Blocking convective loops?

[Kcodyjr]

Meaning the kind where hot water rises up through the otherwise-still pipe, reaches a cold pipe above, cools the water that just rose, making it fall, where it gets heated again; etc etc etc, slowly sapping energy out of the system.

The little info sheets that come with ASSE1017 valves, like the Watts MMV and 1170-M2 series, show pictures of the piping being routed right over to the side of the tank and then down a few inches, hanging in the air.

Is all that really necessary? Space is limited, it would be well if I could minimize the amount of stuff that's physically supported by the tank.

I propose to use a pair of 3/4" cast copper wyes, with the flow coming into/from the 45 degree leg, and into an S made from a pair of long-radius ells. The top position of the wye would have a bottom-entry temp gauge on the hot side, and the vacuum break on the cold side. The "S" pieces would then get copper flex tubes running to the rest of the apparatus, mounted on the wall.

Would that cut the mustard? Eventual tank replacement is already going to be a bastard getting it in and out, even without all that extra piping and hardware causing clearance issues.

 

[Dana]

Have you looked into heat trap nipples?

 

[Sponsor]

 

[LLigetfa]

My former (second) water heater had those heat traps and they were problematic. I am so glad my new tank doesn't have them. My first tank didn't have them either. The piping from my HWT goes horizontal right away and then turns and goes down inside the stud cavity so I doubt convection was ever really an energy robber.

@Dana do you have any reports on studies that show what energy saving there can be?

 

[Dana]

I've never seen third party published studies on heat trap savings, but it clearly depends on how the thing is plumbed, the storage temperature, the levels of pipe insulation, etc. I threw it out there because it's a product specifically designed to limit the type of convection kcodyjr was looking to tame. Many new water heaters come with heat traps already installed, are you SURE yours doesn't have them?

There are numerous heat trap designs, and I would imagine some are more reliable than others.

 

[LLigetfa]

are you SURE yours doesn't have them?

I cannot be sure but the traps on the old HWT make a lot of clunking noise and there was a notable hesitation when the hot tap was turned on. The new HWT doesn't have the hesitation and is completely silent. The old HWT even clunked on cold water use albeit without the hesitation.

Wife has a large soaker tub with high flow roman tub filler. If she opened the cold tap first, it robbed enough pressure from the HWT inlet that when she opened the the hot tap second the heat traps might not open and the tub would fill with cold water. I had to train her to open the hot first and the cold second.

If ever I drained the plumbing to service taps, when I turned it back on, mineral deposits would get caught in the heat traps and I'd have to bang on them while she opened and closed the tub filler to get it going again.

 

[Kcodyjr]

Dana, I had not heard of those; but from what I'm reading, it's not the right solution. A ball and spring is providing orders of magnitude more force than necessary, and introduces reliability problems, especially where pressure is marginal. To actually need that much force to stop natural convection, the water tank would have to be the size of my 12'x60' house trailer.

Tell you what, I'll try out my S-curve approach, and if anyone can think of a way to test whether it's working, I'll do so and report the results. We're only talking 20 bucks per side's worth of copper fittings, and I don't see how the arrangement can harm anything.

 

[Dana]

There is no spring involved- it's gravity doing the work. One side is a float-ball, lighter that water, the other is heavier than water. The forces are quite small for that type of heat trap.

 

[Kcodyjr]

What's causing the reports of issues, then? Are the parts necessarily so light they get banged up as the flow comes on?

Another thing that bothers me is the flow restriction that a ball inside the pipe necessarily introduces... and it just feels like brute-force engineering.

The way I read it, the hottest water rises due to buoyancy, displacing whatever's above it. Near a top-barrier, that displacement flow will be sideways. With the layout I described, the top barrier is a dead-end at the bottom of the temp gauge or vacuum break. That will collect the hottest water, that rose with the most force, and turn itself into a pressure head, lowering the position of the apparent top of the pipe, but making the top cushiony, tending to absorb and stop water rather than encouraging a flow that bounces off the wall.

That should result in a much lower amount of hot water creeping out the split in the wye, where it should settle at the apex of the long ell. With even curves on both sides, and what amounts to a speedbump right below it, it should go no further. Remember, on the other side of the apex is hot water up high over cold water leading down, which won't want to go anywhere.

My reasoning is that by creating that correct non-inverted state, that remains correct even as the water heater heats and the water in the pipes cools, I've stopped any further energy transfer from happening.

The remaining problem is conduction. If the entire "S" is made out of metal, heat will be conducted to the bottom-most point, where it will drive a new convective loop. So I've realized, I shouldn't make a whole "S", I should make the whole piece look like a lower-case "r", where the copper ends in a MIP fitting pointing downward 45 degrees, and connects to one of those stainless-braided, PVC-core water heater hookup hoses.

I'd still be thrilled to hear ideas how I might test whether it's working.

I'll post a picture once I've got an assembly built - got to go pick up some thread fittings.

Last-minute thought: seems to me, the shorter the leg between the wye and the long ell, the less time that water will have to pick up velocity. Meanwhile, the top of the wye wants a little space; ideally, the bottom of the temp gauge should be just about halfway down the opening to the side of the wye. Close enough for a good reading, not sticking into the running flow more than necessary, and providing as much space above for the hottest water as possible.

 

[Kcodyjr]

Here it is, test-assembled (more work to go before everything the tank connects to, goes somewhere).

This convective block is brought to you by the letter "r" and the fact that water is about as bad a heat conductor as PVC.

 

[Dana]

Nice bit o' abstract sculpture!

MA (IRC 2015 ) code requires any hot water distribution plumbing 3/4" or bigger to be insulated to a minimum of R3 (including the flex.) Not required by code (but still "worth it" at MA electricity pricing), the nearest 5-10' of cold feed to the tank could use R3 as well.

 

[Kcodyjr]

Nice bit o' abstract sculpture!

MA (IRC 2015 ) code requires any hot water distribution plumbing 3/4" or bigger to be insulated to a minimum of R3 (including the flex.) Not required by code (but still "worth it" at MA electricity pricing), the nearest 5-10' of cold feed to the tank could use R3 as well.

It'll get it - actually, I think everything in that closet is within 5' of the cold inlet. I was going to insulate the entire apparatus once complete, including the cold feed all the way to the inlet, and the filter setup that's off to the left of the image, to prevent sweating in the summer. Moisture is a mortal enemy in a 47 year old trailer.

To get R3, do I want the stiff styrofoamy stuff that needs wireties or tape, or do I want the floppy rubberfoamy stuff that comes with tape in the seams?

 

[Dana]

The cheap foamed polyethylene stuff is fine for domestic hot water temperatures, even for heating system plumbing <<200F (or <<180F for some vendors.)

To hit R3 polyethylene pipe insulation usually have to have about a 5/8". The half-inch walled goods are usually only R2-ish, though some manufacturers are now using a higher density/smaller cell polyethylene that's north of R3 @ 1/2". (eg Armacell Tundra , trade name "Everbilt" available at the big orange box-stores. )

The more flexible rubbery stuff is usually good for higher temps, but is substantially more expensive.

 

[Kcodyjr]

Okay, since I intend to run the tank at its max setting of 150, and stratification means there could be water north of 175 at the top of the tank, I'll use the expensive stuff between the tank and that tempering valve you can just see peeking out from behind a brass wye in the upper right, and use the regular stuff beyond there.

Thanks for the product reference - I'll go visit the orange guys and get that.

 

[Dana]

The Armacell stuff is rated for an 180F operating temp, which has some margin. Your water heater probably isn't rated higher than 180F. I think you're pretty safe to go with the cheap stuff.

 

[Kcodyjr]

Here's another reason, I'm not sure the polyethylene is going to make those tight bends I've got in those flex lines.

This stuff says it's R3.1:
https://www.homedepot.com/p/Everbilt-3-4-in-x-6-ft-Foam-Self-Seal-Pipe-Insulation-ORS07812/204760815

And for $7, I can spring for one for the flex lines:
https://www.homedepot.com/p/Armacel...-Seal-Pipe-Wrap-Insulation-HST07812/100585113

 

[Dana]

I think you've got this figured out then!

 

[Sylvan]

Easiest fix , install a 27" heatsink on the CW supply (loop 27")

This will prevent starification of hot water molecules entering the CW line without have so much friction loss as going through a rubber flapper

 

[Kcodyjr]

I can report that the convective block seems to be working, at least by gross measurement - when the system has been idle a few hours, the "r" gets hot as the dickens, and the bottom of the curve on those flex lines is as cool as a cucumber.

I can't swear that no heat is leaking past it, because I'd need multipoint calibrated precision measurements over several days with no water use and stable ambient temperatures - which are no, no, and no, respectively.

I can also report that, in my entire rebuild, I only had one sweat joint leak - and I'm ticked with myself about that one. Consequently, at the moment, my two-stage whole house filter is set for bypass. Sure glad I sprang for those 3-way valves to implement bypass, backflush, and purge functions.