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

Direct vent vs. power vent gas water heaters

mpls1921 | Posted in Mechanicals on

A recent blower door test/audit has alerted me to the fact that my natural draft hot water heater can backdraft in worse case scenarios (range hood/clothes dryer/bath fan on). It seems to reverse back to a correct draft after a few minutes. It’s only about 6 yrs old, so I hate to replace it, but not as much as I hate the idea of the back draft.

So, it seems my options are power vent or direct vent (with exterior air intake/exhaust, but no fan/electricity required). I can’t think of a reason why I would want the power vent when the direct vent would be more quiet and not require electricity. Why would I use a power vent unit?

House is in Minneapolis.

Thank you!

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. Expert Member
    Dana Dorsett | | #1

    Power vented unit would still be an infiltration leak in the house, but it can't backdraft when burning. A direct vented burner has no air communication between the house and the combustion air or exhaust venting, and is NOT an infiltration path.

    I've never seen a direct vented unit that didn't need electricity. Can you point me to one? They tend to be quieter than power vent units because there is no direct air path to the room air, but they use blowers to induce the draft.

    The only reason to go with power vent over direct vent would be lower upfront cost.

    An atmospheric drafted HW heater is a 24/365 infiltration leak, and in fact drives infiltration from other leaks even if it's not backdrafting, since it has no flue damper.

  2. mpls1921 | | #2

    Yes, the infiltration with a power vent is also an issue. I had never heard of a direct vent w/o electricity either until I saw Mr. Holiday mention one in an article from a couple years ago that I just read where he mentioned a Canadian product. Looking around after that, I see that similar products seem to be available in the US now. For example:


    Seems a little more expensive, but no noise and no electric use seems worth it.

  3. mpls1921 | | #3

    Forgot to mention something. I can feel the heat of the backdraft, and it will fog a mirror, but I put not 1, but 2 CO detectors right on top of the water heater and neither said "boo". One of them has a digital ppm readout and showed "0". Now, I'm sure these homeowner alarms are not designed for this kind of use, but I would have thought there would be enough CO to set off the alarm...or at least indicate SOMETHING.

  4. GBA Editor
    Martin Holladay | | #4

    Here is a link to info on the direct-vent water heater manufactured by GSW Water Heating:

  5. GBA Editor
    Martin Holladay | | #5

    A clean-burning gas appliance may not emit any carbon monoxide. Just because your CO monitor doesn't go off, doesn't mean that the appliance isn't backdrafting.

  6. mpls1921 | | #6

    Thanks Martin. I knew the CO could be low, I just didn't realize it would be that low. Still a serious problem, obviously. Just recalled that a "chimney sweep" had taken off the old 5" metal flue cap at the top of the chimney and replaced it with one of those cage type things. Perhaps I will start with a proper cap to the top of the chimney and see if that improves anything. Any insight into why those "direct vent" units aren't more popular? At least here in Minneapolis, one of the big pluming/HVAC company reps has never heard of them.

  7. GBA Editor
    Martin Holladay | | #7

    Q. "Any insight into why those 'direct vent' water heaters aren't more popular?"

    A. No, I'm sorry that I don't have any insight into that issue.

  8. bennettg | | #8

    When I looked into the direct-vent (natural convection) types, it was clear that there was a very limited vertical and horizontal range for the combined flue / intake to the outside, plus clearance around the outside cap. I suspect this limits their popularity.

  9. PaKettle | | #9

    A direct vent draws very cold outside air in while a power vent uses the warmer inside air.
    I would easily bet direct venting lowers the overall efficiency numbers....

  10. Expert Member
    Dana Dorsett | | #10

    Briain: You would lose that bet.

    The combustion air temperature has an unmeasurable effect on the flame-front temperature (just shy of couple thousand degrees) or the temperature of the gases going up the center-flue heat exchanger. The raw combustion efficiency is about 80%, independent of whether it's direct vent sealed combustion or atmospheric drafted. But by not drawing DILUTION air from the room air out the draft hood it's quite a bit more efficient when looking at the house and the water heater as a system.

  11. PaKettle | | #11

    Unmeasurable -yes but not incalculable.

    In the end you're right - there are far too many other factors with a much larger influence on the final outcome,

  12. Dana1 | | #12

    And when the order of magnitude of the calculated difference is 0.1%, there's no point to it, eh? The difference in delta-T across the center-flue HX at the bottom is at most a couple of degrees (out of several hundred). The difference in average delta-T over the full length of the HX would be less than a degree.

  13. benrush | | #13

    Dana, Martin, (or anyone else)-
    I'm very interested in Dana's observation (comment 10) that if the water heater and the house are viewed as a system, a direct vent sealed combustion water heater is more efficient than an atmospherically vented water heater, because it's not using conditioned air for combustion. (Did I get that right? Feel free to correct me if I misrepresented something.)
    Can you help a non-physicist to get some idea of the savings the direct vent water heater would provide? I suppose the most important factors are how much the water heater runs, and the temperature difference between conditioned and outside air.
    I'm surprised that the direct vent marketing folks aren't pointing out the energy savings.

  14. user-2890856 | | #14


    A direct vent draws very cold outside air in while a power vent uses the warmer inside air.
    I would easily bet direct venting lowers the overall efficiency numbers....
    Answered by Brian Larsen
    Posted Mar 10, 2016 7:10 PM ET

    Briain: You would lose that bet.

    The combustion air temperature has an unmeasurable effect on the flame-front temperature (just shy of couple thousand degrees) or the temperature of the gases going up the center-flue heat exchanger. The raw combustion efficiency is about 80%, independent of whether it's direct vent sealed combustion or atmospheric drafted. But by not drawing DILUTION air from the room air out the draft hood it's quite a bit more efficient when looking at the house and the water heater as a system.
    Answered by Dana Dorsett
    Posted Mar 10, 2016 7:28 PM ET

    Dana, I may be wrong but I think Brian may have been speaking of thermal efficiency also when he stated overall . The direct vent has almost as terrible a thermal efficiency as an atmospheric . While it does draw outside air while firing it also has a larger flue area through the tank specifically . Larger surface area in contact with tank will certainly equal greater standby losses , No ? Very good convective current happening in there with stack effect out drawing cold air down . No argument on the combustion efficiency though . Not quite sure who loses the bet .

  15. GBA Editor
    Martin Holladay | | #15

    Dana's point about dilution air is that not all of the air that is drawn into the combustion zone and up the flue is required for combustion. The volume of air is more than what is needed for combustion.

    Moreover, air moves up the flue of an atmospherically vented water heater even when the burner is off (drawn in by the heat from the warm water, which warms the flue). For many homes, the atmospherically vented water heater operates as a kind of ventilation system (exhausting air out of the room where the water heater is located, and pulling in outdoor air through cracks in the thermal envelope).

    It's very hard to estimate the effects of this type of air leakage on a homeowner's energy bills, because the answer to the question depends on many factors, including the leakiness of the home in which the water heater is installed.

  16. Expert Member
    Dana Dorsett | | #16

    There is a larger surface area in contact with the HX, but less standby convection due to the typically smaller diameter of the forced draft flue, and thus less convective loss. Some may even have flue dampers to impeded parasitic convection to improve their EF numbers but I can't point you to a particular model that does. As-installed the stack effect pressure of direct vented systems is also typically lower, due to the intake and exhaust being at roughly the same elevation, rather than the exhaust being 15-40' above the combustion & dilution air intakes. Even though it's sometimes vented through the roof, it's more common to see them side-vented less than 10' above grade, with less than 2' of elevation difference between intake & exhaust. Unlike an atmospheric drafted unit there isn't a low-impedance convection path around the outside of the tank, between the exhaust hood and the combustion air intake.

    AND without needing to provide dilution air, a direct vent forced draft unit isn't removing room air 24/365 that has been conditioned by the other mechanical systems. Thus, viewing the house and mechanical systems as a single system, it's still generally more efficient at any given EF rating, even if there may be a few exceptions to prove the rule.

  17. benrush | | #17

    Thanks Richard, Martin, and Dana.
    Fascinating stuff. Making my head spin a little bit- but I like it.
    Thanks again,

  18. benrush | | #18

    Some atmospherically vented water heaters may be a little better than we thought.
    Recently, for a certain situation, I opted to install one of the best (in terms of energy rating and warranty) atmospherically vented water heaters I could find.
    Shortly after opening the box, my plumber told me the WH required an electrical outlet. At first, I thought he was mistaken- and that the electrical requirement was only for an optional Wi-Fi module that I hadn't bought- but it turned out that the water heater had an electric flue damper.

    Though still not as good as a model with sealed combustion, it's a better choice than an atmospherically vented WH without a damper- and I thought some of you might like to know about it.
    By the way- is there a way to "follow" a forum, so that I get an email when someone comments? I feel like Martin and I talked about this a couple years ago. At that time, the answer was "no". Still no?

  19. Dana1 | | #19

    The flue damper stops the convection through the center-flue heat exchanger, which is good for improving the water heating efficiency.

    But the damper doesn't block the infiltration from the dilution hood to the exhaust flue & the Great Outdoors. It's still a 3"-4" hole in your building envelope with 24/365 stack-effect draw, a parasitic heating & cooling load that a direct-vented system does not have.

  20. benrush | | #20

    Not saying it's great. Just pointing out that the best of the worst water heaters is better than we thought.
    Which aspect is more wasteful- the convection thru the center flue heat exchanger that this WH addresses? -or the infiltration from the dilution hood... to the great outdoors that still remains?
    I realize there are lots of variables- but do you have a guess? (Chicago area, WH in basement; +/- 25 vertical feet from dilution hood to top of chimney)
    If I've addressed 30% of the problem, that's kinda cool. If I've addressed 3% of it, I've wasted our time.

  21. GBA Editor
    Martin Holladay | | #21

    I don't have an answer to your "3% or 30%?" question. But as long as your water heater doesn't have backdrafting problems, I wouldn't worry about it if I were you. It's not a huge problem.

  22. charlie_sullivan | | #22

    The convection through the tank is a heat loss all year, working on a temperature difference that is somewhere near 70 F. The leakage through the envelope that Dana points out is only a problem during heating or cooling seasons, and is very rarely as big. So I think the damper does solve considerably more than half of the problem.

  23. Dana1 | | #23

    With the flue damper the standby losses of the tank are smaller, and a greater fraction of that smaller standby loss accrues to the heating & cooling loads of the house. The parasitic loss to the house from the open flue doesn't change, so it really is a net benefit.

    The loss to the house varies a lot by climate, and the particulars of the house, But it's not even close to 50% of the problem, since it doesn't even raise the EF of the water heater by 50%. The steady state combustion efficiency is about 80%, the 2015 EF energy conservation standards the minimum efficiency for gas fired HW heaters under 55 gallons is a minimum EF of 0.675 - (gallons x 0.0015), so the minimum legal 50 gallon unit would have an EF of:

    0.675 - (50 x 0.0015)= 0.60.

    A 50% improvement in standby loss would require an EF of 0.60 + (0.80 - 0.60 )/2= EF 0.70.

    This series of water heaters test between 0.67 and 0.69, not 0.70, and would probably still beat code-min without the damper. While EF 0.67 is achievable even without the flue damper (there are numerous existence proofs), the flue damper is probably a cheaper way to get there. Hopefully the 24/365 electric controls for the damper are sub 1-watt while in monitoring mode too, and not one of those "efficiency" measures that trades one parasitic loss for another of larger size from a different energy source.

  24. benrush | | #24

    Why do we need a dilution hood anyway?
    (Not what I wanted to lead with- but this way it's visible in "recent comments".)

    Thank you all for your feedback.
    I hope to try a direct vent model in the future, but in this situation I didn't want to relocate the WH to an exterior wall, nor did I want to put a hole thru the wall either.
    Still, it sounds like I probably did better than I usually do (which is to buy whatever's available locally when a WH breaks). This time I had enough time to order one.
    At the end of summer I may try to compare this summer's gas bills to the previous year's. (Rental property. Not my gas bill. But same tenants for many years.) Not exactly a scientific study, but with no other significant gas use, it should be somewhat meaningful.
    Last question: Why do we need a dilution hood at all? Why don't we connect the flue right to the top of the WH? (If someone addressed that previously, sorry- I missed it.)
    Thanks again.

  25. GBA Editor
    Martin Holladay | | #25

    There are several reasons why an atmospherically vented water heater needs a draft hood (or dilution hood).

    One document explains it this way: "When properly installed, the draft hood cools the combustion products and allows them to flow safely from the building. It also protects heater operation during updrafts and downdrafts that are caused by indoor/outdoor air pressure differences."

    Another document explains the purpose of the draft hood this way: "The draft will change in the chimney as exhaust vents towards it – especially when going from cold air to hot. A draft hood is placed above the upper most part of the gas furnace to draw air into the chimney and makes it possible to draw more or less air through the chimney as necessary to create a constant flow. This makes it possible for the burner to enjoy consistent air flow without any wind gusts or sudden temperature spikes or drops. Hot air, if not put through a draft hood would create a strong air flow through the burners."

  26. benrush | | #26

    Thank you. Now I get it.
    I didn't realize it- but it sounds like furnaces and boilers have them too. Maybe I never looked- or maybe they're inside the enclosure.
    Also interesting (from the second link) "For every cubic foot of gas burned, the furnace needs to have 15 cubic feet of air for combustion and another 15 cubic feet of air for dilution." That's more than I expected- and a good reminder of how inefficient it is too use conditioned air for combustion (and dilution).

  27. Dana1 | | #27

    Aside from evening out the combustion draft variations, a dilution hood is necessary to lower the dew point temperature of the gas combustion exhaust. The exhuast condensate is mildly acidic, and will corrode non-stainless metal or masonry flues & flue liners. Without the dilution hood copius condensation occurs (primarily in winter). With terra-cotta lined flues this can be an expensive repair, and corrode-through B-vent is a health hazard.

    Even WITH the dilution hood venting a water heater into a grossly oversized masonry chimney has sufficient condensation to become a problem. This is a common occurance when boiler or furnace that had shared a masonry chimney with a water heater gets replace with a power vented condensing heating appliance, leaving the old chimney to the water heater. The water heater doesn't operate at a sufficient duty cycle or BTU rate to keep the chimney liner above the dew point of even the diluted exhaust in winter, whereas a space heating appliance usually does. This phenomenon even has a name, the "orphaned hot water heater" problem.

  28. benrush | | #28

    Very interesting.
    Just to make sure I've got it: The dilution air reduces the concentration of acid in the exhaust- so that it has a higher dew point- so it doesn't condense inside the flue/ chimney?

  29. Expert Member
    Dana Dorsett | | #29

    You have part of it, but some of it is backwards.

    The primary benefit of the dilution air cuts the concentration of WATER VAPOR in the exhaust (even though it also cuts the acid content), resulting in a LOWER not higher) dew point, making it less likely to condense onto or adsorb into the flue materials where the residual acids can cause problems.

    The combustion products of CH4 (methane, the primary component of natural gas) are one molecule of CO2 to two molecules of H2O (water). In addition there is the excess combustion air that didn't react, and other compounds that got formed from that air at high temperature (small amounts of nitric acid, NOx compounds, some carbonic acid etc). At 80% combustion efficiency there is typically between 15-30% of excess O2 that didn't react in the flame, and almost all of the N2 that passes through without reacting in the flame, as well as some amount of water in the combustion air that passed through with no reaction.

    Even with the substantial N2 and the excess O2, fully half of 70-85% of the combustion air O2 that reacted in the flame formed water, making it an extremely water-vapor rich mixture of gases compared to the combustion air that came in. At the 15-30% typical excess combustion air of an ~80% efficiency burner the undiluted exhaust of natural gas has a dew point of about 130-135F. Dumping that directly into even a 100F chimney liner presents a bit of a condensation problem. Dumping it directly into a 40F or cooler chimney will condense copious quantities.

    The water content of room air is MUCH lower than that of the exhaust mixture, with dew point no higher than the room temp (usually much lower than the air temp), and even though mixing with that low-temp room air reduces the temperature of the mixture, it lowers the saturation temperature ( dew point) of the diluted mixture by even more. The lower the dew point temp of the admixture, the colder the flue liner can be without incurring copious condensation.

  30. GBA Editor
    Martin Holladay | | #30

    For more information on the "orphaned water heater" problem and flue gas condensation, see "When a Flue is Too Big."

  31. benrush | | #31

    Ok, now I get it. Wow- my understanding was a long way off.
    I meant a lower dew point- but was calling it a higher dew point.
    But more importantly, I didn't realize how much water is produced by combustion. My cave-man mind just didn't want to accept that fire makes water.
    Thanks everyone.

  32. mpls1921 | | #32

    This whole discussion is a great reminder of why I would love to have a nice PV array and stop having any combustion appliances in my home!

  33. jpritzen1 | | #33

    As someone who had an atmospheric draft water heater & "upgraded" to power direct-vent {intake & exhaust PVC pipes}, the electricity usage has been almost unnoticeable, however, the noise is unbearable.

    I debated direct vent as it did not need electricity & didn't have fans to push air, however the huge hole that I would have in my wall was the deal breaker, in addition to the thought that the plume of smoke I used to see coming out of my chimney would now be at ground level {in addition to being extremely hot}.

    I retracted the thought of a power vent because it would still encourage outside air infiltration when running. In hindsight, it wouldn't have been that much of a negative because it could be serve as exhaust-only ventilation for my basement - and I have the tight house/low ventilation problem.

    My initial push for upgrading the water heater from an atmospheric draft was because I too had a backdraft issue (but no measurable CO). Now I dread the moment my power direct-vent hot water heater comes on due to the noise of the fan (inside & outside the house).

  34. AlbertaCanuck | | #34

    Another perspective on why you might not wish to use a power vent is if you used a direct vent gas water heater, certified for combination use in space heating, as a source for backup heating when the grid goes down during a winter cold spell.

    In Alberta Canada, the overnight low can occasionally get down to minus 30 degrees Celsius (or about -22 degrees Fahrenheit for those in the US) for a week or two each winter. In that scenario, if the local electrical grid goes down for a couple of days, the outcome can be serious w/o a back-up heat source. Common back-up heat sources here include having a gas fireplace that doesn't require power to run.

    Another alternative though is to use one's gas water heater in combination with one or two radiant hot water heating circuits, with hot water circulation through the HW radiant heating circuits pumped by dc pumps connected to a couple of modest solar panels mounted outside.

    The plumbing is a bit complex though as the Alberta plumbing code requires use of a closed indirect system for that kind of set-up, meaning use of a single walled heat exchanger, water pressure reducer, and dual check valve, such that a non-toxic fluid in the heating circuit (e.g.50/50 propylene glycol plus water mix) operates at around 20 PSI with no / v. little risk of back-flow into the potable water supply. A side note is that given a "combination" type HWT, is that local code requires that at least one potable water fixture be supplied from the tank, which is fairly simple to accommodate in order to meat Alberta plumbing code.

    One question that I haven't resolved yet is whether use of a direct vent tank-type HWH as the heat source for radiant heating would be relatively efficient in the sense that the average steady state efficiency might be in the order of 70% and perhaps a bit more. Helpful input on that much appreciated!

  35. GBA Editor
    Martin Holladay | | #35

    You'll need more than "DC pumps connected to a couple of modest solar panels mounted outside." You'll need two types of pumps -- AC for ordinary use and DC for emergency use -- along with valves to be able to convert from one system to another, and switches to turn the pumps on and off. More importantly, you'll need to have a homeowner who remembers in an emergency which switches to flip, which valves to turn on, and which valves to turn off. (Even if one homeowner remembers these steps, that homeowner's spouse may not.)

    Finally, the PV panels won't be enough. You'll also need a battery system and a charge controller, because it's often dark and cloudy when emergency power is needed. These components are expensive and need maintenance.

    The bottom line: buy a wood stove or a gas-fired space heater with through-the-wall venting.

  36. CyrusJuliet | | #36

    One important point is highlighted by the variations in venting. A direct vent water heating system may be installed practically anyplace, making it simpler to install than a power vent system.

    1. user-5946022 | | #37

      Article is wrong, and suspect the above poster resurrecting a 5+ year old thread is spam, trying send traffic to their linked article.
      However, in the event someone at GBA reads this:
      A non powered direct vent water heater has a limitation on the length of the vent. Generally the vent pipe cannot be longer than 6' (depending upon how many turns), thus the water heater needs to be near an exterior wall.
      A power vent system is probably simpler to install because the vent can be a longer run, so installation location is not limited to x feet from exterior wall.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |