Optimal heating/water heating system to install

Discussion in 'Boiler Forum' started by Frozen23, Dec 15, 2017.

  1. Frozen23

    Frozen23 New Member

    Joined:
    Dec 15, 2017
    Location:
    12590
    Hi, i'am remodeling my house in NY 12603. It's a major overhaul. There was an air-forced furnace with ducts, but the house was abandoned for 8 years and i decided not to mess with a system that has no records of previous proper work. I will be installing 2400 ft of radiant pex pipe on top of subfloor and then covering it with laminate and tile, and in the basement walls . Maximum 2 heating zones. The house has 2000 sq ft x 8ft ceilings of heating space. There will be 11 new windows with u-factor of .28. The walls are 2x6 with R19 inside. The attic will be R38 at least and the basement is XPS R10 and fiberglass R15 can be easily added to the studs. There will be natural gas line. The consumption will include boiler, range, dryer and outside grill. CenHud suggested to install Navien 240 with the help of a captive installer. The quote is 9K minus incentives - total 5K out of pocket. Not bad but to me it's not an optimal solution. In search of reviews i found your forum. I've had Viessmann tankless combi-boiler. Non-condensing one. It gave me only headache - skipping switching from water heater mode to heating, some errors, heat exchanger has licked in 5 years. At the same time in my primary house i have Peerless cast iron 'gorilla' 87+BTU with indirect water heater and it works fine since it's been installed in 2004.
    After reading some forum threads i understand Navien 240 is an overkill and it can have problem modulating with all the tees and other gas-consuming equipment. Plus it doesn't leave room for a margin error with an installer. As long as i will combine radiant heat installed by me and their's installation the errors are anticipated. I need simple but effective solution. I open to suggestions
     
  2. Dana

    Dana In the trades

    Joined:
    Jan 14, 2009
    Location:
    01609
    Low mass combi boilers tend to suck for low load houses, and your load isn't very much. They're a much better fit for houses with high heating loads and average DHW loads.

    The min-fire output of the NCB 240 is about 17,000 BTU/hr The house you describe is probably going to have a design heat load of less than 30,000 BTU/hr @ +6F (Poughkeepsie's 99% outside design temperature.) With a min-fire output that's more than half the whole house load it won't modulate much, and will be cycling on/off quite a bit even in mid-winter. And since the load is divided into two zones there is some potential for short-cycling on zone calls.

    At least you don't have it micro-zoned to the nth degree, and there is at least SOME thermal mass in the 2800' of PEX, ~115 lbs of water if it's 3/8" PEX, ~215 lbs if it's half inch, so it's not hopeless. But it depends a bit on how the thermal mass is divided. Roughly how many feet of PEX (of what diameter) will be in the smallest zone?
     
  3. Sponsor

    Sponsor Paid Advertisement

     
  4. Frozen23

    Frozen23 New Member

    Joined:
    Dec 15, 2017
    Location:
    12590
    The pipe will be 1/2' pex-al-pex. There is less sense in creating additional heating zones in one big thermal envelope with no insulation inside. I mean, i lived in Europe with brick inner walls and it helped to keep different temperatures in different rooms with hydronic thermostats. Here i will have same length 9 loops of approximately 250-300 feet long with the same designed ambient temperature of 72F. Two heating zones are for the split-level basement (2 loops) and the main floor (7 loops). After reading the forum Navien i'am sure is not suited for my situation. I have two 12BTU heat pumps SEER 17.5 for cooling and heating (works down to 5F) 1350 ft of the main floor and willing to add a natural gas boiler as secondary (to increase comfort) during 90% of the heating time and 10% as primary heating source. I'am looking to build simple cost-effective system that do not require a lot of attention while maintaining and can be replaced by me. I believe one call to a technician for any Navien issue will kill all the saving it was intended to provide in long run compared to low-cost equipment from DIY stores
     
  5. Dana

    Dana In the trades

    Joined:
    Jan 14, 2009
    Location:
    01609
    A zone with 600' of half-inch PEX only has about 45 lbs of water mass, and will almost certainly short-cycle even if embedded in a concrete slab.

    Using a tank-type condensing gas water heater with an exterior plate type heat exchanger separating the potable hot water from the heating system water should work:

    [​IMG]

    A ~76,000 BTU/hr condensing water heater such as the HTP Light Duty (all stainless, with a modulating burner) or an AO Smith Vertex (glass lined non-modulating, won't last as long) would work, but you either have to find someone to design the system for you or design it yourself. The heat exchanger has to be rated for potable water, and the pump on the potable water side has to be tolerant of fresh water, but these things are pretty simple and effective. The thermal mass of the water in the tank keeps it from short cycling the burner, the thermostats only control the pumps and zone valves.
     
  6. Frozen23

    Frozen23 New Member

    Joined:
    Dec 15, 2017
    Location:
    12590
    Thank you, Dana for reply. My question is about components of the system. I'am going to put an ECM pump (TACO VT2218) for the radiant heating loop. If i find a potable water rated ECM pump for the main loop from a tank will they drive each other crazy constantly speeding up and slowing down ot it's my imagination?. Or i just can do with a regular 1/40 or 1/25 stainless steel pump to simplify the system?
     
  7. Dana

    Dana In the trades

    Joined:
    Jan 14, 2009
    Location:
    01609
    ECM pumps come with a number of different ways of programming the speed. Some you can just set a constant speed, some can operate off a temperature difference, some are just constant gpm, or constant power, but more efficient than older technology electric motors. A regular 1/40 hp stainless pump could do it for the potable side. Some of the small ECM drive pumps used for domestic hot water recirculation systems can probably handle it, but you'd have to do at least the napkin math to prove that.
     
  8. Frozen23

    Frozen23 New Member

    Joined:
    Dec 15, 2017
    Location:
    12590
    Dana, do you ever designed the system that allows weather compensation of supply water inside a water heater? Kind of a dynamic t-stat control with an outdoor temperature sensor. The lower the temperature outside the higher supply water.
    Of course one can control supply water with an electronic mixing valve but the only drawback they cost 400-500 bucks.
     
    Last edited: Feb 4, 2018
  9. Dana

    Dana In the trades

    Joined:
    Jan 14, 2009
    Location:
    01609
    I don't know of any water heaters that have what's called an "outdoor reset" function, where the temperature is adjusted in response to outdoor temperatures, but it is a a common feature of modulating condensing boilers.
     
  10. Frozen23

    Frozen23 New Member

    Joined:
    Dec 15, 2017
    Location:
    12590
    Hi Dana, i designed the system and now I have a conundrum on hands I don't know what to do with. I installed TACO VT2218 and at 130F it returns around 106F on 4th speed (60W or .5 Amp). On a cold day it's kind of not enough and my next move was to install more powerful Grundfos 26-99. I did and return on 3rd speed (197w or 1.7 Amp) is 108 (both readings were taken the same day). I purged the system thinking the air prevents but it didn't help. How come a 4 times more powerful pump with the head twice higher than TACO's can't get higher GPM? I have a GPM gauge on the heater and it shows very little difference - 1.8 GPM with Grundfos vs 1.6 with TACO (WATTS manifold balancing valves supports the flow - almost no visible change all 4 shows a little less than .75 GPM fully opened)?
     
    Last edited: Jan 7, 2019
  11. Dana

    Dana In the trades

    Joined:
    Jan 14, 2009
    Location:
    01609
    The gpm is not strictly a function of power-impeller design matters too. But pumping head increases quickly with higher flow, and conversely falls rapidly with flow. When designing a system it's important to spec both the flow and the pumping head at that flow to find the right pump.

    Comparing the split capacitor Grundfos 26-99 to the ECM drive TACO VT2218 isn't going to tell you very much. The fact that the Grundfos is consuming more than 3x the power is more a function of the motor design than anything else. Unlike the Grundfos VT2218, the speed of the 26-99 doesn't modulate with load.

    If you're getting 1.6gpm out of the VT2218 in speed #4 it looks like you have ~20' of pumping head, and it's a pump impeller & motor designed to be able to drive high head systems:

    [​IMG]

    The curves for the UPS26-99FC would seem to indicate it's capable of about 13 gpm at a 20' of head. (Notice how linear the curves are for the VT2218, and how nonlinear they are with the UPS26-99?) There's something amiss here- that's nearly an order of magnitude less flow than expected, unless the flow is being split, with only a fraction going through the NCB. Perhaps some pictures of the near-boiler plumbing is in order here.

    I assume you went with the NCB 240?

    If yes, did you install the pre-engineered manifold, or at very least another hydraulic separator, or is it being pumped direct (same flow through both the radiation and the NCB)?

    A flow of 1.6gpm is about 800lbs/hr. At a delta-T of(130F-106F= ) 24F that's about 19,200 BTU/hr, which is pretty close to the minimum firing rate of the NCB 240. The heat exchanger on the NCB can tolerate much higher deltas than that (I'd have to look it up to find the max), but don't be timid about raising the output temp to get more BTU out of it. With sub 110F return water temperatures you can easily bump it another 10F and still be in the mid-90s for combustion efficiency.
     
    Last edited: Jan 7, 2019
  12. Frozen23

    Frozen23 New Member

    Joined:
    Dec 15, 2017
    Location:
    12590
    No, I went on a spree and first used 40-Gal Tank water heater. It performed nice until i saw the bill for gas - with it's efficiency close to 62% i switched to Takagi TH3M. First i thought it's Takagi limiting the flow. Nope, it has a 7 GPM limit. Also while purging the system the flow raises to 4 GPM combining the flow of the pump and feed water

    I also tried Grundfos UPS 15-100 which gave the same result. Crazy
    20190107_164210.jpg

    I have raised the temperature to 140F and it's becoming less comfortable - the tile is too hot, and when the outside temp will drop to 6F the return will drop to the same 100F which is to low.
    So the only option is to raise flow rate. The only question is how?
     
    Last edited: Jan 7, 2019
  13. Dana

    Dana In the trades

    Joined:
    Jan 14, 2009
    Location:
    01609
    A tank water heater might have an EF of 0.62 in water heating mode, but it's combustion efficiency is about 80%, not 62%. At the much higher duty cycle of heating applications it's as-used AFUE will be close to it's raw combustion efficiency, not it's EF numbers.

    The high pumping head of the heat exchanger in the Takagi is one limitation- it should really be plumbed primary/secondary, separating the radiation flows from the flow through the tankless. Setting the flow on the tankless to 1.5-2gpm, and setting the flows to the radiation to whatever they need to be allows you to run the tankless at a high delta-T and the radiation at a much lower delta-T. The Takagi is designed for very high delta-Ts- even at it's maximum temperature it won't hurt the heat exchanger. But running the TH3M at 4 gpm or higher will wear out flow sensors quickly.

    With the pumping head of the heat exchanger you're better off pumping the primary loop TOWARD the tankless, or the pressure at the tankless can drop low enough during pumping to cause flash-boil in the heat exchanger.

    The expansion tank should be on a straight section of pipe close to the intake side of the pump to mitigate against cavitation risk. It's not clear where the pipe from the air scoop & expansion tank dissappear to at the top of the picture, or how it's connected to the rest.

    Calculating why you can't pound more flow through the radiation with a monster sized pump may be more than what's reasonable on a web forum, but it's clearly at the radiation plumbing end.
     
  14. Dana

    Dana In the trades

    Joined:
    Jan 14, 2009
    Location:
    01609
    Note: At 4 gpm the Takagi alone would be inserting 33 feet of pumping head, and at 5gpm it would be 55 feet of head. See the figure at the bottom of page 4. This is why it needs to be set up primary/secondary, running low-gpm /high delta-T through the tankless.

    At your 1.6-1.8 gpm it's still a non-trivial but mangeable 6-8 feet of head.
     
  15. Frozen23

    Frozen23 New Member

    Joined:
    Dec 15, 2017
    Location:
    12590
    I think 2.5-2.6 GPM will be enough for the system
    It's really noticeable difference - about 50 ccf/month less. And the other thing with a water heater of 40 KBTUs multiplied on .62EF - it doesn't warm up the house in colder weather.


    So yor are suggesting kind of this scheme.
    The only change the pump goes on supply side of the primary loop with the scoop before the pump, right?
    system-boilerlayout-3zones-1temp-wm.jpg
     
  16. Frozen23

    Frozen23 New Member

    Joined:
    Dec 15, 2017
    Location:
    12590
    There is nothing fancy there
    20190108_181150.jpg 20190108_180853.jpg
     
  17. Dana

    Dana In the trades

    Joined:
    Jan 14, 2009
    Location:
    01609
    A 40,000 BTU water heater has an output of about 32,000 BTU/hr, not 0.62 x 40K= 24,800 BTU/hr. The EF number is based on substantial idle time, and factors in the standby losses, not it's burner output. When the burner is running it really will deliver the full 32K.

    It's possible, even likely that your heat load is over 32,000 BTU/hr and running an atmospheric water heater that way will destroy the water heater in one heating season unless taking precautions to prevent condensation on the center-flue heat exchanger.

    Yes, you can use closely space tees to separate tankless flows from heating system flows. Keep the tees closer than 4 pipe diameters, with straight runs on both ends to minimize turbulence on the secondary.

    [​IMG]

    USe fatter pipe than 3/4" if you're looking for higher flow.

    [​IMG]
    [​IMG]

    But there are now TWO pumps, one between the hydraulic separator and tankless, the other driving the radiation flows:

    [​IMG]

    Note the pump driving toward the boiler/tankless drive the closely spaced tee loop, and separate pumps driving from the manifold to the radiation zones. (This could be one pump & zone valves on the secondary).

    Placement of the expansion tank & air scoop is ideally on the return side on a straight section just ahead of the closely space tees. (Which isn't how that is drawn.) If your radiation pump is on the outbound side of the tees and the boiler circulator placed as-drawn the air scoop and expansion tank would go about 8-12" before the return side tee.
     
  18. Frozen23

    Frozen23 New Member

    Joined:
    Dec 15, 2017
    Location:
    12590
    20190109_195709.jpg
    One mystery is solved - it's not manifolds. Grundfos 26-99 is pumping like crazy - all 9 manifold balancing valves show 1.25 GPM. TACO on cold water started at 2.6 GPM but on return above 100F dropped to 1.1 GPM (viscosity, i guess). And now i have the same issue i had before - low return temperature. The supply on the secondary loop is 140F but after mixing with primary decreases to 118F and return on manifolds is 108F. Should i consider a pump with a higher head (like Grundfos 15-100)?

    UPDATE - after 20 sec rebooting TACO pumps back to 1.5 GPM. Could it be the air that prevents flow?
    UPD2 - and in 10 minutes flow is back to 1 GPM.
    UPD3 - i swapped the pumps and now TACO pushes the same flow (it shows 6 GPM but on 9 valves i see >1 GPM/each) on primary loop and Grundfos gives 2.6 GPM on secondary. We'll see in couple hours what the flows are
    UPD4 - in couple hours I'am back to page one - Grundfos is pumping 1.7 GPM vs 2.6 GPM from the start. I switched TACO to 3rd speed (shows 4 GPM) and now the supply jumped to 130F and return to 115F

    Still don't get it why flow in secondary loop deteriorates over time?
     
    Last edited: Jan 10, 2019
  19. Dana

    Dana In the trades

    Joined:
    Jan 14, 2009
    Location:
    01609
    Since the secondary loop's intake is a mixture of tankless output and return water, it's temperature will always be lower than the raw tankless output. If you need more BTU/hr, increase the temperature of the tankless.

    With 115F in, 130F out it's a lower than optimal 15F delta-T, and 4 gpm is too high a flow rate. Ideally you would have at least a 25F delta-T at the tankless, even 75F deltas are just fine. Tankless water heaters are designed to work at high delta-Ts, and will be somewhat more efficient when it's >25F, at any flow rate. The 4 gpm through the tankless it will probably wear out the flow sensor. Running it at 2-3 gpm flow at a higher delta-T would be easier on the equipment.

    At a 15F delta & 4 gpm you're looking at about 30,000 BTU/hr. A 30F delta and 2gpm is also 30,000 BTU/hr. So if you bumpted the output temp to 145F and pumped at 2gpm it would be the same heat rate. Bumping it to 145F and pumping the primary at 3 gpm would be 45,000 BTU/hr assuming the same return water temp. I don't know what the max setting is on that model, but see how the system behaves set to the max, with the Taco set to it's lowest speed- 2gpm much more "polite" flow to be driving through a tankless.

    Air in the system can reduce flow. Crud in the little screen filters on most Takagis (don't know much about your model) can also reduce flow.

    The drop in flow on the secondary over time is a mystery not easily solved via web forum.
     
  20. Frozen23

    Frozen23 New Member

    Joined:
    Dec 15, 2017
    Location:
    12590
    I think, thanks to you, mystery is solved.
    Today the flow in secondary loop dropped to 1.1 GPM and I decided to flush the water heater backwards (from supply to return). I believe it was crud that prevented the flow. And it helped - now Grundfos pumping at 3.2 GPM. So i think I'll go to the primary loop only with one TACO pump by putting it before the heater on return side plus installing inline wye strainer with a fine mesh
     
  21. Frozen23

    Frozen23 New Member

    Joined:
    Dec 15, 2017
    Location:
    12590
    Yes, mystery is resolved. I put the water heater at 135F and TACO tracks return water at 100F. And today at outside 25F it pumps at 1.1 GPM using only 8 watts. But it can pump at 3 GPM when start the cycle
     

Share This Page