Recirc loop insulation
Space constraints are driving me towards a gas tankless water heater. I’m looking at the Rinnai RSC160in b/c it’s on a big sale, plus I could get rebates, and thinking I would utilize the recirculation feature.
Does anyone have any ideas for affordable ways to boost insulation of the hot water loop? I see the regular foam pipe insulation that’s like R3, but I’d prefer to increase that substantially if possible. My reasoning is, if I’m gonna have recirc, I’d like it to be available pretty regularly, and I don’t like having to push buttons 30 seconds ahead of using water. So, I envision keeping the hot water loop “charged” for a good portion of the day. And to do that I want it to be insulated well enough that standby losses are kept manageable.
Or if this is a bad idea all in all, please let me know why.
Thanks!
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Affordable..
Just brainstorming...
Wrap a batt around it and tape it then wrap with plastic...
Probably not workable, but maybe someone will chime in with a less hokey idea
haha, that's what came to mind for me too! There must be a better way, but it's also nice to know I'm not the only one to have that thought
You can get commercial pipe insulation made from a formed rigid fiberglass with a white outer wrap that is available in higher R values. The stuff is commonly used in commercial applications, in my case usually on chilled water lines in datacenter facilities. You can order it at any commerical mechanical supply house. R5 to R8 should be available.
In my own home, I just used the R4 poly foam stuff from the box store. I did do one thing to help limit losses a little though: I insulated the supply line (a 1" line in my case), and the return line (a 1/2" line) seperately, then I taped them together for the entire length of the run with the return line directly beneath the supply line. This helps to reduce overall thermal losses a LITTLE more than the R3 poly foam pipe insulation would alone, and it's cheap and easy to do. I just used Scotch 88 electrical tape to tape the lines together, being careful not to overly compress the poly foam pipe insulation. This has held up for 8 years now with no issues.
Bill
Great, thanks. I will check into the commercial pipe insulation. And now that you mention it, I guess it makes sense that I could also downsize the return side of the hot water loop for further efficiency (was planning 3/4" for the whole thing but could do 1/2" for return). I also like the idea of keeping those lines together.
Even with an active system using a pump, the flow rate on the return line is low. I don't see any reason you'd need larger than a 1/2" return line. If you're using PEX, you'd likely be fine using a 3/8" return line. The supply line needs to be sized for the number of fixtures and the length of the run, but the return line only needs to handle the flow of the recirculating pump or convective flow, which is small -- and pressure drop there isn't a big deal as long as you can still move enough water to keep the supply line temperature up where you want it.
Use smaller pipe for the return to both save money and reduce energy loss. There is no need for a parity sized return pipe.
Bill
Better ? Sure!
But Inexpensive? Yeah
We've also gotta keep in mind that we might increase heat loss of we go too thick. I forget what that limit is, and I've always wondered if it's borne out in reality and not just a mathematical outcome of 1D Fourier heat transfer
I don't see any way you're going to increase heat loss if the insulation is too thick, assuming you're using the same insulating materials. I've also heard people worried that insulating the lines will stop the circulation in the return loop (with thermosyphon systems), but I've never seen that, either, in my entirely insulated loop. No matter how well you insulate the system, there is going to be some small temperature difference which will cause some convective flow.
The only issue I saw on my own system was air lock in some fittings I used to get around some obstruction. A quick blast of the flush valve into a bucket took care of that.
Bill
Q=UA delta T
U decreases linearly while A increases exponentially
At some point with thicker insulation Q increases with more insulation
yes, I remember being surprised by that in engineering school
That doesn't apply here, at least not in the way you think. Added EXTERNAL surface area of the insulation does not increase heat flow through the insulation. Insulation is not a heat sink. Added INTERNAL surface area (such as using a larger pipe) would. For a given size pipe, more insulation means less heat loss. For a given R value of insulation, larger pipe sizes mean more more heat loss.
Bill
That makes a lot of sense, but it also seems totally arbitrary.
I don't see how the physics of conduction would know or care what's ,'internal' or 'external'
But as in my original mention of this, I do still wonder if it's generally apt.
It's a 1d algorithm in a 4d world...
I wonder if what I was recalling from e-school was this?
https://www.quora.com/Why-does-heat-transfer-increase-initially-with-radius-of-insulation-and-then-decrease-beyond-critical-radius-Expecting-a-conceptual-answer
But regardless, I'm pretty confident in this case adding the thicker insulation is indeed better
Reply to post #13
It's not the internal/external part. My explanation maybe wasn't as clear as I'd hoped. What matters is the total area of the surface through which heat moves. More total square units of area for energy to MOVE THROUGH (i.e. BOTH sides) means more energy can move for a given R value. More R value alone only reduces the amount of heat transfer.
With a flat plate, it's simple conceptually. With a pipe, not so much. More R value only increases the external surface area over the insulation, not the area from which the heat is originating, which in this case is the area of the pipe. Increasing the R value of pipe insulation increases ONLY the outer area of the outer surface, not the area of the inner surface. You're using a material that resists heat flow (insulation), not a material that will "draw heat out" such as an aluminum heatsink would. Hopefully that makes more sense. If we were increasing the area of a heatsink, THEN I would agree that more area would mean more heat flow, but in this case it's an insulating material we're working with.
Maybe there is some limit at which point the total outer surface area becomes significant in comparison with the amount of heat flow the insulation can support, but I don't think we're going to run into that in an any reasonable amount of R value here. I've actually done somewhat related measurements to this with temperature probes for some electronic designs, and I have not seen what you describe with thermal insulators, but I most certainly have with metals.
Bill
Thank you Bill
I have some fundamental conceptual gaps here I think.
I just don't understand the notion, or significance of a 'heat sink'. That is, in my conceptual framework, both metals and foam are insulators with mass, conductivity, transmissivity, etc.
There aren't materials which "draws heat out" and materials that "resist" heat flow. All materials (and least in STC environments) do all of the above.
E.g.: steel is a relatively good insulator... Compared to aluminum.
I can understand if it's a bit too involved to correct my notions in this comment section, but it you happen to have any links that would teach me, I'm interested!!
That's Generally why I can't see why the exterior of the insulation wouldn't count as a heat transfer surface for the purposes of this conversation. I mean, heat IS lost through the surface of the insulation, right? And that is driven by the surface film coefficient multiplied by the surface area.
Assuming (!!!???) that heat flow is roughly equal in all directions around the circumference and within the insulation, the heat loss through the surface of the insulation must be equal to the net flux at all point in between the surface and the exterior of the pipe. (In this 2D framework)
Thus, if we increase the surface area, we increase heat flow.
This is how heat sinks, in the sense that I think you're using that term (e.g. a heat sink hunk of metal with fins for my raspberry pi) work right?
Reply to post #19:
I think what you're thinking about is a sort of limit-like situation with either very thin or (possibly) very very thick materials like this. In normal practice, as you increase the amount of insulation over the pipe, you decrease heat flow. Within the range of reasonable R values, you won't ever see a situation where more insulation results in more heat flow. While I haven't made measurements looking into the specific concept you're talking about, I have made measurements to evaluate insulation effectiveness (i.e. "did we waste money with the extra R value, or are we saving enough to make it worth doing for some design lifetime?"), and I've only ever seen LESS heat flow with MORE insulation.
To use your 'Raspberry Pi' example, think about the die temperature of the chip under the heatsink. I'm not sure if it can measure that or not internally, but if it can, you could test this. For a given amount of heat from the chip, a larger heatsink will REDUCE the die temperature. Now try putting an insulating material in place of the heatsink, such as a piece of polyiso. You now will RAISE the temperature of the chip. Usually this is expressed as 'thermal resistance' given as some number of degrees C per watt.
I suppose another way to look at this might be that insulating materials have so little heat flow compared to thermally conductive materials (aluminum, copper, etc.), that they don't behave the way you'd expect when dealing with reasonable thicknesses.
In quantum physics they want to find a "universal theory of everything", since right now, the world of the very, very small has rules that don't fit the world of normal size stuff and vice versa. I think we're getting into something similar here, where we are so far beyond the critical thicknesses for these materials that different rules apply since the rules you're talking about are effectively swamped by other effects that affect the system much more.
Bill
@bill reply to #22
Yes... My question was academic, but I was fishing for reel (pun intended) experience
And you delivered. Anecdata, for sure, but good enough to test a Centuries old algorithm
@ben87
That link is exactly what I'm thinking about! Thanks for posting
The simplest is not to have your return line hot. There are some recirc systems that have a temperature sensor to shut the pump down once the supply is up to temp.
There are also the thermostatic valves such as Watts sensor valve which you can repurpose to turn off the flow on the recirc line once the supply gets hot.
Overall, the best is to run the pump the least. I have found that a motion sensor trigger in the bath gets you hot water by the time you need it with minimal runtime.
P.S. You can always double wrap the insulation. In case of 1/2 pipe typical insulation, I think 1 1/4" pipe insulation will fit over it.
Thanks. I spoke with a tech from Rinnai who says that the pump will stop once the supply is up to temp & then it will periodically circulate a little bit and retest the temp until the temp drops beneath a certain threshold. Then it will trigger the burner & run the pump in order to bring the whole hot water loop back to temp again. So it sounded like the time the pump would have to run should be kept pretty low, especially if the lines are well insulated.
And good, I will check for some 1 1/4" insulation locally. (Or larger b/c I think I would need a 3/4" hot water line since if I loop it for dedicated recirc it would feed all of the hot water fixtures)
There is no isssues with 1/2" or even 3/8" line. I've no issues with a 1/2" recirc.
In terms of tankless it is good get one with a built in buffer tank (ie NPE-A2) or install a small one downstream. With recirc setup, the buffer tank doesn't need to be powered as it will get refilled every time the recirc runs. This will prevent delays and cold water sandwich.
Oh interesting. I was thinking I would need 3/4". Do you have a rough estimate of how many gpm 1/2" would be good for? Looks like the tankless heater can do 6.1gpm at a 50 deg rise, so as long as it's above that, should be OK. Guessing it would be a total length of 60ish feet with several corners.
And yes, I would prefer the Naviens for that reason. Problem is they are like double the cost. So I've also considered just putting a small electric tank like eemax on the outlet of the rinnai and not doing recirc. But is cold water sandwiching a problem when recirc is active? My thinking is that it would keep the water in (and just before) the unit hot, so by the time fresh cold water actually comes through the gas has been fired up and begun heating.
1/2" is going to handle plenty of flow for a recirc system. Remember that those are always pretty low flow by design, because you only want enough flow to keep the main line hot. Excess flow here justs wastes energy.
Make sure to put a check valve on that return line so that you draw water only from the main line at the fixtures. I usually put the check valve near the source of the hot water, so the "near end" of that return line relative to the heater. I like to put a valve at the "far end", right near where the return line taps off the main line, which allows you to isolate the return line if you ever need to do maintenance.
Bill
The 1/2" is for the return line from your last fixture. For a typical trunk and branch setup you want a 3/4" trunk to handle the flow from all the fixtures especially if running multiple showers off it.
The recirc can work as a kindof input buffer but you need to tie the recirc feed in well away from the tankless. Essentially the volume in this cold water supply becomes a mini buffer tank. The two issues with it is that it is a very small volume, so it might cover a startup delay but once it is cold, it doesn't work anymore. It also won't work for cold water sandwitch.
The buffer tank on the output can be anything with 3/4" fittings. A small 6 gallon tank will work just fine.
Can I convince you to change your plans?
The gas tankless has a few down side in the it is the least efficient in terms of the percentage of BTUs it gets into the water. It will mix the most conditioned air from your home with the fuel and expel it from the home. That air will be replaced with unconditioned air you will pay to condition. Tankless gas is the most complex and maintenance intensive water heating appliance on the market. If your fuel is propane, it likely to cost more to operate than electric would. If your water is hard make sure you understand how to delime the heater.
I have a recirculating loop on my electric tankless it only circulates on demand not 24-7 and not on a timer. I turn on any hot water faucet for a few seconds and then back off. That flow turns on the pump only if the loop is cold. My loop, pump and check valve compo takes about 2 minuets to heat the full loop. The pumps intended for recirculation are small about 1/8 horse power so the flow in gallons per minute is pretty small and the check valve can be a restriction.
If I were redoing the system today, I think I would go with a heat pump water heater.
Walta
The main thing pushing me towards gas tankless is the space savings. Need to fit HVAC, water heater, and laundry in some tight closet spaces. But your point about complexity and maintenance is well taken, and I will weigh it. If I could put a heat-pump unit upstairs, it would be a good spot to add cooling, and the supplemental dehumidification would be nice.
But besides the space savings for a gas unit, I'm also on city natural gas, which is quite cheap. I think about 1/5th the price per BTU compared to electric. And I would get a condensing unit b/c I don't want any possibility of backdrafting. So I think that means the loss of conditioned air would be a non-issue.
My local water is soft and I had good luck with tankless units. Very efficient and do take up a lot less space.
If you are tight on space, you can look at a slim ducted unit or a multi position air handler (ie Mitsubishi PEAD or SVZ). These can be mounted on the ceiling which would free up the space under it.
P.S. Are are also some hydro coils systems that can use your tankless as a heat source for space heat. Would still need an A coil for cooling as well.
You can get gas tankless water heaters with "sealed combustion" setups that use outside air for combustion air. There is no "sucking out conditioned air" issue with these.
Bill
If you need a recirc line and don't want to push a button, then simply wire the recirc pump to some motion sensors. As soon as anyone enters any bathroom, gets near the kitchen sink or near any other device that needs hot water (laundry sink) the motion sensor triggers the recirc pump to run for 30 seconds or 1 min, and set it up to only run max of once every 10 min. Easy and no need for manual intervention.
Do you have recommendations for the specific sensors and switches to set this up?
I would add a feature to this that the recirc pump have a built-in thermostat like a lot of them do (Grundfos definitely does). Once the sensor determines the water in the line is hot, the pump will stop running. No need to limit it on a timer.
john, do you plan on using any home automation system? I would send the motion sensor alert to a smart hub and make a rule that kicks on a smart outlet where the recirc pump is plugged into. When motion is detected in that bathroom/kitchen, the pump is energized and if the temp sensor says it's cold it will then circulate the water until the temp sensor says it's hot enough.