Sprayer construction

Since the region between the water inlet and the drain was evidently fairly
clean I wanted to be able to aim water at different regions of the interior,
where the remaining sediment must be, and to be able to do so without smacking
into the thermostat probe or the dip tube. (Note, there was a problem with many
dip tubes in the era when this unit was made. They tended to fall apart. This
one was still intact in a boroscope view through the pressure valve hole, but
exposing it to any large lateral forces seemed like a very bad idea. The
access point would be the hole for the the pressure relief valve since it was
relatively easy to remove (14" pipe wrench and a couple of smacks with a 1 lb
rubber mallet and it began to turn, and was then removed without further
complication). That hole threads were cleaned with the NPT tap, driven by the
12 point socket, while the valve itself was cleaned with a brass wire brush.

The first sprayer was made from a piece of steel brake line. These come with
two flared ends and male flare nuts. One end was cut off so that the remaining
length was not enough to hit the bottom of the tank from the access hole, and
was short enough so that it didn't hit the ceiling when placed vertically over
the tank. The male flare nuts were removed and a female flare nut run up to the
flare. Then a 90 degree bend composed of two 45 degree bends separated by a few
inches was made. My son made this piece for me as he has a metal tube
bender. (Previous experience had shown that a single 90 degree bend tended to
pinch the tube nearly closed.) This initial version was then threaded down
through the hole. It was found that the access hole constrained the tilt so
much that basically it just shot straight down, which was not very helpful. So
a second compound 90 degree bend was made in the same plane as the first, and
then a ~45 degree compound bend out of that plane was added. (A single 45 degree
bend would not have passed through the access hole.) This allowed the business
end to be threaded through the access hole, and using both hands it could be
aimed at various angles between the flue and the wall in the direction of the
hot water outlet. A few test blasts in the tank were made with it, but it was
essentially abandoned as soon as its problems became evident.

This sprayer had some major problems. When attached first to the Y
adapter (to use as a valve for turning it on/off), then to a hose to an open
outside spigot, it kicked like a mule when engaged. So there was a high
probability of it banging against the threads of the access hole - not a good
thing. It took two hands to control it, one at the Y splitter valve and one at
the access hole. The stream came out narrow, like a laser bean, and the force
of it against an open hand bordered on painful. Finally, it turned out that the
inside of the steel tube rusted after a single use (surface rust). It was
possible to defang it somewhat by very carefully setting the handle on the
outside spigot to reduce flow. In short, it was a touchy beast and seemed
as likely to cause damage as it was to clean out the sediment.

The second sprayer was based on PEX tubing. The 10 ft piece which had been used
to vacuum out the tank was cut in half and attached to the adapters listed in a
previous post. This was long enough to reach the bottom of the tank, plus a bit
more, resulting in a sort of J shaped piece of tubing aimed more or less
parallel to both the vertical wall and floor of the tank. The PEX was flexible,
which was very handy for getting it into the access hole since there are various
pipes, a flue, and a ceiling in the way above the heater. The PEX is soft, so
where it rubbed up against the threads of the access hole the PEX suffered
small scratches, but the threads would not be damaged. The diameter was larger,
so while there was a lot of flow, the velocity/force was lower. Finally, with
the Y adapter used to modulate flow (one male end to the PEX, the other valve
off and capped, toggle the valve on the PEX branch with a thumb) it was
possible to control it with one hand. This was very convenient when using a
bucket, since the drain is too low to go over any of our buckets sitting flat on
the floor - it was necessary to tilt the bucket with the other hand.

 

[pasadena_commut]

PEX Sprayer use

It was determined that with the outside spigot open exactly one full turn that
10 seconds open on the Y hose splitter valve released 2 gallons of water. This
would not overflow the oblong bucket even though it was tilted. With the PEX
sprayer inserted into the heater it was also determined that it took 1 minute
and 45 seconds for the water to stop dripping out of the tank. (Remember, wood
floor, putting in 3 gallons would have been a problem.) Initially many
cupfuls of sediment appeared with each shot. I didn't want to empty that
sediment into any other pipe in the house, so the bucket had to be lugged out to
the curb each time and emptied. This was repeated 10 times. My hand waving
estimate is that near the end of that series the amount of sediment on the N+1th
pass was around 90-95% of the sediment on the Nth pass. The amount of sediment
at the end was perhaps 1/3 of a cup per shot. It was also observed that the
sediment fell to the bottom of the bucket in seconds, and that most of the
sediment came out shortly after each shot of water. That suggests that some
sediment was probably being displaced initially and then falling back onto the
bottom of the tank as the water flow decreased.

At this point a hose was attached and the tank sprayed ten times for 10s each
separated by 105s pauses. (I didn't want water to accumulate in, and possibly
overflow from, the tank.) Lots of sediment came out. The hose was detached and
the bucket tried twice more. In each case it was still about 1/3 of a cup of
sediment. Oddly, a quick shot of water from the cold water inlet produced
almost no sediment (here the drain valve was closed after 2 gallons out). So
the total amount of material coming out per shot had leveled off during the
initial 10 shots. There were some differences though, with some very fine sand
like material having been deposited at the curb. The hose was reattached and
another 20 shots at similar time intervals applied. For some of these the
PEX was repositioned to aim at slighthly different directions. At the end of
that there was still a little sediment coming out of the hose, but it wasn't
really possible to estimate how much, like it was in the bucket.

Done? The boroscope was put through the access hole but all it could see
straight down was its own reflection, either in the glass lining or in a thin
layer of water on that lining. Based on the small amount of sediment at the
last PEX spray there might still be a bit of the sand bank at around 10:00 (on
the WH diagram, thinking of it as a clock) but I could not guide the boroscope
over there successfully. That meant that at least half the bottom was clear
(from the cold water inlet to past the pressure valve hole). If it had been
earlier in the day I would have washed some more, but the sun was going down, so
it was time to stop.

Wrapped the pressure valve in 1.5 turns of teflon tape, then put pipe dope over
that, and screwed it in. Closed the drain. Filled the tank. Turned on the gas
and lit it. No bumps. No bumps the next day. Two weeks later and still no
bumps. So while it might not be entirely sediment free, it is now symptom
free.

 

[pasadena_commut]

Conclusions and miscellaneous

The PEX sprayer worked very well. It might be helpful to make the PEX tube a
bit longer so the water exit point could be advanced a bit clockwise around the
tank from the current location. But I don't think it would be a very big
difference, since the water shooting out of the sprayer is going in that
direction anyway, and the water probably migrates towards the outside of the
tank and circulates around there. The Y splitter valve used to control the
sprayer was OK, but any other valve which could be threaded on and turned on or
off with one hand should work just as well. I didn't need to use it for this,
but because the PEX tube fits into the brass drain valve, it would have been
possible to clear out the initial massive jam by catheterizing that valve and
blasting water through the sprayer. (Probably with a rag wrapped around it to
prevent water from shooting back out again in the narrow space between the valve
and the tube.) In other words, it would have worked as well as the initial
backflush method, but without all the futzing around with FH to FH adapter
at the curb.

It would be very helpful to have some sort of clear sluice or filter between
the drain and the hose. Using just a hose one must keep running out to the curb
to see what is coming out. Using just a bucket provides direct observation of
the sediment coming out, but requires too many trips carrying a full sloshing
bucket to the curb. The ideal would be to see the sediment as it exits the
drain, and capture it there, while releasing the water (only) to run off to the
curb. The tank drain in this instance is pretty close to the ground. To do
this job I picture a plastic trough about 2 feet long, 1 foot wide, and 7 inches
high (~8.7 gallons) into which the drain empties, with a mesh screen basket
about half as long under that point to catch/filter the sediment, and a fitting
at the bottom of the "far wall" to let the water drain out through the hose.
When the basket fills up it could be removed and dumped into a plastic pail for
later disposal. In most instances there probably wouldn't be ~4 gallons of
sediment, so the basket would only need to be emptied at the end.

Lastly, a gripe on gas water heater tank design. The manufacturer installed
drain valvess seem to have been purposely designed to be useless for removing
sediment. Not only do they constrict the pipe width at the valve to a tiny
easily jammed diameter, but the exact radial orientation means that a back flush
goes straight into the flue rather than being directed around the tank where it
could do more good. Whatever their other pros and cons might be, an electric
water heater would be much easier to clean out. Removing an element provides a
larger access hole, and the lack of a central flue means the entire floor of the
tank can be accessed from that port.

Please post comments.

 

[wpns]

Wow, that's a lot of work, congrats on getting it working.

As I like to say after completing that kind of project: "Good thing my time is worth nothing."

I'd be concerned about the remaining life of a 28-year-old water heater, but IIRC from pervious postings it'd be really difficult to replace, and you have been replacing the anode, so it may last a while longer.

Hopefully you've learned your lesson and will hook up a hose to the curb and clean it out on a more regular basis.

I really like your drain mods, I'll have to do that for my electric water heater in my BoatHouse, draining thru that stupid plastic fitting is a real PITA.

Thanks for sharing your project, I learned a lot!

 

[pasadena_commut]

The tank has been bumping for a couple of months, and it is a bit past a year, so cleaned it out again yesterday. Since I was going to do this anyway, it seemed like a good opportunity to see if a "power flush", (flushing the tank with the inlet open) was significantly better than just draining the tank and letting gravity do the work. The outlet material was in some instances collected in a dog pool at the curb, and in other instances in a bucket inside.

Three sediment collections were made:

The first collection was from power flushing the tank for 1 minute continuously (Tub1A).
The second collection was from power flushing the tank for 2 minutes continuously (Tub2A).
The third collection was from an initial drain (inlet closed, T&P valve open, only a little sediment came out, no picture), then 5X inlet on for 10s then off 30s (Tub3A), then close T&P, open inlet for 30s, then open T&P valve after 30s (Tub4A), then inside, using a bucket, remove T&P valve, 3 shots with my PEX cleaner (see previous posts), then 7 times pouring 1.5 gallon buckets in through a funnel (last one had almost no sediment).

It is hard to see in the collections picture, but the ratio of sediment volumes is roughly 5:2:3. The total amount of sediment collected was roughly 750 ml, measured after decanting as much water as possible without pouring sediment out with it.

The Tub6A picture is from a final 5 minute power flush, after the T&P valve was reinstalled and the tank refilled. There are still some rust particles, but only a few scattered calcium carbonate chunks. This house has old rusty galvanized pipes. Whenever a very high flow occurs those get knocked off the pipes, and this was a very high flow situation, with two fully open ball valves at the inlet and outlet. So the rust went through the tank, but I'm not sure how much, if any, came from the tank.

Results:

The power flush works faster than the usual gravity drain, getting out in just a couple of minutes 70% of the sediment that came out in total. I think though that it has the same problem as the gravity driven flushes, not cleaning the parts of the tank that are far from the inlet/outlet/T&P valve hole well, because the tank still bumps. There is probably still a "sand bar" in that region.

Notes:

The original plan for the third collection was to use the PEX cleaner exclusively, but my second hose sprung a leak between the threads and the crimp, and since this was inside on a wood floor, even a small leak is a problem. So I had to resort to the bucket and funnel method.

There is no Tub5A picture because the original was a duplicate of Tub4A, and I didn't want to renumber.

Before putting the T&P valve back in a boroscope was put into the tank through that hole. The area immediately under that hole had no visible white sediment pieces. However, moving clockwise from there I could see an increasing number of such particles at the junction of the floor and vertical inside wall, but still none "up the slope" on the floor. I don't know exactly where the tip of the boroscope was, but I don't think it even reached 9:00 on the diagram in post 2.

When the PEX tube was first inserted it was pushed all the way to the bottom of the tank, which was hard. If there had been significant sediment there it would have felt squishy.

This was not a completely successful flush, the tank was still bumping when it was turned back on. This is the same as what happened last year: to get all the sediment out the PEX sprayer has to be used to reach the far parts of the tank.

The dog pool was used, instead of a bucket, because the flow rate was very high during the power flush. It wouldn't have been possible to collect the sediment because the roiling water would have kept it floating and sediment would have washed over the top. With the hose placed at least a foot from the edge of the pool, and the pool already full of water, the sediment had time to settle out, and water overflowing it at the low point was essentially devoid of sediment.

 

[axxell]

Fascinating post. Wanted to ask your thoughts on doing a chemical soak/flush? Looking to try calci flush or Mag Erad.

 

[pasadena_commut]

Yes, one can pour acid into the tank and it will eventually dissolve the calcium deposits. It could also burn holes in the metal, or open one up which had not quite corroded through. I prefer the flush since it is really unlikely to damage the tank.

I have used vinegar twice (just table vinegar, it comes in cheap large containers from Smart and Final) and it didn't start any leaks. It is only going to get all the deposits if there isn't much to begin with. Neither got all the deposits. The second time I tried it it turned out the tank was completely packed up with sediment (at least 4" thick) and all that happened is the deposits neutralized the vinegar. Quickly. Luckily I had some pH paper on hand to confirm this. That is probably what happened the first time too. Not sure how much vinegar it would require to remove a thick deposit, but "a lot" seems like a good estimate. Knowing what I do now, I would mechanically flush the tank until no more sediment came out, and then if it was still bumping, try vinegar to get into the regions which were still somehow covered in deposits. Note, "no more sediment" is only good if there is a good flow in the tank. Without using my PEX washer gizmo, just blasting the inlet and using a big outlet, it clears out the region between those points but leaves sediment in other parts of the tank.

I cannot prove it, but think that the junction between the floor of the tank and the wall can trap sediment, at least in our water heater, so that even when the tank appears completely clean via the parts reachable by a boroscope it can still bump a little when the flame is on.

 

[sarg]

Why I prefer an electric heater ............ a wet vac through the bottom element hole & done.

 

[SGkent]

1994 - replace it. Average life is 10 years with you maintaining good anodes. I get 20 years and then replace on good practice. You can choose to replace it on your timing, or wait until it forces you to do it at a time when it will get in the way of other things more important. With inflation going hog wild, you probably could have done it for 30% less three years ago. Wait three more years and it will add another $350 to the cost.

 

[pasadena_commut]

Why I prefer an electric heater ............ a wet vac through the bottom element hole & done.

Agreed, electric would be much easier to clean out.

Cleaning out a gas WH is like doing laparoscopic surgery, and having to build the instruments too.

 

[BlueSkyHigh]

Agreed, electric would be much easier to clean out.

Cleaning out a gas WH is like doing laparoscopic surgery, and having to build the instruments too.

I just hooked up a 50 ft. garden hose to the drain connection and "power flushed" the unit until the white stuff stopped flowing. Probably wasted a lot of water and, using this method, is a wetvac really necessary?

 

[cgilley]

What I took from all this is I need to flush my water heaters. Excellent write up.

 

[pasadena_commut]

I just hooked up a 50 ft. garden hose to the drain connection and "power flushed" the unit until the white stuff stopped flowing. Probably wasted a lot of water and, using this method, is a wetvac really necessary?

The advantage of an electric WH over a gas WH with respect to cleaning it out is that you can see into it without heroic measures. With our gas WH even with a boroscope I could not examine all of the bottom from a single access port on top (the TPR or anode holes). In addition in an electric WH you can reach any and all parts of the bottom easily. If the sediment is dry(ish) a shop vac is pretty quick and easy way to remove it. That isn't a good option on a gas water heater. In fact, I tried it, but had to hook the hose to a piece of pex to get it through the small hole. The shop vac did not like sucking wet sediment up that high (5'?), but it did a good job sucking out the water on top of the sediment. Going in through the drain hole with the PEX vacuum only accessed part of the bottom because the flue up the middle is in the way.

 

[Ludington]

Why I prefer an electric heater ............ a wet vac through the bottom element hole & done.

What size / type hose are you hooking to your vac ? My vacs hoses are way to big to be going through the element hose. And how are you guiding that hose around once you get it through the element hole ?

 

[sarg]

What size / type hose are you hooking to your vac ? My vacs hoses are way to big to be going through the element hose. And how are you guiding that hose around once you get it through the element hole ?

I utilize an auto heater hose .......... approx. 3 feet. I tried some clear tubing to monitor the removal but it turned out to be less flexible.