GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X 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

Do I need to insulate duct work that is inside the envelope?

user-1140356 | Posted in Mechanicals on

I am considering adding a geothermal heating system to my house in South Jersey. The utility room is inside the house. The trunk and duct work branches are very short and they will be inside the envelope in soffits, pretty much as detailed in “Keeping Ducts Indoors.” I plan to seal the trunk and branches with mastic. Do I need to insulate them as well? Is it better to insulate the soffit or the duct?

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. GBA Editor
    Martin Holladay | | #1

    If all of your ducts are inside your home's thermal envelope, the ducts don't have to be insulated. Meticulous air sealing is more important than insulation. Make sure that:

    1. The soffits are meticulously air-sealed before the ducts are installed. Remember, if these soffits are pressurized (due to supply duct leaks), and if your soffit has leaks to the exterior, the air delivered by your air handler will be forced outdoors.

    2. All duct seams are meticulously sealed.

  2. user-1140356 | | #2


    Thank you once again for your advice. Your website is a godsend for me in South Jersey.


  3. user-945061 | | #3

    I think you should reconsider your choice of GSHP. I know your area fairly well, and have had interactions with many dissatisfied GSHP owners. Massive risk, massive expense. You could install an air-source heat pump, it would be more reliable, similar or better performance, and you would spend a fraction of the money. The money you save could be spent on upgrading the building shell.

  4. user-1140356 | | #4

    Hi Jesse,

    I apologize for not getting back to you as to the dense-pack cellulose. I believe I have upgraded the shell as much as is prudent with R20 - 40 - 60. The house is tight.

    The GSHP will cost about $4000 more than the mini-splits (maybe less than that if I went with multihead minisplits). The Feds are generous at tax time with GSHP. For that $4000, I will be able to clean and circulate the air in the house better through a return in each room (including circulating the wood stove heat). I should be able to hook the HRV to the system if I choose. I think the GSHP may be more robust long term.

    I am not convinced that mini-splits provide even heat throughout a house that has separate, closed-off areas. On the discussions on GBA, I note that differences of 5 degrees in temperature are expected between rooms. I am older and less tolerant to such differences. I hope to get even older and I doubt that I will get more tolerant.

    I do thank you for your advice. I must admit the choice was not an easy one.

    thanks again,


  5. Expert Member
    Dana Dorsett | | #5

    In more conventionally insulated houses with ducted air heating and cooling it's pretty common to see room-to-room temperature differences in excess of 5F at heating & cooling temperature extremes. Unless you're clairvoyant about the plug-load use and occupancy of the rooms, even in a much better insulated house, you can still expect to see temperature differences of 5F or more, even with a balanced ducted GSHP system. (Two gaming teenagers with an X-Box with a big screen TV will exceed the average mid-winter heat load of many family rooms in a high-R house, causing the temp of that room to rise when used in that mode.)

    The only way to get much tighter than that a 5F room-to-room delta is by micro-zoning the heating & cooling system on a room by room basis, which presents both significant design issues and a VERY significant upfront system cost uptick. Most people (even cold-blooded Floridians :-) ) can find a thermostat setting that works fine for comfort, even if it's 2-3 degrees warmer in the warmest room than the whole-house average, and 2-3 degrees cooler in the coolest room. Without micro-zoning the solar gain even of low-E triple panes will easily boost high-gain rooms 5F over the temp of the lowest gain rooms for hours at a time. on any day of the year, winter or summer.

    The design risk of GSHP systems is also very real- every system is a custom system with many not-easily measured variable, and as a result there are MANY instances of real-world systems underperforming expectations from both an efficiency & comfort point of view. By contrast, ductless heat pumps are "system in a can", with all of the hard design work already done, with a limited number of ways to screw it up, which makes them an attractive solution for low-load houses. But micro-zoned ductless with a head per room is expensive and usually pretty silly, buying very little in the way of real comfort. The real comfort-investment is jn the high-performance building envelope, not the mechanical systems.

    Last (and probably least), to get high efficiency in heating mode out of GSHP with air rather than hydronic heat distribution requires running them at air temps well below human body temp, which presents a wind-chill issue. Duct placement has to be considered very carefully to keep humans out of the air stream, which adds a constraint to the furniture placement, etc. Ductless systems were originally developed for smaller Asian houses, where it's assumed that the air currents will be blowing on humans at least sometimes, and by-design the exit air temps are above human body temp as a rule, with less of a wind chill issue should you find yourself in the breeze of the ductless head on the coldest day of the year. The wind chill effect is also mitigated by modulating variable speed blowers in the ductless heads keeping the cfm (and blower power use) to the minimum required for keeping up with the load.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |