Supply next to Return
![](https://secure.gravatar.com/avatar/2b69fc037dd19a9f5f5e9fb53bbe4f7a?s=52&d=https%3A%2F%2Fimages.greenbuildingadvisor.com%2Fapp%2Favatars%2F200x200%2Fl.png&r=g)
In typical projects, I like to locate supplies right next to returns to keep duct layouts short. In particular, I like to have the air handler up in a conditioned attic space and then distribute supplies on interior walls out towards the exterior walls (i.e. the Load). And, if it makes sense, then I put the returns on that same interior wall, and sometimes it right next to the supply.
I’ve noted that this approach is controversial. Some argue that there’s going to be short-circuiting.
Can someone describe the physics of this short-circuiting?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
![](https://www.greenbuildingadvisor.com/app/themes/greenbuildingadvisor/dist/img/modules/detail-library/thumb1.png)
![](https://www.greenbuildingadvisor.com/app/themes/greenbuildingadvisor/dist/img/modules/detail-library/thumb2.png)
Replies
I don't know that you need to bring physics into it. If you have a vent pushing air into a room, and right next to it is a vent sucking air out, it seems pretty obvious that some of that air is going to go right from one to the other without much interaction with the rest of the air in the room.
Yes-- it does seem obvious, especially from the movies. But is it real?
Is there a suction force, like gravity, that is drawing the air from the supply towards the return?
What are its physics-- equations and such?
"Is there a suction force, like gravity, that is drawing the air from the supply towards the return?"
Am I missing something? Yes. Yes. The return is literally sucking in air. Is it specifically targeting the air from the supply duct? No. But the closer the two are to each other, the less time the supply air has to mix with the rest of the volume of air in the room, and a higher percentage of the air that is being sucked into the return is fresh supply air rather than "well mixed" supply air.
I believe I've read you typically want them at least 6 ft apart.
You shouldn't design systems that way. Ideally, you want your supply and return diagonally opposed across the room, but that's usually difficult to achieve in practice. Next best is to have the supply and return on opposite walls, in which case it helps to have them placed to work with convection. That means that if you're in a heating dominated climate zone, you want your supplys down load and your returns up high. If you can't do that, then put them both down low.
What happens if you put the supply and return next to each other is you don't stir the air in the room much, so you end up with a localized spot of air near the vents and not a lot of air movement further out in the room. This means you won't get as even of a temperature throughout the room as you would with more seperation between the supply and return vents, which would "stir" the air more.
For an extreme example, in datacenters we bring cold air in through floor grates, and exhaust hot air out the top of the "hot isles", with partitions between hot and cold areas to force all the cold air into the electronic equipment. By PREVENTING mixing of hot and cold air as much as possible, we maximize effective cooling of the equipment, which improves efficiency of the facility. This IS very noticeable, so it's not an imaginary effect.
Bill
Data center take away-- supply and return opposite each other minimizes mixing. Is that correct?
I can understand a little about having the supply/return diagonally opposite, but would that still reduce possible mixing, due to the same effect of the Data Center? You're throwing air out towards a return, which can then catch the air on the other side of the room before it can diffuse to the other corners (especially on the along the plane of the supply).
In a datacenter, we do NOT want mixing. What is typically done is "cold isle containment" (or "hot isle", sometimes both), which is done by partitioning off isles between rows of equipment racks, usually using sliding doors. The floor is pressurized with cold air, so we can put "perf tiles" (floor tiles with vent holes, basically) whereever we want cold air to come out. The air "return" is either inside the drop ceiling (best), or just follows allow the ceiling (simpler), back to the air handlers, and the fans in the air handlers run continuously. The cold air rises up from the floor by air pressure, gets sucked into the fronts of the equipment racks, gets blown out the back of the equipment racks by the fans in the equipment in the racks, then the hot exhaust air gets carried upwards to the ceiling return by convection. It is not uncommon to have a 50*F thermal differential between cold supply and hot return.
In a residence, you WANT mixing, because you want EVEN temperatures throughout the space. By spacing the supply and return vents out across the room, the air handler in the HVAC system acts to "stir" the air, which helps to prevent the room from having hot and cold spots. This makes for a more even overall temperature, and better comfort. There will be some impact on system efficiency too, but not nearly as much as is seen in a datacenter, which I used as a sort of extreme example to make it easier to understand the point I was trying to make.
Bill
ASHRAE has some tables showing effectiveness of various configurations.
(there is one in the 62.1 ventilation standard). It can be significant.
There is more to it than location. Temperature deltas and air velocities play into it.
Supply and exhaust close to each other are rated at only 50% effectiveness.
You run a high risk of drawing in warm air to reheat instead of the cool air and vice versa in cooling mode. Leaving much of the room without adequate air movement.
Elegant duct design with high effectiveness should be a top priority, it helps tight sizing work, is more comfortable due to more even temperature distribution with fewer cold corners or exterior zones. It also lets you get away with a lower ventilation rate to achieve the same air quality. Or in other words you have to over-ventilate to get the same air quality with a subpar ductwork design because you're not changing the air.
I'd suggest playing with a smoke pencil.
In playing with a smoke pencil, (or really a theatrical fog machine) in the past, I haven't seen much of a 'suction' effect. That said, I haven't been able to play around with a full setup where I can adjust the supply and return.
Intuitively, I wet my hand, and blow air at it from my mouth. I feel air movement. At the same distance, I suck air into my mouth. I feel no air movement (or no change in evaporative cooling). What's going on here?
If you put your lips together and blow, the air is directed in a small stream and easily felt. If you open up wide like a yawn you will feel a lot less air pressure. Accordingly it is possible to push the air out of the supply at high velocity to achieve mixing of air to make up for return and supply being close together. But this is more complicated and it wouldn't work well for setting the fan on low to circulate air.
BTW-- I'm trying to find the ASHRAE tables-- maybe I have to go to the library.
I realized that I have a perfect experimental setup in my house. big room (500 sq.ft.) and a window heat pump in the middle of the exterior wall. Supply is directly above return (like any ductless minisplit cassette). If there's short circuiting, it's imperceptible in this configuration. I'm getting great air mixing.
I'd rather have a longer duct run than a room that's uncomfortable/not entirely efficient in the grand scheme of things.
Not heating supply and return but a good visual none the less.
NS Builders range hood smoke demo.
https://youtu.be/u2uVaTMXd_8?t=143
Interesting how the makeup air is actually coming into the hood with a downward direction before getting entrained upward towards the exhaust fan
This is not the first time this type of MUA design came up.
If you look at the video, it clearly shows that it is actually NOT working. The makeup air gets sucked right into the hood and exhausted out. This means very little of the thermal plume from the range is actually vented out (shows that only the back half of the hood is venting kitchen fumes). This is a very good way to exhaust all your makeup air but a very poor hood setup.
You want the flow from the makeup air to contain the thermal plume, this mean it should at very low velocity and supplied from all the sides of the range.
There has been a lot of research on MUA for commercial kitchens. Most recommend only a small fraction of the makeup air to be supplied by the hood as in the video (about 15%) the rest needs to be supplied elsewhere. Since most houses won't have multiple MUA feed locations, the best solution is to not do this at all and supply the makeup air either behind the range or the toe kicks on either side.
Interesting point here Akos
Why isn't the exhaust fan's suction force sufficient here to exhaust the plume itself? Couldn't the makeup air be anywhere (for barometric relief)-- so long as the exhaust fan has the localized effect of sucking/drawing/forcing the pollution out?
This video does seem to demonstrate more complex physics, which is exactly the spirit of my question. How do we understand the suction force in this application?
To your question about the physics of short circuitng: Short circuiting is when some of your incoming air is getting sucked back into the outgoing air. Or that suction is influencing the throw of the supply Ie: your system is not as effective as it could be.
I assume the supply velocity and stream characteristic plays a big factor in this side by side scenario. It's just a rule of thumb, that you're not supposed do that, and as you may have noticed, not many people on this forum like the rule of thumb approach. Fluid dynamics is really finicky and there's lots of variables in homes. It's hard to say for sure or even guess that you're short circuitng with little system information and no testing.
So get out the flow test box and anemometer, play with a smoke pen, but better yet, place a smoke bomb upstream of your supply and look at it. If you do the smoke bomb, video tape it and post it here. I'm sure we'd all like to see it for knowledge sake.
Other things to consider is the actual location of the return in terms of returning the stagnant air. Assuming you're in zone 7, based on your name, you're in a heating dominant area. You might get better efficacy from having your returns low on the wall to suck cold air. But that also makes more work in terms of ducting, which I assume was your original concern.
So, Does your system work? Maybe it works good enough? How's the efficacy of the air distribution ? You tell us.
Here's a fun video and document on this stuff.
https://youtu.be/5ReA3B2elNM
https://www.priceindustries.com/content/uploads/assets/literature/engineering-guides/air-distribution-engineering-guide.pdf
Jamie
Jamie-- these resources are excellent, but also present a different set of paradigms here. The video says that "return air inlet location has very little effect on room air diffusion regardless of inlet type or location". And then in the next sentence says to put it a 'sufficient' distance away to avoid short circuiting.
What is a 'sufficient' distance? (general question)
Smoke test
If you read the document, it also says on page 3 that's the location of the return doesn't matter, but not close enough to short circuit it.
So I think the crux of it is, don't short circuit the airflow.
Now, this document addresses systemes generally in commercial with higher air velocities.
You've yet to mention what kind of systems you're putting in. Or what the blower specs are or your duct design etc. If it's your typical high volume-low velocity home HVAC system, I believe both the supply and return locations plays a bigger role Vs a high velocity system, the locations matter much less.
Just my thoughts,
Jamie
The HVAC subs I work with generally do Manual T, so the throw is managed.
I'm working with heat pumps usually. Usually ducted medium static VRF units.
you need to terminate the supply and return in such a fashion that when system is operating there is flow through the room. with the registers close to each other some of the air will short circuit and the lower the speed the more it will be so.
you can experiment very easily. put a temp sensors in the supply, return and opposite side of the room (total of 3). run the system and observe. if in heating mode the return temp will be closer to the supply temp rather than the temp across the room- that is your short-circuited air in action.
ideally you want your return air to be the same temp as your room temp and create air flow in the process so you dont have hot/ cold spots in the room. ideally this means supply and return registers need to be in the opposite sides of the room- supply usually close to the outside walls (under/ above windows) and return registers on the inner wall across the room.
This, it's not just about direct short cycling. It's about moving circulating all the air, especially in outside corners and against exterior walls.
Diffuser/register selection and setup can make a big difference in how much throw/reach the diffuser has. Sometimes smaller is more appropriate due to higher velocities.
+1 This can work but you need high throw (usually means high velocity and restriction) supply register.
In cold climate, you can't go wrong with supplies under windows, even above windows can create some comfort issues if the registers don't have enough throw.
Thanks everyone for the responses:
The conclusions and inferences I'm taking away here:
--Return adjacent to supply is a bad idea and will cause short-circuiting. The return's suction force will (like gravity) draw air towards it and have a significant effect on the throw of the supply, thus reducing mixing.
--Ideal locations for return are in a surface (wall, floor, or ceiling) opposite of the supply to effect good air mixing.
--This has interesting implications for hvac selection. I would infer from this that ductless mini-splits are inherently flawed in their mixing potential since the supply and return are essentially colocated. But in a pinch, it'll work well enough. Also, central returns (say in a core hallway in a house with supply runouts to perimeter bedrooms) are also flawed since they will draw air out of the bedrooms (through interior door undercuts or bypass ducts hopefully) before that air can fully mix. A better return location would be dedicated to each bedroom on a surface opposite the supply.
Thanks everyone who responded!
Ductless minisplits are stuck with that tradeoff, since they are packaged units. What they typically do to try to offset this is to run with relatively high air velocities, and a directional grille that can "throw" the air farther compared to what you'd get with a normal HVAC supply register. With a ducted system, you don't have to make the same design tradeoffs and you have more flexibility with the placement of vent openings.
Bill
Thanks Bill.
I feel like I'm gradually getting to a better understanding. That is-- with sufficient throw (Manual D/T) then it should be fine to co-locate a supply and return and achieve good mixing.
I'm still wondering about the suction force itself-- shouldn't it have a simple equation a la the gravity (9.8 m/s^2) but one that's relative to the negative pressure at the return inlet? Not asking you-- just thinking out loud.
an example-- we can calculate the kinetic energy of an air-stream via 1/2 mv^2. Then we can model the force vector of the suction force (most simply via the model of F=ma where 'a' is the aforemention expression of the suction force relative to return inlet pressure delta)
There is a rule in science: "Science never sucks". there is no "suction force", not really. The force is other stuff pushing towards an area of low pressure. You're very much limited with how much you can do with "sucking" compared to what you can do with pressure, by "blowing".
I think you'll find it much easier, and you'll get better results, by locating the supply and return at opposite ends of the room. You can easily do this by running the supply duct and return duct around opposite sides of the floor, for example, or "leapfrogging" branch ducts off of a main plenum if you have to run the two main runs of ducting together. There are various ways to do it depending on your layout, but it's not usually difficult to route the supply and return to opposite sides of the same room.
Bill
Great Bill
There is no suction force. Others are saying there is...
In the case of my projects, the solution of running ducts to the opposite ends is a non-starter. Slab-on-grade, vaulted ceilings. Not technically impossible, but I don't like the idea of running ducts underground or on the roof (nor does the architect). the notion of keeping a small/compact duct distribution system seemed to check a lot of good boxes, so I was exploring the paramters of best practice in this scenario