Expansion tank location with multiple buildings

GroundUp

I'd like to hear your thoughts on the subject. My own home, I have the house and 2 outbuildings. The boiler is in one of the outbuildings and I am heating all 3 via insulated underground lines to the other 2, with individual circs in each building. The boiler is always hot, and any call for heat in any building starts the respective circ and heats the space, then sits idle again until the next heat call. When I put this together, I installed a single #30 expansion tank at the boiler thinking any load circ would be "pumping away" from it albeit 200 feet away.

I was having a discussion with a colleague yesterday who is doing the same type of system with the boiler in an outbuilding feeding two houses, but was going to mount the expansion tank in house 1 with that load circ pulling away from it, as you would with any system if the boiler were right there. Then house 2 would still have its own circ, but he was relying on the expansion tank in house 1, 250 feet away that may or may not be circulating at the same time. We debated adding another tank in house 2 so each loop would operate independently but then there would still be nothing at the boiler.

What's the proper channel here? I feel like multiple tanks in the same system might screw with each other, but at the same time have my doubts about the tank at the boiler with all the underground piping between. Moving the circs to the boiler is not an option, there is no space nor is there a conduit in place for signal wire.

ch4man

as you know pumping away with the circ right after the x-tank all the pressure differential is added to the outlet with no drop in the inlet to the circ.
in your case the delta P the distance between the circ and the x-tank even though its still "pumping away will alter the ratio of pressure rise and drop at the circ even though the delta P remains the same.
as to multiple x-tanks remotely installed in the system its too early for my brain to process but...
i was at one install where it used a compression tank (old school tank in the ceiling) and one CI rad was full of air (making is a kind of sort of compression tank ) at just the right distance where air and water would surge and move between both at times based on the circ on/off and the current boiler temps. the "point of no pressure change would fight each other.
very weird.........

Jamie Hall

When you really come down to it, @GroundUp , what we are trying to do is to make sure that the pressure everywhere in the system is at least adequate. Generally -- in most "normal" setups, it's easy, since the lowest pressure when the system is operating will be at the inlet to the circulator. So... put the tank there, and pump away. Voila.

In your particular situation, the lowest pressure will still be at the inlet to your circulator, 200 feet away. Your's works. It would be useful to figure the head loss in that 200 feet of pipe, and subtract that from the tank pressure, more for curiousity than anything (it's working, don't fix it!).

Now 1250 feet of pipe is another story. I would be very much inclined to put the tank at the boiler, like yours, but honestly if it were at al possible I'd put both circulators there, too, as a first thought. However, the only really reliable way to do this -- and to size the circulators and determine what pressure to set for the tank -- is going to be to construct a head loss diagram for the system (I see things graphically, and there are other ways) to see what the pressures will be throughout the system under four different conditions: nothing pumping, circulator 1 pumping, circulator 2 pumping, and both circulators pumping.

I'm a little concerned, too, that with adequate pressure at the inlets to the pumps the pressures at the outlets may be rather high when operating; you need to check that. Another subtle effect which I have seen on water supply systems is that with that long a pipe the transient pressure at outlet when the pump starts may be very high -- that column of water won't start moving instantly! -- or the pressure at the inlet at the same time may be too low (same reason) (on water supply systems controlled by pressure with the switch near the pump but the expansion/control tank a distance away, this shows up as a rapid sequence of the pump starting, pressure spiking, pump stopping, pressure falling, pump starting... etc. which is really hard on the pump). Another reason to put the circulators and the expansion tank close together at the boiler.

If you can.

Zman

Are there any issues with your existing system?
The issue you may have now would be low pressure at the inlet of the remote circs. This may or may not be an issue depending on the pipe and circ sizing.
When you add remote expansion tanks the pressures around the system kind of average themselves out. This may or may not be a good thing. I would not recommend remote expansion tanks unless there are issues with your existing setup.
For your buddy with the new system, locate all the circs at the main boiler location and run control wires in conduit.

GroundUp

Thanks guys. The system has been working great, at this point in time both the house and garage are kind of temporary systems as I just piped the boiler in November when it was already cold out and just threw together some mishmash of pex to get the buildings warm so nothing is very professional at this point. I only have one pressure gauge in the system, on the outlet of the house circ, which will rise to 20 psi on circulation with 15 psi static- it does not cavitate in the least and I'm cool with that for right now but when it warms up again and I pipe the system correctly, I will add gauges everywhere. I did screw a gauge on my XT drain valve and it holds right at 13psi whether the pump is running or not (house is lower than the boiler) so I think it's doing what it should. My house loop is currently 340ft of 1" ID pex and 80ft of 1" nominal pex with 8 elbows, a 30 plate 5x12 HX for DHW, and a 16x20 W2A HX in the furnace plenum (converted from atmospheric OWB system, yes it's all o2 barrier) with a Grundfos 26-64 in the basement. My math shows roughly 6 GPM at 20ft of head with 150 degree PG at 30% and judging by my heat loss calc and delta across the HX that's about spot on. The garage loop is 200ft of 25mm pex (11/16" ID) and 7 feet of 3/4" nominal pex with a Grundy 15-58 and single surface mount CUH at this point, calculation says approximately 4 GPM- this will be changed around and flowmeter added when it warms up and I add the radiant floor/ceiling to the system.

@Zman as I mentioned earlier, circs at the boiler is not an option. These are both atmospheric outdoor wood boilers with plate exchangers in the rear cabinet to close and pressurize the rest of the system. The cabinets have very limited space and cannot fit 4 pumps along with all piping, HX, XT, controls, etc. Pumps will be in each building no matter what.

Jamie Hall

Can't put circs at the boiler? OK, not really a problem -- but do the system pressure diagrams I suggested. Find out what your worst case pump inlet pressure is (that's really what you are worried about) and set the system static pressure so that that inlet pressure is within the specifications for that pump under the worst conditions. If you have significant elevation changes, include those in. Yes, it is complicated a little. It's an interesting exercise (at least, that's what I always told my students!).

GroundUp

Jamie Hall said:

Can't put circs at the boiler? OK, not really a problem -- but do the system pressure diagrams I suggested. Find out what your worst case pump inlet pressure is (that's really what you are worried about) and set the system static pressure so that that inlet pressure is within the specifications for that pump under the worst conditions. If you have significant elevation changes, include those in. Yes, it is complicated a little. It's an interesting exercise (at least, that's what I always told my students!).

I'm not sure what you're suggesting, with the pressure diagram. The circ has no inlet pressure "spec" according to Grundfos. Static is 13-15psi everywhere in the system, and as the temporary setup now is operating, there is no cavitiation which tells me inlet pressure is sufficient but when it gets piped correctly come spring, I will be adding gauges everywhere to monitor pressures. Come to think of it, I believe I have a plugged ball valve on the bottom of the SuperVent on the inlet side of the 26-64 that I could screw a gauge into. Maybe I'll do that and gather a little more info.

pecmsg

Can you post a diagram of the piping?

Im having a hard time picturing it.

hot_rod

Basically you have a district system, common in cities in Europe and Scandinavian countries.

Grundfos UP series have a min. inlet pressure
1.3 psi or 3'at 140°
4 psi or 9' at 190°
15.6 psi or 36' at 230°

So the problem with unpressurized OWF circulators becomes obvious especially at high operating temperatures. Taco has or had a 3 piece, I think non ferrous, circ better suited for that application, that is the best option.

Good that you isolated with HX at un-pressurized boiler, but probably still running the boiler circ out of spec, unless the water level is about 20' above the boiler Knowing that OWF sometimes boil

As Jamie noted the PONPC will be somewhere out in the piping for the remote circs, it would take some number crunching to determine that point.

I would not use expansion tanks in multiple locations.

Jamie Hall

That's really useful, @hot_rod ! Are those gauge or absolute? I'm thinking absolute, but it would be helpful to me to know...

And @GroundUp … I'm terrible at describing these things. Perhaps just thinking about numbers would be better. What I would do is start somewhere in the system -- the easiest is the outlet from the pump -- and assign an arbitrary pressure. Let's say 15 psig. Then follow along the water flow in the pipe to the next interesting point -- and, at the flow you're interested in, figure the head loss and subtract that from your start. Then to the next -- such as it might be a manifold or something. And keep going that way with the flow and pressure drop until you get back to the inlet to the pump. Then you have three numbers at the pump -- the inlet pressure and the outlet pressure and the flow. The difference in pressure is what the pump has to add to keep the flow going, and the flow is just that.

If there is a place where the water can go two ways such as a T, remember that what goes into the T has to be same as what comes out -- so be sure that you have the correct flow rates in each pipe! Loops get very interesting (such as where you have a flow splitting out there somewhere and then coming back together -- for what we are doing here just ignore one side of that).

Now you can raise or lower that arbitrary pressure so that it matches reality. Say, for instance, that you have an expansion tank set at 20 psig at the inlet. Your diagram came up with let's say 5 psig. Just add the difference -- 15 psi -- to all your pressure points around the loop and you will have the operating pressure at each point.

Or you can look at Bob's numbers above, and just make sure that your pump inlet has the minimum pressure it needs.

Confused yet?

pecmsg

This sounds lime a primary only.

Not a good plan with multiple buildings. Primary loop from the boiler looped past each building ther secondary to each building.

hot_rod

That pressure info is, or was in Grundfos installation manuals. Still is in the online manual I found at Supplyhouse.com.
Tech info is often easier to find at SH compared to manufacturers sites😳. Especially if you stumble in to the GRUNDFOS international website.

I agree with Jamie, gauge up the circs. If you are determined to know actual as built operating conditions
A few brands of ISO flanges now have 1/4” gauge ports, finally!

Or invest in a top quality gauge and and needle then stick PT ports all over the system. Engineers love PT ports. They allow easy access to pressure and temperature measurements. ISO valves with PT ports would be sweet, eliminate gauge errors especially with low variances that you will be dealing with.

I’d use gauge port valves on all circs if they are only a few bucks more.

Larger circulator have gauge ports in the volutes, often times.

jumper

A big enough tank located high enough above any pump provides water at sufficient pressure for pumps. Modern diaphragm tanks are a poor idea.

hot_rod

IF it were piped primary secondary, a big loop, which was the original intent of P/S, pipe sized correctly, then every set of closely spaced tees becomes a PONPC. Of course the entire loop would need to circulate to put heat in or take heat out.

Zman

If desired, you could use delta p circs at the boiler and zone valves in the remote buildings. The circ will detect the valves opening and respond accordingly.

GroundUp

Zman said:

If desired, you could use delta p circs at the boiler and zone valves in the remote buildings. The circ will detect the valves opening and respond accordingly.

This would not work, I have a mixing valve in the garage that needs to be pulled through. And again there is no space for another circ in the boiler cabinet

GroundUp

This is a rough sketch of what is currently in place, for those needing visual aid

hot_rod

This would give you a true P/S, establish the PONPC for all connections. There will be some temperature blending, can one of the secondaries get by with lower SWT?

GroundUp

I contemplated doing exactly what you have there, but the reality was that there was absolutely no space to add another circ for the primary loop. And while I could get away with lower SWT in each building, the whole reason for piping the system this way was to eliminate all underground circulation possible and the higher the SWT, the lower the underground circ time so I was not okay with lowering it.

hot_rod

GroundUp said:

I contemplated doing exactly what you have there, but the reality was that there was absolutely no space to add another circ for the primary loop. And while I could get away with lower SWT in each building, the whole reason for piping the system this way was to eliminate all underground circulation possible and the higher the SWT, the lower the underground circ time so I was not okay with lowering it.

The primary loop is just in and out of the plate hx, the only time the underground needs to flow is when there is a heat call somewhere out in the district. That doesn't change from the way you have it piped now, does it?

The primary circulator, the boiler circulator, and the zone circulator all all need to run to move the heat around with a series loop primary piping.

Doesn't matter I suppose if you cannot fit it into the enclosure.

Get some matching metal and extend the enclosure

Or stick with you you have. Throw in a handful of PT plugs and you can measure pressure and temperature at the various point to see how the pump ∆P is showing up with the single expansion tank location.

If you believe the heat loss numbers from the insulated piping folks, doesn't cost much wood to keep an underground loop hot.

GroundUp

hot_rod said:

GroundUp said:

I contemplated doing exactly what you have there, but the reality was that there was absolutely no space to add another circ for the primary loop. And while I could get away with lower SWT in each building, the whole reason for piping the system this way was to eliminate all underground circulation possible and the higher the SWT, the lower the underground circ time so I was not okay with lowering it.

The primary loop is just in and out of the plate hx, the only time the underground needs to flow is when there is a heat call somewhere out in the district. That doesn't change from the way you have it piped now, does it?

The primary circulator, the boiler circulator, and the zone circulator all all need to run to move the heat around with a series loop primary piping.

Doesn't matter I suppose if you cannot fit it into the enclosure.

Get some matching metal and extend the enclosure

Or stick with you you have. Throw in a handful of PT plugs and you can measure pressure and temperature at the various point to see how the pump ∆P is showing up with the single expansion tank location.

If you believe the heat loss numbers from the insulated piping folks, doesn't cost much wood to keep an underground loop hot.

No, flow under heat call would be the same but the way it is now I do not need a primary circ, and I do not have space to add one, and I am not willing to hack up my $10,000 boiler to add space. I'm honestly not overly concerned with the delta P, as long as the circs are not cavitating, it is "fine" for my liking but the reason for this thread was to see what you guys thought of the situation, if maybe there was a correct answer that everybody besides me already knew.

In my experience, the heat loss numbers from the underground line folks are actually quite accurate BUT, even at 1/2 a degree per 100ft at 180* and 5 GPM like they claim, my Logstor lines to the house are pumping approximately 10,000 BTU per hour into the surrounding dirt. If I can cut that to 12 hours of circulation instead of 24, it adds up to roughly 2 cord of wood every winter. Considering my worst winter to date was 10 cord, that's at least a 20% savings. Also the 25mm Thermopex I have to my garage has a considerably higher heat loss due to lesser insulation (I needed the added flexibility to get under my shed foundation where the OWB is located) and according to my FLIR, the loss in that run is actually comparable to the 17k building load at design temp so there's another 3-4 cord just lost in the dirt for 12 hours a day extra circulation. This garage was new this year, so I have never heated it in the past and have nothing to compare it to but the ground loss would be quite substantial on 24/7 circ.