Welcome! Here are the website rules, as well as some tips for using this forum.

# A REALLY basic hydronic question

Member Posts: 202
In a simple residential reverse or direct return system, what causes the water in the hot water pipe to divert to the radiator? In other words, what would cause the hot water to exit the 3/4" pipe at a tee and flow through the radiator? I understand how the water in the 3/4" pipe would want to divert enough flow to fill the radiator, but I don't understand the dynamic that would force it to circulate through the radiator and out into a direct or reverse return pipe with any velocity.

Numbskull question....but I don't know the answer.

• Member Posts: 2,917
Resistance to flow, or path of least. The circulator produces pressure, which results in flow, the resistance to flow in the pipe and fittings results in influence through a given circuit.
Contact John "JohnNY" Cataneo, Master Plumber for Consulting Work
Or for plumbing in NYC or in NJ.

Or take his class.
• Member Posts: 202
On pages 160 & 161 of "Modern Hydronic Heating" (Siegenthaler), 4 diagrams show different uses of a Single Diverter Tee. According to the diagrams, these can be placed in the main supply pipe either just past the rad inflow, just before the rad outflow, or in both locations simultaneously. What I don't understand is the pressure threshold at which one or both if these tee's are needed. Complicating the equation are the TRV's that I'll have on every rad, each likely applying a different flow constraint through the various rads.

This is the point at which I begin asking myself if I have sufficient technical knowledge of fluid dynamics to build a well balanced, efficient system. On the other hand, the 2 local hydronic "professionals" seem to have even less knowledge than I, so I carry on.

In this case, it sounds like technical expertise can be trumped by just getting an oversized pump?
• Member Posts: 202
I'm designing for a 1500 sq ft 3/2 residence. I've done the heat loss calculations so have a working knowledge of boiler/panel sizing for my home. My home is built on a slab and with post and beam construction (i.e., no crawl space and no attic) I've decided on a single loop system due to the logistical challenges of running the pipe. I've been told by an "expert" that a single loop is more than sufficient on a modest sized home.

It's come down to a decision between the following three system types: One pipe, two pipe direct return, or two pipe reverse return.

I estimate the total loop length at about 175', with a total of 9 radiators, each equipped with a TRV.

A one pipe system is attractive from a cost perspective, but I'm concerned about heat loss toward the end of the loop; I know that TRV's will mitigate this to some extent, but I have no easy way of knowing to what degree.

If I go with a 2 pipe system, a reverse return seems to make sense since it will be a single loop in a circle configuration, supplying all rads.

I'm also at a loss as to loop pipe sizing. On a loop this short, is there any need to alter the pipe diameter at different points along the route? Would a 3/4" continuous loop with 1/2" rad connectors be sufficient for my design?

Thank you...
• Member Posts: 7,330
edited July 2015
That is NOT "A REALLY simple question".

You just mixed direct return, reverse return and diverter tees in one question.

Direct return needs to be balanced.It is best to use real balancing valves.

Reverse return will balance itself. This is true because, by design the water in all loops will have the same amount of resistance.

Diverter tees are going to drive you nuts with zero upside.TRV's will make it worse.
If you decide to go that way, pay alot of attention to air removal and purging valves.

You have been looking at this for a couple years. How about posting a sketch?

I still suggest homeruns or reverse return.

Carl
"If you can't explain it simply, you don't understand it well enough"
Albert Einstein
• Member Posts: 202
Zman, it did indeed start with a simple question and has migrated as you point out. I'm open to changing the title if possible...just need instructions on how to do that.

As I indicated, a reverse return home run system seems to make the most sense for the reasons given. And it sounds like a reverse return system obviates the need for diverter T's. The TRV's are staying though, and I had planned to install air removal and purging valves.

Any thoughts on my question: "On a loop this short, is there any need to alter the pipe diameter at different points along the route? Would a 3/4" continuous loop with 1/2" rad connectors be sufficient for my design?"

On the subject of air vents, Siegenthaler's book mentions they should be placed at hight points in the pipe, which makes sense. My loop pipe will run on top of my flat roof, with connector pipes to each rad dropping down and then rising back to the loop. Siegenthaler seems to suggest a vent for every emitter. What's not clear from the book is whether the vent should be just prior to the emitter input pipe's takeoff point, or just after the emitter output pipe's return to the loop. Do you have any thoughts on this.

I've posted several diagrams of my proposed system on prior, older threads. None of them show the home run loop that I'm now leaning toward. I'll update and post later this evening.

And yes, I've been kicking this around for a couple of years. Paying construction projects have prevented me from finalizing this one, but I've recently purchased property out of the area and am putting my house on the market soon so it's time to get this done.

Thank you...

• Member Posts: 202
Hatterasguy, as I indicated I'm leaning toward a single loop two pipe reverse return system. Any thoughts on whether a 3/4" loop is sufficient for my planned system?

The panel rads are all at the low point.....see above, my loop is overhead, rad inflow/outflow drop down then rise back to the loop. It stands to reason that the air vents will need to be on the roof...the question is where should I position them in relation to the panel inflow and outflow pipes?

Looks like the Taco HS007 is variable 0-20 GPM so sounds like that's the one I want, right?

Thank you...

• Member Posts: 7,356
Again, mixed metaphors. I doubt I would design a new diverter tee based system unless someone put a gun to my head. Reverse return is cheap enough and easy enough using modern polymer pipe (most common translation = PEX) that I have trouble seeing a good reason not to do so.
• Member Posts: 7,356
Is this by any chance an Eichler?
• Member Posts: 202
"Again, mixed metaphors. I doubt I would design a new diverter tee based system..."

I don't get the comment. I'm planning on a 2 pipe reverse return home run system with panel trv's. I asked about diverter tee's back at the top of this thread. Looks like you may have missed my later posts.

Yes this is an Eichler.
• Member Posts: 7,330
Most panel rads have the option for TRV's, isolation valves and air purge valves.
I would do either reverse return or home run depending on the layout.
Don't overthink the vents. A manual vent on each rad and a well thought out boiler piping plan is all you need.
"If you can't explain it simply, you don't understand it well enough"
Albert Einstein
• Member Posts: 202
edited July 2015
Anyone have an opinion on whether 3/4" loop pipes are sufficient for my system, and on where to position air vents?
• Member Posts: 7,330
You will need to provide more info for piping sizing. The pipe sizing is based on the amount of energy you are trying to move and how far you are trying to move it.
"If you can't explain it simply, you don't understand it well enough"
Albert Einstein
• Member Posts: 202
edited July 2015
How do you define "how far you have to move it"? Again, this is a 175' reverse return single loop with 9 rads. Is it 175' total? Or do I also add in the return pipe distance? Do I need to add the cum panel inflow/outflow lengths?

Zman, this confuses me: "I would do either reverse return or home run depending on the layout." I'm doing a reverse return home run. I don't see an either/or in this.

I don't know how to calculate "the amount of energy" I'm trying to move.
• Member Posts: 16,301
The loop needs to be sized to the amount of BTUs it has to move.

I'd assume a heat load of 20 but/ sq. ft, just for an example. So 1500 sq. feet of building =30,000 BTU/ hr which would be a 3 gpm flow.

You need a room by room heat load calc to size the radiators, how did you come up with 9, and what are their outputs?i

I think you have the piping systems confused.

A home run is where each radiator has a supply and return that goes to a central manifold location, typically at the mechanical room. So you would have a 9 port manifold up top or at the equipment. Possible 3/8 or 1/2 pex would feed each rad.

Reverse return is a loop around the home with take offs for each radiator. A third pipe assures equal distance for balancing. A true reverse return would have the pipe size change as every load is branched off, but that rarely happens in a job like yours. (page 14 in the link below)

Or a basic two pipe with balance valves at each branch. You can get panel rad valves with balancing function built in.

This link may help explain the different types of piping and the need for balance.

http://www.caleffi.com/sites/default/files/coll_attach_file/idronics_8_0.pdf

Sounds like the piping is exposed on the roof top? That would make a home run system a bit awkward.
Bob "hot rod" Rohr
trainer for Caleffi NA
Living the hydronic dream
• Member Posts: 202
Thanks Bob. Yes, I'm using incorrect terminology. Let me clarify. I'm installing a parallel reverse return system. It will run from the boiler room up onto the flat top roof, around the roof perimeter and returning to the boiler. Positioned along the 175' run are 9 Runtal wall panel radiators connected to 1/2" inflow/outflow pipe that will enter and exit the exterior frame wall cavity through holes drilled in the top plate.

I've completed a room by room heat loss calculation, using 2 different commercial calculators: 1) Slant Fin Heat Loss Calculator software, and 2) Siegenthaler's Heat LoadPro software. I'm using 20 btuh/sq ft and sizing my radiators for 150 degree water. Each of the 9 radiators will have a TRV to allow individual rooms to be set to a desired temperature.

From this description, do you have an opinion on the impact of using 3/4" parallel piping the entire length?

I had planned to install air vents at the location of each radiator branch. Do you have any suggestions as to where I should place them (either before the rad inflow branch pipe or after the rad outflow pipe)?

• Member Posts: 7,330
Your heating lines are all on the interior of the building, yes? I start hearing about running lines to the roof and drilling top plates and I wonder...
I sounds to me like you are getting carried away with this venting thing. You should be able to put one manual vent on the upper loop for purging the original install and you will be fine. putting hidden vents all over will lead to leaks down the road.
I agree that 3/4" piping should get the job done. I have no idea how Hatterasguy has signed off on the 007 without knowing anything about the near boiler piping, sizes and length of piping or the resistance characteristics of the radiators and TRV's.
Looks like it time to start drawing.
Carl
"If you can't explain it simply, you don't understand it well enough"
Albert Einstein
• Member Posts: 202
Zman, it looks like you've missed a couple of my posts. I said above: "My home is built on a slab and with post and beam construction (i.e., no crawl space and no attic)" I also identified the house as an Eichler, which is a well known Western mid-century style characterized by post and beam construction, with exposed tongue and groove roof decking, typically surfaced with a tar and gravel buildup.

You don't show your geographical location so this style of construction may be unfamiliar to you.

It does present challenges in running piping. I have no interest in opening up the drywall around the entire interior perimeter of my house to run Pex, so the only options left are to 1) run the copper over the roof, encased in poly sleeve insulation & covered to protect from UV, with rad connectors entering and exiting the wall cavity through the top plate, or 2) run the copper in a soffit against the exterior wall, up tight to the eave, with the rad piping entering the wall cavity through the outer wall. I stated above that the single loop is ~175 lineal feet, however last night I decided to eliminate 2 radiators from the design, leaving 7 radiators and shortening the parallel piping to ~150'. This length doesn't include the 1/2" panel piping.

I'm not sure how to answer your question about the resistence characteristics of Runtal rads & TRV's. Their website has flow rate and pressure drop calculations, accessed through this link: http://www.runtalnorthamerica.com/commercial_institutional/technical_calculating_same.html. I'm not sure how to apply this info to determine pump size.
• Member Posts: 7,356
Don't worry about the pressure drop in the rads, it's effectively zero at those flow rates. Just figure the GPM needed for each one at design conditions and add them all up. I would use a 30°F ΔT on that, and I would step the trunk lines down from 3/4" to 1/2" where the velocity falls below ~4 FPS.

Are you planning to box the pipes on the roof, or just insulate them individually? Either will work, but the details are important and you want plenty of insulation.
• Member Posts: 202
Thanks SWEI. I'll work on the GPM info and post it later.

If I run the pipes over the roof I had planned to use poly sleeves with R=3.21, and then cover with rain gutter turned upside down (that's become popular out here to block UV from breaking down the poly sleeve). I'd line the inside of the gutter with an additional layer of poly.

If the pipes run in an eave soffit, they'd also be covered with poly sleeves, and the 2 soffit sides would have rigid foam glued to the inside surfaces.

I'm leaning toward the soffit method due to 1) aesthetics and 2) running pipe over the roof would require installing sealed roof jacks at each point of penetration. Doing this on an existing t&g deck requires a multi-step process involving fiberglass scrim and a lot of Henry's. I've done this before and would like to avoid it on this job.

thank you
• Member Posts: 202
edited July 2015
Swei, I'm having a difficult time understanding Runtal's formula for calculating GPM:

GPM = (BTUH/FT X FT of Radiator) DT (DT X 500)

I'm not understanding the algebraic bracketing syntax on this formula. They claim that with a 445 BTUH/FT, 10ft panel length, and a DT=20, the flow rate is .445. I don't understand how their answer is derived with their formula. I checked Siegenthaler's book but found no reference to a panel GPM calculation.

The first part of the formula is obviously (445 x 10ft)= 4450. But how does the balance of the formula, 20 (20x500), get me to .445 GPM?

Although the formula bracketing doesn't make sense to me, it seems that what they're saying is to calculate rad btuh/ft times the number of panel feet, and then divide that number by the product of DT times 500. In their example this would be (445 times 10) = 4500 divided by (20 times 500), or 445 divided by 10,000= .445. I'll apply that logic to my situation and report the GPM's shortly.
• Member Posts: 7,356

I'm having a difficult time understanding Runtal's formula

... [snip]

it seems that what they're saying is to calculate rad btuh/ft times the number of panel feet, and then divide that number by the product of DT times 500. In their example this would be (445 times 10) = 4500 divided by (20 times 500), or 445 divided by 10,000= .445. I'll apply that logic to my situation and report the GPM's shortly.

You have to start with the design day heat loss for the room or area served by the radiator. That determines both the sizing of the radiator and the required flow rate. The Runtal nameplate ratings assume an Average Water Temperature (AWT) of 180°F (supply at 190°F and return at 170°F) which you really don't want to use for a high efficiency boiler -- or even a conventional boiler in a house that has (or might have) kids in it. Make a spreadsheet with the design day heat loss of each room in one column, and then look at the lower temp ratings for the Runtal, e.g. (for the Contractor Series):
Outputs at 140° are 350, 450, 550, 850 and 1050 Btu’s/Ft on UF-2, UF-3, UF-4, UF-6 and UF-8 respectively. For outputs at various temperatures, see the complete Runtal catalog.
140°F is a reasonable AWT for a high efficiency boiler (lower AWT will increase boiler efficiency, but not all that much.) With a 30°F ΔT, that would imply a 155°F supply temp and a 125°F return temp. Use those ratings to determine how many feet of panel you need for each room. Now go back and measure the rooms to be sure those lengths will fit. You may need to move to a taller model or even add a second radiator in some rooms. Once you have those parameters all optimized, you determine the flow rate for each emitter using the formula:

[room heat loss] ÷ [500] ÷ [30] = GPM
• Member Posts: 202
Swei, I've done precisely as you've described right up to the point of calculating rad GPM's: I've completed a detailed design day heat loss calculation for every space in the house including hallways and closets. I've plugged that info by room into Excel and determined how many feet of a given radiator I'll need to heat each room assuming 150 degree entering watering temp (Runtal's default sizing is based on 180 degrees; they specify using a .67 factor for 150 degree water). I know precisely what rads I need for each room. What I hadn't done is calculate individual rad GPM's, which I'm calculating today. I feel like I'm close to moving forward with an installation, but you guys keep throwing more questions my way (for which I'm grateful, BTW).

Thank you
• Member Posts: 7,356
Are you sure the Runtal numbers are EWT and not AWT? Your numbers above are for a 20°F ΔT, but you should be able to get by with a 30°F ΔT so I would check that as well.

Whatever you settle on, take the GPM numbers and start adding them up around the loop. At 4 FPS, 1/2" PEX will carry 2.2 GPM (22,000 BTU/hr at a 20°F ΔT or 33,000 at a 30°F ΔT) and 3/4" PEX will carry 4 GPM (40,000 BTU/hr at a 20°F ΔT or 60,000 BTU/hr at a 30°F ΔT.) You can safely exceed any of those by 10-15% if needed.
• Member Posts: 202
edited July 2015
Swei, Runtal states: "The BTHU outputs are listed for 180 degree water"

I interpret that as EWT=180 degrees. Would you agree?

Runtal says to multiply 180 degree BTUH/Ft rad output by a factor of .67 to derive BTUH/Ft output at 150 degrees, which I did for sizing the panels.

Here are the GPM numbers calculated for each panel, in sequence within the loop:

GPM Panel #
0.307 1
0.130 2
0.223 3
0.147 4
0.161 5
0.307 6
0.307 7

1.582 GPM cum all panels. Does mean I can get by with 1/2" pipe all the way around?

It sounds like the 007 pump is more than sufficient.

Been a long day, brain getting fuzzy. I'll be back on here tomorrow morning to try and wrap this up.

Thanks for the help
• Member Posts: 15,331
Does it ever get below freezing in Mountain View ?
All Steamed Up, Inc.
Towson, MD, USA
Steam, Vapor & Hot-Water Heating Specialists
Oil & Gas Burner Service
Consulting
• Member Posts: 202
Our coldest winter months avg low is in the mid 40's; our summer months avg lows are in the low 60's
• Member Posts: 7,330
Stick with your 3/4" plan. Pushing that much water through a 1/2" line will require a larger circulator.
"If you can't explain it simply, you don't understand it well enough"
Albert Einstein
• Member Posts: 7,356

Swei, Runtal states: "The BTHU outputs are listed for 180 degree water"

I interpret that as EWT=180 degrees. Would you agree?

I would interpret that as AWT (which they explicitly state in some of their documents.)
Here are the GPM numbers calculated for each panel, in sequence within the loop:

GPM Panel #
0.307 1
0.130 2
0.223 3
0.147 4
0.161 5
0.307 6
0.307 7

1.582 GPM sum all panels. Does mean I can get by with 1/2" pipe all the way around?
In theory, yes -- but assuming those numbers are accurate, you won't find a pump that will really do the job right, and you won't find flanges smaller than 3/4" anyway.
• Member Posts: 16,301
perfect application for a delta P circ, like the Grundfos Alpha.
Bob "hot rod" Rohr
trainer for Caleffi NA
Living the hydronic dream
• Member Posts: 202
Swei, you are correct....I went back and reread the Runtal literature and in the Calculating Flow Rate section they do state that nominal panel btuh/ft are based on 180 degree AWD. I've made the corrections to my spreadsheet and will post shortly, but before I do I have a question: I have a central hallway that accesses all 3 bedrooms and 1 guest bath. The design day heat loss for that space is 960 btuh. I don't plan on a radiator in this space, so how would one typically account for this? The hall is accessed from the great room so would i just add the hallway heat loss to the great room and size the great room rads accordingly (i.e., slightly larger)?
• Member Posts: 202
edited July 2015
Hattersguy, that post is way over my head. What precisely does "if only one or two rads call" mean? Could you put that in layman's terms?
• Member Posts: 4,469
I'm still trying to figure out a single loop, reverse return. What you describe sounds like a series loop, and that won't work. I haven't had a beer yet, but it sounds like a good idea.
• Member Posts: 202
I messed up on the terminology earlier in this thread, but clarified it a couple of days ago: I'm designing a parallel reverse return system. Is that a valid description within the hydronics lexicon?

I've also been referring to it as a "single loop" when I believe I should have been describing it as a "single zone". By "single zone, parallel reverse system" I believe I'm describing a single hot water pipe feeding all of the radiators in my house, with a single return pipe back to the boiler containing all of the outflow from all of the radiators, controlled by a single thermostat located inside the house. Does make sense? I want to get terminology correct before I make all of us crazy.
• Member Posts: 202
edited July 2015
hot rod said:
"perfect application for a delta P circ, like the Grundfos Alpha."

Hattersguy responded:

"What I'm struggling with about that pump is its inability to get below 6' head using constant pressure or 4' head if you let it "auto-adapt" (if it could actually do it). If one radiator calls, and it might be desirable to flow about .5 gpm through that radiator, it would appear that the Alpha is going to send at least 2.5 GPM to that single radiator before it reaches 6'.

There isn't enough resistance in the 3/4" loop until a flow rate of 2.5 GPM is achieved and there is negligible resistance in the 1/2" supply and return from the radiator (presuming a total length of six feet).

Now, it may not be an issue with those rads but it does appear that the flow rates through them will be quite a bit higher than the calculation shows if only one or two rads call."

My comment:
Not sure if this is germane to the above, but the way I'm setting this up is a single thermostat in the great room mounted to the wall just adjacent to the hallway entrance that accesses the 3 bdrs and 2 baths. With this setup I'd expect the call to always be initiated by a drop in the great room temp below the thermostat setting. I'll have 3 radiators positioned around the great room, so it stands to reason that all three would be triggered within a very short time assuming their settings were similar, which is the plan. I cannot imagine a scenario where a temperature drop in a rear bedroom(s) would trigger the great room thermostat to send a call, unless all the TRV's on all 3 great room rads were turned off which is not likely to occur. Does this make sense or do I have to eat my words yet again? In this scenario, would there still be a concern about the resistance issue mentioned above?

• Member Posts: 202
edited July 2015
Got it...thx.

Attached is my radiator sizing worksheet. 1.88 total GPM. Fire at will.
• Member Posts: 7,330

There isn't a "concern". It's just an interesting characteristic of the Alpha when very low flow rates are required. It's still far better than any fixed speed circ.

An excellent point, it is sad commentary on hydronics in America that the alpha is the smallest option.

I reality you are going to end up with higher flow rates than your design, this will be especially true when only one or 2 trvs are open.This is not a big deal, you will just end up with a smaller delta t than your design.

What are you using for a boiler?
Primary/secondary piping?
"If you can't explain it simply, you don't understand it well enough"
Albert Einstein
• Member Posts: 202
edited July 2015

I'm installing the parallel reverse return piping and radiators myself, I had planned to pay a hydronics contractor to set up the boiler room piping and the CDN 040....unless you think I can do it myself using the installation instructions/diagram and factory tech support...and helpful tips from anyone here familiar with that boiler.
• Member Posts: 2,740
R.A. Calmbacher L.L.C. HVAC
NJ Master HVAC Lic.
Mahwah, NJ
Specializing in steam and hydronic heating
• Member Posts: 5,839
Simplicity is beautiful... Go two pipe parallel reverse return, using 3/4" S&R with 1/2 branches. Install TRV's on all radiators except for the "reference" radiator, which is usually found in the worst case scenario heat loss room (thermostat location), and may not be the largest room in the home, and let it rip.

Having a low delta T is not neccessarily a bad thing if you've done due diligence (which it appears you have). It ends up giving you a higher average surface temperature, which equates to better output from the emitters overall. It is what it is.

You can waste a lot of time on crunching numbers and fine tuning every detail, but bottom line, the difference in comfort and parasitic energy consumption will be barely noticeable.

I'd go with the DCECM circulator because it will have a lower parasitic power consumption factor than anything on the market.

WILO makes a 5 watt circulator that is DC powered, but it complicates things because now each radiator requires its own circulator and room sensor AND power source. Great concept, but complicated install. K.I.S.S.

Based on your original input, when you talk single loop, are you talking a series loop? If yes, like baseboard, it can be done, however you must increase the length of the emitters based on the fact that the water temperature will drop each time it passes through an emitter. It also creates more resistance to flow because all pressure drops are additive in a series circuit, whereas in a reverse return circuit, it is the pressure drop of the S&R mains at design flow, plus the worse case pressure drop on the branches added to the mains loss.

If you have SIggys software (comes in his MHH books) you could perform a calculation of a series circuit and see what it comes out to.

A one pipe or monoflo or diverter tee system is a cross between a parallel reverse return system and a series system and as others have pointed out, if improperly installed can be a bear to purge. That is what I have in my home. It works great, but filling and purging can take hours. Start pump, stop pump, burp ALL radiators, repeat, continue until all free air is gone. I had to add dish detergent to my system to finally get all free air out.

When the main is higher than the emitters it is recommended that you use two diverter tees, one as a diverter on the supply and one as a venturi on the return, thereby compounding the problems associated with pressure drop on the loop.

As the water moves through the loop, it gets cooler and cooler, again requiring an increase in emitter length to compensate for drop in fluid temps.

The other distinct advantage of the parallel reverse return system is that all emitters see the same supply water temperature (minus minor piping losses, which are essentially a gain to the conditioned space, so may be ignored.

ME
It's not so much a case of "You got what you paid for", as it is a matter of "You DIDN'T get what you DIDN'T pay for, and you're NOT going to get what you thought you were in the way of comfort". Borrowed from Heatboy.
• Member Posts: 7,356
At those flow rates, you should easily be able to direct pipe the CDN040, BTW.