Recirculating Hot Water System - Duty Cycle Options

I have a recirculating hot water system with a return line driven by a Grundfos UP10 (96433899) pump. This pump can be set to turn on or off every 20 minutes. The house is a single-story, flat-roofed house on a concrete slab in Palo Alto, CA (read, a mild climate) and the water delivery lines are copper in the slab. The master bath is approximately 90’ from the water heater! Energy was cheap in 1955. The return line is on the roof under 3†of urethane foam insulation. Instant hot water is wonderful.

I am thinking about changing from a duty cycle of 5 AM to 9 AM and 5 PM to 10 PM to 20 minutes on and 20 minutes off from 5 AM to 10 or 11 PM. I believe we will find that the water temperature will not drop significantly in 20 minutes once the encasing slab is warmed and that gas usage to keep the water hot will be about the same. I am concerned that cycling the pump on and off every 20 minutes might trigger a more rapid failure of the pump and/or switch components. The number of on/off cycles each day would be about 24. If this works, we would have access to hot water from dawn to bed time and not need to use our water heater to warm up the slab twice a day.

Thanks

baumgrenze

Instead of just running it in fixed cycles, you could add an aquastat to the system, and then, only turn it on when the temp had dropped enough to be an issue.

In reality, it probably does not need to run for 20-minutes to get the hot water to the end point.

The system I have only runs when the aquastat senses it's cooling, then runs the pump and shuts off. Typically, runs maybe a minute every 15-20 minutes or so. It is also on a timer, so only runs during my typical up times, and shuts off at night.

The problem with an aquastat is that it senses the water temperature at the water heater, NOT at the end of the line where the water would be used. Those two temperatures are very different.

hj said:

The problem with an aquastat is that it senses the water temperature at the water heater, NOT at the end of the line where the water would be used. Those two temperatures are very different.

Click to expand...

That does not need to be the case, but is for most of the prepackaged units. The RedyTemp is a self-contained unit that is designed (at least one model) to where it can be installed under the furthest sink, and senses the water temp right there at the point of use. If you cobble up your own, you can mount an aquastat anywhere - just make sure you run a big enough gauge wire so it can still control the relay. Think of it like wiring up a door bell switch - low-voltage works fine with a low current relay as long as the wire gauge and load are matched to the distance you need to run the wire.

Thank You hj

Thank you hj

hj said:

The problem with an aquastat is that it senses the water temperature at the water heater, NOT at the end of the line where the water would be used. Those two temperatures are very different.

Click to expand...

Mine is a recirculating system. Hot water from the heater is pumped 90' to the bathroom and then 90' back to the water heater. The more I think about it, the more I realize that I could use this solution:

1) Set the pump on its 'run continuously' setting

2) Put a $15 on/off timer switch between it and the 120 V supply

3) Put a simple open-on-rise disc thermostat, open at 120 F, close at 100 F, mounted with good contact on the return hot water line, between the timer and 120 V supply. I've made such a mount from aluminum stock, bored to fit 1/2" copper pipe, split and bored and tapped for screws to clamp it on the pipe and also tapped for screws to mount the thermostat.

Are there cheaper 'aquastats' than the ~$100 Honeywell devices I found searching the web?

Thanks again,

baumgrenze

There are strap-on aquastats already designed to just snap in place. Unless your WH is set higher than normally suggested, if you measure at the tail end of the loop, it may never get to your 120-degrees. Most of the aquastats I've dealt with are only designed for low-voltage control, and can't be used to switch a 120-vac load without a relay or something similar in between.

Some of the prepackaged kits use a thermostatically controlled valve at the point of use(s), and you could still use the pump on a timer, but it won't move water unless one or more of those valves opens. Won't hurt the pump, and since those are typically quite small, the energy use isn't a big deal most of the time. Your bigger loss is what is radiated from the pipes in most cases. Good insulation helps a lot.

Thanks, Jim

Unfortunately I can't do anything about 90' of pipe in the slab. I should send the problem to an engineering prof somewhere and ask if the task of heating up the slab that the pipe passes through twice a day looses less heat than running hot water through it the 4.5 hours between my current morning circulation time and my evening circulation time.

The urethane insulation on the roof can't be doing a spectacular job. We ran some cold water lines over the top to add some garden hose-cocks and on a sunny summer day the water that comes out first is HOT. If heat can get into water that way, it can get out, too.

Grundfos support replied that they think their pump/timer can take 18 on/off cycles a day so I will try to find time to reset the timer. If 20' on and 20' off supplies water in the master bath at 100 F (that is where the Hans Grohe thermostated shower valve is set) then I will try 20' on and 40' off next. As a retired organic chemist I'm always ready to try the empirical method, especially if all it costs is the time needed to reset the timer.

Thanks for your insights,

John

Insulation only slows the transfer of heat...it cannot stop it, so water sitting in a hot attic will get to water pipes, and if they were heated, will leak out if that attic is cold. Slowing it depends on the quality of the insulation, its thickness, and the delta-T of the two. Moving air speeds up the heat transfer.

Not an Attic - Not a Cellar (or Crawl Space)

jadnashua said:

Insulation only slows the transfer of heat...it cannot stop it, so water sitting in a hot attic will get to water pipes, and if they were heated, will leak out if that attic is cold. Slowing it depends on the quality of the insulation, its thickness, and the delta-T of the two. Moving air speeds up the heat transfer.

Click to expand...

In the 1950's a developer in our area hired some architects to mine the work of Frank Lloyd Wright. With a limited number of plans (and reverse plans) in any development, Joe Eichler built some interesting houses. All of them featured hydronic radiant heat (and other utilities) in a concrete slab with a 14" perimeter footing, open-plan post-and-beam construction combined with some walls of conventional wood frame construction. All featured a roof deck of 2x8 or 2x10 tongue-and-groove redwood covered by a built-up (tar-and-gravel) roof, some flat and some slightly pitched. In 2010 we did a significant remodel and ran a lot of utilities on the top of the roof deck. We covered that with 3" of rigid urethane insulation. The hot water return line is set on 2x4 (1.5" high) blocks nailed to the roof deck, so it is more or less in the middle of the insulation. The delivery line for the hot and cold water is either in the concrete slab or just beneath it.

I took a minute and changed to 20' on and 20' off from early AM until latest bedtime, ~5:30 AM - 12:00 midnight. This approach reduces the hours/day that the pump runs by 1 hour. We shall see if the water remains hot enough. With luck maybe 20' each hour will be enough.

Thanks for the interest,

baumgrenze

Works Effectively 20' on and 20' Off - Any Estimate of My Flow Rate?

baumgrenze said:

I took a minute and changed to 20' on and 20' off from early AM until latest bedtime, ~5:30 AM - 12:00 midnight. This approach reduces the hours/day that the pump runs by 1 hour. We shall see if the water remains hot enough. With luck maybe 20' each hour will be enough.

Thanks for the interest,

baumgrenze

Click to expand...

We've been using this setting for a good month and it seems as effective as the previous 'always on' setting was for the chosen periods.

In a brochure, Grundfos says that the flow range is 0 - 3.5 U.S. Gal/min and the Head Range is 1-5 U.S. Ft. They also show a 'Performance Curve' showing flow as a function of 'head.' Understanding the whole concept of 'head' looks like a serious study in hydraulic engineering. The physical difference between the highest point (over the roof deck) in the loop and the lowest point (top of the water heater) is ~ 3 feet, but then there is the concept of the 'friction head' (the resistance to flow in the ~ 90 feet of half-inch copper piping.)

I did do a quick calculation that I could do. In my estimation, 90 feet of half-inch copper piping has a volume of about 1 gallon, so at 1 gal/min the pump would replace the water in the pipe every minute. If I measure the outgoing temperature (water heater feed temperature to the pump) and the returning temperature (circulation loop return temperature) I think I should be able to get a sense of the heat loss, i.e., how hard the water heater is working to provide hot water at the far end of the loop.

I'd put in a surface mount aquastat, but it is now a big project. The delivery end of the loop is at a shower valve. The shower valve is in a wall between the master bath and the master bedroom. I'd need to cut open the wall again to install (and maintain) any aquastat (thermostat) used to control flow. I believe that if I had one, in principle I could use an X-10 communication device to send the on/off signal from a simple disk thermostat mounted on an aluminum block bolted to the pipe and wrapped in insulation to a control relay in the furnace room where the water heater resides. Unfortunately I don't have the 'bandwidth' just now to tackle the project.

I promised to indicate how my changed duty cycle is working. It works.

thanks

baumgrenze

In a closed loop, the falling water helps to 'pull' the water going up. It can get complicated. The head required to supply something like a fountain is quite different than what is required in a circulator of a closed system.