Please see below to bypass the Hot Water Heater section, only partly successful, and see the Heat Exchanger approach.
The goal of the project was to heat my hot tub in an economical way. The existing built-in electric heater used almost 6000 watts when running and at the new electric rates, keeping the tub up to 104 F made the utility bill very high. I attempted for a while to keep the temperature at a lower setting and turn it up only in anticipation of using the tub. This was unsatisfactory since it makes spontaneous use of the tub impossible and still used expensive electricity for heat.
The idea began when my 21 year old household gas water heater appeared to develop a leak. The leak seemed to grow so I bought and installed a new tank gas water heater. I then had a somewhat viable gas water heater on hand. Why not use it to heat water for the hot tub? The hot tub is located in the back yard lower patio, and the only apparent place for the extra gas water heater would near the gas source next to its replacement in the basement, about 40 feet distant from the tub.
The hot tub has provision for an ozone generator, but this option was never installed. What this means though is the tub has an unused jet with stubbed off 3/4" hose connected, and a stubbed off hose on the manifold which comes after the hot tub pump and electric heater. The manifold has several tubes which direct water from the pump to the various jets which shoot water into the tub. My plan was to connect the stubbed ozone tubes to the tank heater in the basement and disconnect the electric heater.
The hot tub sits at almost the same level as the base of the water heater. But I wondered whether the 5 foot tall heater if filled to the top would siphon into the tub and cause it to overflow. I therefore planned to fill the water heater only partway leaving an air cushion at the top. Tank water heaters have a tube on one side that leads the cold water to the bottom of the tank. On the hot side, the assumption is a full tank and that hot water will be taken off the top. Anticipating an air cushion in the tank, I soldered in a 3 foot copper tube so hot water would be drawn from further down in the tank. One half inch copper pipe soldered neatly into the brass nipple I had.
The hot tub pump runs at two speeds. It comes on slow when it comes on automatically twice a day for an hour of filtering, and slow also when the thermostat tells it it needs heat and the heater circuit is turned on. It comes on fast only when manually directed by push button and then only stays on fast for 20 minutes. I wondered if the hot tub pump at slow would have enough pressure to move the water over the relatively long path to and from the gas water heater.
I installed a tee and new valve to provide a gas source. I decided on PEX tubing to carry the water between the heater to the tub and bought a 100 foot length at Lowes. I didn't want to invest in special crimping tools for connecting the PEX so I opted to use hose clamps, thinking the pressures would be less than houshold water pressure and I could always do it the official way later if needed. 3/4 inch PEX doesn't bend to any less than about a 3 or 4 foot radius, so for the final 6 feet where the tubing goes through a basement vent then to the hot tub I used 3/4" clear plastic tubing.
When the water tubes were connected, I disconnected the electric heater, filled the tub and tried it. I found that on low speed, the hot tub pump didn't have enough power to start the flow to the water heater, but on high it did. I got it started and let air out of the tank using the pressure relief valve. It was hard to estimate how full the tank was, but I let it fill so the tank was quite heavy and air was purged from the tubing. But with insufficient pressure to start the flow at low speed, it became clear I needed a separate pump.
In fact the separate pump was a fine idea because I really only wanted water flowing to and from the tank when the hot tub was trying to get hotter. I ordered a pump and isolation valves and installed them. I used the switched leg of the 220 circuit to power the extra pump. That way the extra pump would come on only when the hot tub wanted more heat. The pump is a Taco model 007-F5, requiring flanged connectors. I got connectors with isolation valves.
I fired up the water heater, but noticed immediate boiling sounds. My conclusion was there wasn't enough water in the tank. So I let more air out via the pressure relief valve, not knowing how full it was. Eventually water came out. So there is now precious little air left in the tank. Siphoning back and overflowing the tub did not seem to be a problem. I increased the tank thermostat and though it still made boiling sounds, at least I knew there was plenty of water. I let it run and eventually the hot tub shut off at its trial low setting of about 80 F. I turned it up to its max and desired temp of 104 F and let it run, also turning up the tank heater to the high side of "warm".
In the evening I heard the hot tub running for its hour of filtering. And in the morning I found it had shut off and the tank heater had shut off. But the tub temperature was about 108 F, too hot. The temp in the tub seemed to indicate that even on the filtering cycle, when the extra pump does not run, water flows to and from the tank heater. So to make sure no water flows unless the tub calls for heat, I put an electric valve in line with the extra pump.
The heater works fine. I'm glad to have my hot tub back. The apparent leak seems to have disappeared, perhaps because there is less pressure in the tank. When the tub calls for heat, it takes about a minute to detect hot water coming out the "ozone" jet - probably because the pump must build up pressure in the heater tank before the flow can start. The water coming out seems very hot, but comes out at low volume so it seems to run a reasonable time period before shutting off (about 10 minutes). The only problem now is that sandy grit keeps appearing in the hot tub, I assume from the water heater. There was a lot initially, but after siphoning it out and washing the filter a few times, the amount has diminished. I guess I should have flushed out the old heater before hooking it up.
