GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Hydronic Heating System Hydraulic Buffer Tank vs Hydraulic Seperator

Hotsauce1183 | Posted in Mechanicals on

This question might be more appropriate in a plumbing forum but I have had such great and thoughtful responses in the GBA forums I look to here first. This is also a hydronic heating mainstream topic which should be appropriate for GBA.

One of the primary assumptions in a well balanced hydronic heating system is that not all heating zones will ever call for heat at the same time. Based on that assumption the system designer can select smaller and more efficient components. However there will still be circumstances when the heating load can conceivably exceed the supply capability of the boiler. I see a lot of designers who deal with this issue by specifying primary and secondary loops. This design approach requires multiple (expensive) circulating pumps and increases the complexity of the design by a substantial amount. An alternative design approach would be to employ either a hydronic Buffer Tank or a purpose designed hydraulic Separator (such as Caleffi or Taco for example). The Buffer Tanks start in the $600 range for a smaller but well insulated tank while the Separators are about half that amount.

So does anybody have and thoughts or experience on dealing with this choice?

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. Jon_R | | #1

    Incorrect - under many conditions, all zones will call for heat. But so what?

    If heating load exceeds the supply capability of the boiler (or the radiators), the building temp will drop. Fairly slowly, depending on thermal mass.

  2. charlie_sullivan | | #2

    Edit: I answered the wrong question originally. Somehow I read the question as a concern about supply exceeding demand with only one zone on. That's a legitimate reason to use a buffer tank, to avoid short cycling, but it's not the topic of the question.

    But the real question is about demand exceeding supply. As Jon points out, that's not a problem.

    Also, neither a hydraulic separator nor a primary/secondary loop configuration addresses that issue. Rather, they are intended to ensure that the flow in one zone isn't affected by whether or not another zone is on. I think the best way to address that problem is to use a constant-pressure ECM pump and zone valves. The pump energy consumption can be much less.

    A buffer tank, could, for a very short time, allow demand exceeding supply. But that's not really any better than allowing the temperature of the water to droop a bit. In most cases, only the temperature of the water droops, not the temperature of the house.

  3. Hotsauce1183 | | #3 - Thanks Charlie. You are quite right and you nailed the other side of the equation: There is demand exceeding supply and, Nailed that one (thanks much Jon), and there is supply exceeding demand which causes short cycling. It seems to me that a buffer tank would address both problems nicely albeit a bit expensively (That begs the question of how many years payback?).

    But I guess the real question boils down to: "Would a hydraulic separator, at half the cost, perform nearly as well as the buffer tank"?

  4. Jon_R | | #4

    You can add a partial bypass to each zone valve such that the zone valve isn't on/off, it's high/low flow. This addresses the supply (min fire) exceeding demand (eg, just one zone) issue.

    Hydraulic separation provides independence for the flow rates but doesn't solve either issue.

  5. user-2890856 | | #5

    Closely spaced Tees , Hydraulic separator or buffer are all ways to employ P/S piping . Only a buffer can address short cycling due to load being lower than minimum firing rate of a boiler . Properly sized and smart controls (not expensive , smart) can eliminate all or most of your short cycling issues and save loads of money . Both by allowing the boiler to run for a minimum amount of time and also by guaranteeing the lowest return water temps for the largest % of the heating season .

  6. Expert Member
    Dana Dorsett | | #6

    Hydraulic separation (buffered or not) is about managing differing flow rate requirements between what the boiler needs for efficient and safe operation and the flow the radiation needs to deliver the heat.

    There are a growing number of fire-tube boilers out there with sufficiently high turn down ratios that only ridiculously micro-zoned & under-radiated systems would need to be buffered as long as the boiler isn't ridiculously oversized for the design load. Many will work just fine pumped direct in typical systems, without needing hydraulic separation of any type.

    There are no one-size-fits-all answers, you have to do the math. (Hydronic design by web-forum can be fraught with pitfalls.)

  7. user-2890856 | | #7

    In many cases a Buffer tank is designed into and installed in a system for the sole purpose of managing RWT and to prevent short cycling of a boiler . Closely spaced tees and hydraulic separators cannot perform this function .

    Zoning is necessary in all houses due to SHG and other factors such as wind direction and infiltration . Comfort leads to efficiency , design for comfort first and well . Even the lowest TDR boilers will have a certain amount of short cycling . In the old days everyone thought mass was to be found in the big rock under your feet . Turns out mass is better off leveraged near the boiler and conduction is king .

  8. Hotsauce1183 | | #8

    Much thanks to all those who took the time to respond to this topic. Great conversation with much food for thought and all replies were greatly appreciated.

    To Jon I would ask, regarding having only one zone: how does that style of design reconcile with the objective of not heating or cooling an entire house regardless whether or not a room is in use? I ask that because I worry that my design takes the pendulum too far in the other direction... I have a zone for every room employing the philosophy that some rooms like a guest room are seldom used, other rooms like bedrooms are mostly only used at night and at a much lower temp, and only a few rooms are used during the main part of the day at higher temps.

  9. user-2890856 | | #9

    If I may Roger . Zoning rules to be followed regardless of what type of hydronic heating .

    1. Rooms on the same zone MUST have similar BTU per sf requirements

    2 They should have a similar use pattern . Bathrooms and bedrooms should not be on same zones unless rule number 1 is met . Wet , naked people need warmer feeling .

    3. Rooms that will experience solar heat gain should never be on the same zone as rooms on the North and West of the home . During the day time you will usually have no more than 60% of zones calling . Use that guide to size a buffer and use the lowest BTUh requirement as opposed to design . Buffer tanks really do work and most make the mistake of sizing them to design loads which can still leave your boiler short cycling for 90% of the heating season .

    Couple questions . What type hydronic system is this / baseboard , panel rads , radiant floors / ceiling ?
    was a room by room heat loss calc performed ?
    What type boiler are you planning on using and how will you make hot water ?

    Your philosophy is correct although many don't see the value in that .

    A buffer has no ROI but it will keep your boiler in good shape for many more years than without a buffer . Buy the buffer , maintain the boiler properly and it will last much longer than many believe it will . Replacing the boiler years before you must due to abuse (short cycles) costs much more than a buffer tank

  10. Jon_R | | #10

    IMO, zoning is a plus, but is overrated and overdone. And I should know - my house (and many that I know of) is all one thermostat zone (but manually adjusted per room).

    Solar heat gain can overheat a room if you don't leave the door open - without or without zoning.

  11. Hotsauce1183 | | #11

    Thanks Richard - great thoughts. I have attached (I hope) a PDF of the hydronics plumbing plan. There you can see that this is a mixed system with one manifold supporting a number of radiators and baseboards and two others basically splitting the day use rooms from the night-time use rooms. The reason for the radiators and baseboards in some rooms is that this is an older house and some of the rooms only have about a foot of crawl space so no underfloor access. Also, for some rooms raising the floor an inch or two for an above sub-floor installation was just not practical.

    We did a Manual-J using LoopCAD. The results were very satisfying as they were directly in-line with experienced pro recommendations from friends with experience in the heating/cooling field. This is high mountain semi-arid desert in heating zone 5. The primary day-use rooms are on the southern exposure and experience very high UV heat gain year round. When the winter weather finally clears I am planning on new windows and new polyiso sheeting under new siding to establish a much better and more air tight WRB than we get with the 15# felt that they used back in the day.

    We are using a Navien NCB-180e Combi Boiler. Its a compromise in that its a little too big for heating and potentially a little too small for DHW. "too small for DHW" was based on the January observation of 42F incoming supply water. but now that we are starting to warm up the supply water has already gone up 10 degrees which makes me think the supply pipe from the well isnt buried as deep as it should be.

    I might also add that the house has an existing, 5 year old, A.O.Smith top-of-the-line electric water heater. It is so well insulated that I am considering utilizing it as the buffer tank for the hydronic system. The primary downside that I can see is its 80 gallon capacity which is much larger than I would reasonably size a buffer tank.Any thoughts on this idea?

  12. Jon_R | | #12

    IMO, a two pipe connection to the buffer tank offers a small efficiency improvement. It tends to send the hottest possible water to the radiators (perhaps allowing a slightly lower reset curve) and sends the coldest possible water to the boiler (increasing condensing).

  13. Hotsauce1183 | | #13

    Thanks Jon - I agree with you that a four pipe system is obviously going to encourage more effective stratification. I did read Siegenthaler's article, which I found very informative and educational. Thanks for the link so others might benefit from it. In John's article my system would probably run closest to his Figure 4. My tank is close to a thousand dollar tank and would certainly hold the water effectively. The question that remains is, not knowing the exact internal plumbing there is no way to predict how well is would stratify. Do you see anyother problems I might be overlooking?

