Diy radiant...

raided

Hi all,

I am expecting some pretty...honest opinions on what I should not have done.  I am hoping for some feedback that will get me on my way - if bruised.

Here goes.

We contacted an online radiant floor company after deciding to replace the existing radiant floor heating system.  We ran all the tubing - +-10%/circuit in a given zone, less than 300' per circuit, etc. 

The rest should make more sense with the attached picture.



When we fired up the system, the first zone (all the way to the right) got warm as expected.  Interestingly, zones to the left of the running zone didn't seem to change much.  Shutting off the right most zone resulted in the now right most zone suddenly getting warm.  All other zones did not change.  And so on.

When closing the ball valve on the primary loop, all zones supplies immediately get warm as you'd expect.  However, the pumps sound a bit labored and like they have air in them.

The system was designed such that the ball valve on the primary loop shoud be normally open.


I fully recognize the system pump is on the wrong side of the expansion tank - that was my stupid.  I contacted the radiant company and their advice has been very disheartening to the point I realize they are sales - not support.  They explain a lot of bad or unexpected behavior as "phenomena".

The water heater company said the unit is fine for heating and they warranty for that.

Find a contractor yielded a result ~1.5 hours away.  I will try but that seems like a lot to hope for.  I also am trying to find local companies on google.  I am wondering how far off we are and would love some advice.

Thank you.

DZoro

Did you fill / bleed each zone with water, one at a time. You must get every bubble of air out of all the lines and the boiler.

Ironman

What size pump is that on the primary loop?

A tankless water heater has far too much head to be used in place of a boiler. It's also not controlled or approved to be installed in place of one.

That being said, you may be air bound, but your piping is not correct. Study the attached drawing to see how closely spaced Tees are necessary to create primary/secondary piping.

Correcting the piping, along with a large enough primary pump, MAY get the system functional, but not totally correct without a true boiler.





In your scenario, run the boiler loop through the bull (side) of the Tees. The drawing shows zones valves, but that's where your pumps would go. And disregard the one pump shown near the expansion tank.

Radiant Wreck strikes again!

Jamie Hall

All the pumps are the same? Primary circulator and secondaries? I need to know before I comment further... but I have a suspicion.

raided

Thanks for responses - very appreciated.

Is it possible this is simply an air problem? Based on the pic, should that bv on the primary be open or closed - assuming it can even stay?

DZoro - we went zone by zone, circuit by circuit for a while. We will bleed again to be sure. I can't be sure the air we hear is from cavitation due to the misplaced system pump. I've read "you'll bleed til the cows come home!"

Iron man - grundfos 15-58fc on the primary. Am I reading you right that we essentially need to rip out the heart and repipe the whole thing as a proper p/s? That sounds painful but not as much as a long winter with minimal heat. "Radiant Wreck" - I am rolling.

raided

Apologies. The zone pumps are 15-58's. The system pump is bigger - i dont have the model handy but can get it this evening.

All variable speed and all set to high.

Ironman

I don't think you need to rip it all out, but you do need to make it conform to the drawing.

You'll need AT LEAST a ups26-99 on the primary to overcome the high head of the water heater's heat exchanger.

Jamie Hall

OK. Looking at the pumps and the arrangement of the piping, what I am suspecting is that whichever loop is first in line is stealing most of the flow from the boiler. Going to be interesting to balance without some changes.

raided

Agreed. I am open to your ideas. I am going to sit down with the feedback and sketch something while in front of it. So far it sounds like air is a distant possibilty with repiping to create proper secondaries. Without being able to fully soak in the scope (at work) it sounds considerable. Looking to get a solid plan.

Tim Potter

What is the GPM flow rate per per pumped zone?
What size "Big, Short, Fat" Header do you need for all that GPM?

Tim

hot_rod

Here are a couple options. Using a hydraulic separator gives you some added features, a small buffer volume, air, dirt, magnetic and hydraulic separation. In either case, the piping for the zone pumps needs to be sized to provide adequate GPM when all are running 1- 1-1/4 probably.

Zman

I think you would be amazed by how much better your system would work if you did the following
1. Move the expansion tank to the bottom of the air separator. This will solve your air elimination issues and make you circs run better.
2. Turn the zone circs to speed 1.
3. Repipe the zone returns in reverse order (reverse/return). First zone in is last zone out. This will help even out the flows.

raided

The pipe on the primary is 1". The radiant company provided gpms per zone and came up with that size. Although oll take ot with a grain of sand.

I can't simply move the expansion tank because I do not believe the eliminator is outfitted for it. That would also take some repiping. That's fine and i expect to have to do that but i am hoping it to be done in one big hoorah.

hot_rod

Zman said:

I think you would be amazed by how much better your system would work if you did the following
1. Move the expansion tank to the bottom of the air separator. This will solve your air elimination issues and make you circs run better.
2. Turn the zone circs to speed 1.
3. Repipe the zone returns in reverse order (reverse/return). First zone in is last zone out. This will help even out the flows.

With the ball valve in that loop closed, the high head "boiler" circulator is in a sort of push pull arrangement, boiler pump pulling from returns, zone pumps suppling. I think?

With the ball valve open you have some bastardized parallel piping, pumps in series.

Without closely spaced tees there is no variation of P/S going on in that confusion.


Either way some very un-predictiable flows going on and undersized piping adds to the confusion. Wonder if there are checks in any of the circs?

raided

They do have flow checks in them. The tankless adapts down to 3/4", also. Guessing that is very wrong.

hot_rod

7 circulators for a small application like that seems way overkill. I believe those are 70- 80W circulators, probably more on the boiler circ. So that adds to over 500W! The pumps alone may heat the building on mild days

What model water heater? Be good to determine the pressure drop through it. In some cases you cannot pump them enough to get full output.

What is the heat load for the home?

raided

Oh interesting. I did head loss calcs to double check what i was being fed and the pumps seemed to be correct, if big. Next step down was too small. Unit is a gt540 -American "powered by takagi". Im not sure the relationship there.

I need to look at my notes (not with me) to answer on heat load.

hot_rod

I think that is a 199,000 BTU unit? What size home are you heating?

The manual doesn't give any info about then pressure drop in that tankless or the info you need to size the pump when they are in combined systems.

I think yours is dedicated just to the heating load?

Zman

I agree that my solution is less than perfect and ideally you would take a sawzall to much of it. I have worked on systems that were intentionally piped as I described and was surprised that they really did not run that badly. I would leave the valve open, as I see it, that would be the boiler loop, and yes it could effect flow to the zones to some extent.

Hilly

The picture is a little fuzzy but is there a "drain" on the side of the gas valve - meaning not a gas valve? Again it's fuzzy on my phone and blowing it makes it harder for me to see.

raided

Hot rod - roughly 2600 square feet heated. It is a closed loop system.

Zman - bv open means it's heating one zone at a time although the curcs sound better.

Hilly - not quite sure i follow you. The black iron is all gas - installed professionally. There is a drain on it, assuming all to code. Or am i missing you and that was a ref r nce point ?


This all said, we have a radiant contractor coming friday morning to review and recommend remediation. I think we are a but over our heads and really need to pull the ripcord. October in the northeast and we are running against the clock. Sincerely thought we had already paid for this level of support but clearly a case of "oh you can do it" and we bit. Thank you guys for the help so far. I am not personally done with this thread and would love to hear feedback once we hear more.

Again - thank you. If you have more thoughts - i am a student. But, i also dont want to detract your attention from others more deserving.

hot_rod

I suspect the heater is grossly oversized for that size home, unless you don't have doors or windows installed.

But it is possible that you can not flow that heater enough to get full output.

A heat load calculation is always the best first step to designing and sizing a system.

Hopefully a pro can get it sorted out with the least amount of $$ pain.

rick in Alaska

I just don't see how it can work at all with that piping arrangement. If the boiler wasn't a high head unit, you could remove the system pump, close the ball valve in the loop, and it would work fine. But, having to have the high head pump to overcome the boiler resistance, it needs to be piped primary /secondary. Right now, those pumps are confused as to what to do.
A hydraulic separator would probably be the easiest to fix without having to re-pipe it all.
And since you asked for it, I would have put all the zones on a manifold and just used one pump for the system. As Hot Rod implied, that is a lot of extra energy to use that is not needed.
Rick

raided

Interesting. The original design was a tank water heater without system pump and that primary loop was not connected (capped on each end instead making the turn ). They adapted that design when suggesting tankless and im wondering how much testing was done. To a relative layman, it seemed simple/elegant and functional.

DZoro

Agree with the above, if that is a side kick of a takagi. Then good luck, super high head. Must be done P/S and at least a 0011 pump would be the minimum. What size pipe is coming off of the pumps?

raided

The system pump is 26-99fc. Crossover chart shows equivalent to a 0011. Zone pumps are on 3/4", system pump is 1" but adapts to 3/4" on supply and return to the takagi.

hot_rod

Here is a pressure drop curve for a competing brand tankless. I have not found a curve for the Takagi brand you have?

As you can see in this curve compared to what your pump is capable of. I suspect 3-4 gpm is about what you can expect. No where near the 199,00 output that heater is rated at.

raided

We were seeing around 3 gpm on the unit. Does that mean we need a giant pump? I assume there going to be other physical limits that begin to factor (pipe sizes, etc.).

Giant pump with zone valves or continue with the zone circs? It sounds like maybe they are overkill and system pump is undersized.

hot_rod

raided said:

We were seeing around 3 gpm on the unit. Does that mean we need a giant pump? I assume there going to be other physical limits that begin to factor (pipe sizes, etc.).



Giant pump with zone valves or continue with the zone circs? It sounds like maybe they are overkill and system pump is undersized.

Therein lies the problem with tankless heaters trying to be a boilers
The type with restrictive heat exchangers, I'm assuming yours is, just will not give you the rated, or anywhere near the rated output.

You just can't pump them adequately. A pair of high head circs in series doubles head, but the water will be screaming thru those small HX tubes.

Tankless water heaters are designed to supply a low gpm across a wide delta T, from 50F to 120F for example, under occasional flows. The small tubes depend on the water pressure in your home, maybe 60 psi or more to overcome that pressure drop. Your circulator will not provide anywhere near that kind of delta P.

So it is completely opposite of what you want from a boiler.

You already have a big pump, the 26-99 is a high head model. It may or may not be adequate once you get it piped better with primary secondary or a separator.

raided

Gagh. That's kind of what i was thinking - limitations at the unit itself. The pro coming tomorrow said he doesnt prefer tankless but does see them work. He actually called out the radiant company by name when i explained the general scenario of why I was calling. I am hoping we can make it function with tankless but also realize limitations of square pegs and round holes.

Again - thank you for the advice!

Zman

Your biggest issue is the expansion tank location. The system will function if you run it as piped with the tank moved. Right now your pressures are looking something like this.

Zman

If you move the tank, it will be more like this

hot_rod

Zman said:

If you move the tank, it will be more like this

With the ball valve in which position? What is the pressure drop thru the tankless, that you used for the calculation?

Without knowing the heat load and how many rpm each loop needs we are all speculating to some extent.

Here is what I think happens with the ball valve in the loop left open.

With the ball valve closed the boiler circulator is in series with zone pumps. Possibly the first one to call on gets excessive flow and other get little, if any flow.

Since we don't know the required gpm, it could be the loop is undersized, even if piped correctly to supply design day loads.

I'd take a hack saw to the mess and pipe it in a known predictable way After you have a room by room, zone load calculation.

Send a copy of this journal to the company that sent that piping example.

https://www.caleffi.com/sites/default/files/file/idronics_15_na.pdf

raided

They didnt send an example - they sent that manifold prefab. It's the crux of their systems which is bad news for people like me.

We had eyes on this am and they propsed cutting out the supply/returns to the heater and piping in a proper secondary -closely spaced tees. And that bv to the left remaining closed. Also said that system pump is quite oversized - even on low - and to replace with a pump the same as the zone pumps. Seem fair? It sounds like with a secondary in there the system pump doesn't need to work as hard as the current would be fit for.

Ill try to dig up the numbers the online company said for heat load by zone.

modconwannabe

Homeowner here with radiant flooring as a zone in a 3-zone home--is there a reason not to buy a pre-fab manifold to deal with distribution to all those zones, maybe with zone valves, rather than the mess of Ts and piping and individual circ pumps? Not sure if that's useful, or more or less work.

hot_rod

The system, or boiler pump needs to be sized to the pressure drop of that tankless. I suspect it is close to what I sent a few posts ago. If so a 15-58 probably will not give you adequate flow? The only way to correctly size a circulator is have the required gpm, and the pressure drop of the circuit it is attached to.

The boiler, or primary pump just moves flow thru the boiler and the primary loop, period.

The zone pumps size to the load (gpm) they need to move. That is the whole concept of primary/ secondary piping, the two circuits are hydraulically isolated from one another.

Again pure guesstimate without have some, any, data. You might contact American or Takagi for the spec on that tankless, you need a chart like this. This is for a Rinai.

I'd get your ducks in a row before you make changes that may or may not fix or improve the system.

Zman

I thought the boiler loop circ is a 26-99

The rest is a swag. Just trying to demonstrate how much the expansion tank location affects the system.
I would never suggest piping a system this way. It is a common setup where I am and they do work, usually

Zman

I thought the boiler loop circ is a 26-99

The rest is a swag. Just trying to demonstrate how much the expansion tank location affects the system.
I would never suggest piping a system this way. It is a common setup where I am and they do work, usually

raided

They had figured at 11.25 gpm across all zones... Now im no math whiz but I do believe 10gpm < 11.25.

I found the chart for our model. Attached.

hot_rod

raided said:

They had figured at 11.25 gpm across all zones... Now im no math whiz but I do believe 10gpm < 11.25.



I found the chart for our model. Attached.

Where did they come up with 11.25 gpm? What's the plan to move 11.25 gpm thru the boiler, looking at the chart you attached.

raided

Maybe i am making assumptions. The heat load for each zone adds up to 11.25gpm (3, 3, 3, 1.5, .75).