Update some years later: The old water heater finally gave up the ghost. The symptoms were that I would find the hot tub very low on water all of a sudden, with no sign of water leak. I'd top it up and life would proceed. Later I'd find that the level was very high. I had to drain some out to get a reasonable level. I tried the pressure relief valve on the water heater and only a small amount of air pressure came out, even when the water was being heated (boiling). Then listening at the top I heard a small hiss - not from the relief valve. The crisis came when the heater pilot wouldn't stay lit. I replaced the thermocouple but that did not do the trick. So I decieded it was time for it to go. It had been in service since 1989 so it had earned it's burial with honors.
I priced new gas heaters but couldn't bring myself to pay the $400 they wanted for the cheapest 30 gallon model. So I bought a used one advertised in Craigslist. I spent a lot of time flushing it out because I didn't want a lot of rust, etc. It was really quite rusty and I concluded (too late) that I should have spent the money on a new one. But it worked, held water (and air) and is now in service doing well. I do get rust coming into the tub however, but it's a different color rust.
I just changed the water in the tub and I make this note so I'll remember for next time. When the water is drained, the water also drains from the overhead pex plumbing and air is introduced. When the pump is then started at low speed, it can't overcome the air pressure and no water flows to/from the heater. The solution is to get the hot tub to run at high speed. Lots of bubbles come out and it starts flowing again. I noticed this before, but it took me a few worried minutes to remember it this time.
The cheap used water heater didn't last long. The symptom was first that the overhead pex would lose its water and wouldn't pump on low (heating) speed). I lived with that for a couple of days, then the heater did let go with a puddle in the basement. I bit the bullet and spent $400 on a new 40 gallon heater, exactly the same as the house water heater. So far it works great, and with no rusty dust coming into the tub.
Observations from a year or two later. It appears all hot water tank (HWT) heaters give off rust. The rust dust appears in the corners of the tub, so I siphon it out when it builds up.
Also, bubbles appear in the line between the hot water heater and the tub and unless the flow is strong enough, they accumulate to the point of "vapor lock" no flow. My solution was to eliminate the very restrictive electric valve I had put in the line. Apparently the valve was not necessary. The flow increased nicely and the bubbles purge themselves most of the time. As for the source of the bubbles, the latest theory is that it's the result of the air pocket at the top of the HWT and some of this air dissolving into the water. Then as the water moves into the cooler line, the air comes out of solution forming the bubbles.
If I were to start over, I would try a heat exchanger instead of a HWT. The household HWT would supply the heat. It would take up less space than my current setup with two HWTs, would be cheaper to buy, wouldn't have an air pocket to create bubbles, and wouldn't have rust.
I tried a bit more with the second hot water heater tank (HWT), adding an air vent, second pump, and check valves to eliminate the bubbles problem. The result was that the HWT started leaking a lot. It had rusted out from the chlorine and having air inside the top third. I got rid of the failed HWT making the basement a lot less crowded.
It was time for a Heat Exchanger (HEX). I got the idea from a chap who found my web site and wanted to pursue the idea of using a HWT to heat his hot tub. He ended up getting a store bought stainless steel HEX for his system.
I priced them online, found out about sidearm style, which parallels the HWT and sometimes use convection to move the hot water through. I decided to build my own using copper. Based on my struggles with the dedicated (now dead) HWT approach, I had a pair of one way valves and a extra Taco 007 motor. Here is my picture blog of building the HEX.
The auxiliary pump in the hot tub enclosure is still used to send the water through a PEX run into the basement. To run the second pump, located at the HWT, I ran a wire paralleling the PEX and fed by the same hot tub circuit which comes on when its thermometer calls for heat. The one-way valve in this picture was put in earlier to address the bubbles problem and has been removed. There is however a swing type check valve at the bottom of the HWT, to prevent household hot water from being drawn from the bottom of the tank through the HEX.
A neighbor suggested this approach for heat transfer fins. 4 inch lengths of 3/4 copper, slit lengthwise but leaving 1/2 inch uncut to make 4 bendable fins. One slit is cut through so the circle can be stretched to fit over the inside tube. I have them all facing the same direction and soldered in place at even intervals.
I used 5 feet of 1/1/2 inch heavy copper tubing for the outside. The end pieces are 1 1/2 reducers that happen to be bushing style, which means they fit into the tee. If I had used regular non-bushing reducers, I would have had to cut some 1 1/2 nipples from the 5 foot piece. Notice that the bottom of the HEX has a street 90 soldered in instead of having the threaded fitting come straight out. I should have done the top this way too. I had two temperature gauges left over from another project, and I put these into the top and bottom of the hot tub circuit so I could see the temp. differential. Also they look cool.
After running for a few days the new system appears to be working well. Due to heat loss at the hot tub, the system calls for heat about every 1/2 hour and the system runs for about 10 minutes before reaching 104F. It would have been "unfortunate" if the HEX was too small to bring the tub up to temp. But 5 feet of length seems sufficient. Probably larger would be better.
Initially, due to plumbing confusion, I had the flows through the HEX both running the same direction. Published lore says they should be flowing opposite directions, so I changed it. Based on my log of electric energy used to maintain temperature during the night, it only works marginally better flowing opposite. But in sum, the HEX is superior to the HWT approach. There is NO MORE RUST in the tub. I suspect the natural gas usage will be less since I'm not maintaining a second large tank. And space in the basement has been reclaimed. Cost not including pumps was about $200.