  14. user-2890856 | | #14

    Beg to differ . a 4 pipe buffer always sees all the flow from the boiler and the system , all the time . It does not , cannot stratify as effectively as a 2 pipe which will never see the full flows . During full load conditions the buffer will see it's lowest flows through the middle portion of the tank as SW moves straight through the top and RW flows straight across the bottom . With a 2 pipe buffer the tank sees little or no flow during these times allowing the layering in the buffer . A 4 pipe buffer during low load times will mix the tank since the system side needs little . The boiler circ draws that excess down to the return of the tank and right back to the boiler . Your boiler will first see low delta then shut off , the tank stratifies less in a 4 pipe configuration . Please see IDronics 17 , you can register through he following Link . And I assure you Siggy agrees with this .

  15. SwitchgrassFarmer | | #15

    I believe we experienced a similar phenomena to what Richard is talking about in post 14 when we first started up our hydronic system with 4 pipe buffer tank,

    Below is the actual email exchange that I sent our contractor. (BTW, we never did change out the supply side pump, increasing the Tekmar 261 Boiler Controller differential setting was an adequate fix.)

    On 12/16/2014 05:53 AM, Andy Bater wrote:

    Have watched the geo unit pretty closely I believe I understand why it is short cycling. Looks like we may have too big a pump between compressor and buffer tank.

    What appears to be happening is the buffer tank is stratifying due to low load requirements for the two primary zones. Recall that the mix valves on each zone control how much water gets pumped from the buffer tank versus how much is recirculated through the slabs. During low load requirement times very little circulation occurs through the buffer tank and the level of hot water slowly moves up the tank. Once the geo/buffer tank pump comes on that warm water is suddenly pushed past the (temperature probe) well in the center of the buffer tank, the temp quickly rises, and the Tekmar trips out. Moments later the Tekmar calls for heat again, and once the anti short cycle timer on the Climatemaster expires the compressor comes back on.

    BTW, I have tried all three speeds on the existing pump, with the fastest setting actually limiting the short cycling problem the best. You can actually see the temp on the buffer tank sensor quickly rise and fall before it goes into normal operation. Sometimes that quick blast of warm water gets over with before it hits the Tekmar target set point, sometimes not.

    As you recall we scaled the compressor and buffer tank to the actual calculated load of the home. However we didn't scale down from John Siegenthaler's specified 26-99 pump (0-33 GPM) between the two units. A five ton Climatemaster TMW requires 15 GPM full tilt, three ton unit only 9 GPM. When you are out here next time let's take some measurements of the piping, review the pressure drop spec through the geo unit, etc, so we can do a head calculation to determine if we can downsize to a smaller pump. I think that might be a bolt for bolt and nominal cost ($100 - $200) change.

    On Tue, Dec 16, 2014 at 10:28 AM, Andy Bater wrote:

    As usual after I wrote this email a possible temporary fix dawned on me. The Tekmar's differential setting was at 10 degrees, a +/- cut in and out of 5 degrees. Raised it to 15 degrees, and it can go higher. Think we can get past the temporary warm water slug temp rise that way.


  16. GBA Editor
    Martin Holladay | | #16

    Thanks for your comments.

    Whenever I see a long Q&A thread on hydronic design issues, or on problems related to the oversizing of pumps on ground-source heat pump systems, I think, "Here's another example of why heating appliances like ductless minisplits -- systems that don't require engineering -- make so much sense."

    -- Martin Holladay

  17. Hotsauce1183 | | #17

    Thanks Martin - Wow, I have been wondering about that myself. My hydronic system design has gotten more and more complicated, And expensive, as I worked on it. Originally I had hoped and expected that my hydronic system wouldnt cost as much as a mini-split but now they are about equal. So now, I have to decide if I want to spend that $10k or $15k (materials only) on just heating (hydronic) or on a system that will provide both heating and cooling (mini-split). My original goal was to NOT have any forced air circulation which made dismissing the Mini-split fairly easy but now I am just not so sure.

  18. Expert Member
    Dana Dorsett | | #18

    I've seen recent existence proofs that ~3000' sub-code-min houses can be comfortably heated in a US zone 5A climate with less than $15K (all in, installed, pre-subsidy) of ductless heat pump.

    At the high cost of hydronic GSHP it can even pencil out favorably to improve the building envelope, install ductless, and install rooftop PV sufficient to cover the annual power use (all of it!) for the ductless systems. This is becoming more true year on year as the cost of PV.

    In high-R houses with a lot of doored off space ducted mini-splits can sometimes be a better solution, even though they are somewhat less efficient (buy another PV panel or two!). If you're thinking of ducted mini-splits as being akin to the ridiculously oversized scorched-air gas burners with comparably oversized split AC, don't. The modulating and small air handlers on ducted minisplits have nowhere near the noise & wind-chill of oversized air handlers, and are usually quieter than a refrigerator in the next room. There's none of that hot-blast followed by a chill going on.

  19. user-2890856 | | #19

    You could also think about your hydronic system without the need for a buffer and boiler and just purchase a very good Tank type mod con water heater . Built in mass , a couple circs , an ODR mixing valve and a FPHX . It's that simple actually , your hydronic system need not be so expensive . I would have brought it up long ago but that was not your question .

  20. user-2890856 | | #20

    Entirely too much complexity in your drawing . Could you describe what exactly it is that you wanted to do and why ? Maybe I can clarify a few things and help in bringing costs down

  21. Jon_R | | #21

    As Dana says, previous bad experience with oversized ducted systems doesn't apply to properly sized, variable speed systems in an energy efficient house - even if it's still gas fired with conventional central AC.

  22. Hotsauce1183 | | #22

    Thanks to all who are contributing to this discussion.

    Dana - I think you nailed it when you mentioned the unfair comparison between the scorched air burners and mini-splits. However when I first started this project it was always mini-split versus hydronics. I didnt want the draftiness of the forced air systems which I had enough of early in adult life in California tract homes. Also, not long ago I had a fine 6 zone mini-split in a two story 2900 sq ft home in Washington that I loved. But in Southern Washington Columbia river produced hydro-electricity was much more cost effective than the price of Electricity here in Northern Nevada. Thus the annual cost of operation would become a major concern for my location now. Also, the need for serious air conditioning is somewhat limited here as at this elevation it cools off at night down into the 40's and 50's the vast majority of the time.

    Richard - I did think of getting a mod-con water heater early on but I just couldnt find a product that I liked. Perhaps I overlooked some good brands because they didnt google up as prominently as the ones I did see? My basic current goal is to produce a system with two temp zones: one at 130 F for radiators and baseboard rooms and one at 115 F for in-floor heating rooms. I am also dividing the day use rooms from the night time use rooms with thermostats and zone valves with many of the zones being ganged to a single thermostat. So on the diagram (see post 11) the manifold on the left is the 130 F distribution and the two manifolds on the right are the 115 F distribution.

    Jon - I couldnt agree more. When I first started this project I said to heck with doing it myself. I am retired and I am just going to hire it done so got several quotes from local HVAC dealers. The numbers and the sizes they were quoting didnt even come close to passing the giggle test so then I decided that I would have to do the engineering myself. I have a bachelors in physics and a masters in CS so the concepts and tools are second nature to me. I also own a number of rental properties where I always did most of the maintenance, upgrades, and repairs myself so this really isnt my first rodeo. I used HeatCAD to do the first Manual-J and then found out that LoopCAD would have done the same thing AND will also produce the hydronic circuit layouts. As a learning curve tool I also used two other HVAC CAD programs to model the space. It was very gratifying to see that all 3 programs came within ~%5 of each other in modeling the room-by-room loads... The total of which was about half of what my early quotes were for. Thats when I moved on to doing the Taco HSScad that produced the plumbing layout posted in post 11 above.

    The bottom line is that I am very confident in the sizing and heating/cooling load modeling for this house. What remains is the implementation and I truly look forward to seeing what Richard comes up with.

  23. user-2890856 | | #23

    Forward your LoopCad heat loss and everything else that program spit out to

    [email protected]

    or you could just post it here for others to see also . Is this in floor heat embedded or some other install method ? What type of rads , CI or panel , what type baseboard ?

    Sorry that our HVAC trade , hydronic and air side has so few qualified folks , it is frustrating . Let's
    see if we cannot get you into exactly what you want . Unfortunately we cannot train everyone , especially thlse who think they already know how and those that do attend training have no idea to take notes , read them for a couple hours a week for a month and the effort was wasted . They think just being there and being able to tell people they were qualifies them .

  24. Hotsauce1183 | | #24

    Thanks Richard. Much appreciated.

    Answers to your questions...
    1. Floor heat for most rooms will be above the subfloor using omega heat transfer plates with plywood spacers.
    2. Radiators are Myson T6 IVC Panel radiators
    3. Baseboards are Slant/Fin BL-2000 although I am considering upgrading to Smith Heating Edge High(er) Capacity. The Smith products are 25% more expensive for 50% more heat output. Seems like good bang for the buck.

    By LoopCAD heat loss reports I am assuming you are referring to the Manual J (MJ8) reports. I will attach those to this posting so others can follow the reasoning based on the reports. I am also attaching a LoopCAD floor circuit layout for other readers benefit as well.

  25. user-2890856 | | #25

    Might I suggest that you call me ? 732-581-3833 . We can cover alot of ground in a shorter period . We can return to share with the community also , I have no issue with that .

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |