Hi all, I'm a new homeowner dealing with a hydronic heat system that services both radiant heat downstairs (trays installed over subfloor with wooden flooring on top) and forced air upstairs. The radiant floors are divided into three zones, with a single pump and zone valves controlling water delivery there. The boiler is an on-demand Munchkin type.

We've been having some trouble with the system since moving in. I'm learning about these systems as I go so please do question or correct me.

Initially, the radiant downstairs was working reasonably though some spots in the home had trouble coming up to set temps. However the upstairs system had zero heat when activated, and the supply and return pipes at the air handler remained cold. Previous owners did not run the upstairs heat often or at all it turns out. A service company came out and bled some air out of the water loop running up to the attic unit.

This got the upstairs unit pumping some hot water, but led to a second problem: the system would heat until all zones shut down, then the pressure in the system would drop close to zero, which in turn caused the boiler to shut down. Adding a little make-up water would bring the pressure back-up, and the cycle would repeat. Another new thing: psi would swing around a lot more, and hit 30 psi at the top end.

As I continued researching how these things work, I realized that there was no expansion tank installed. As I thought about this my going theory was that excessive air in the upstairs had allowed expansion previously, and now that we had eliminated it we were hitting excessive pressure levels and tripping the PRV in the boiler at 30 psi. A cup under the relief pipe proved that we were in fact losing water (about 2 cups with every cycle I described above).

I asked the service company about this and they contended that the radiant system PEX tubing should accommodate the expansion and that a tank wasn't necessary. This seems to contradict just about everything I can find online on the subject. The problem continued - and on their next visit the service company installed one without my asking them to do so. Now pressure in the system stays at approx. ~20 regardless of activity, without adding make up water.

But the latest problem: we were away when the tank was added, and came back to the house after about a week. The tstats had been left at 50, we walked into the home that was about 45 degrees or so. I didn’t think about it and turned up the thermostats. After running for about two hours we realized nothing was happening. Pressure was correct, boiler was running. Feeling around on various pipes in boiler closet no heat was detectable on the supply side of the radiant system at all, but the heat loop off the boiler was hot as it should have been. The two pumps in that closet were running. The boiler was not cycling much, tell me that not much cold water was coming in.

The next day I attempted to bleed and refill the radiant zones. I didn’t hit any noticeable air bubbles from the downstairs pipes. I purged the upstairs loop, and here I did see a lot of air come out. I’m assuming and hoping this came in while the expansion tank was being installed. Thinking I had the problem licked, I let the system run, but same issue.

I started to wonder next about the piping. The radiant side has a bypass between supply and return below the heat loop connection, and I started thinking that it was just pumping back its own return water. It had not seemed to do this before, but when the service company added the expansion tank, they tee’d it off the supply line for the radiant loop - so this was a change (more on that later).

Pursuing this, I started partially closing the valves on the radiant return lines to see if I could restrict the water flowing back in, and with a delicate adjustment suddenly hot water starting coming out of the heat loop and into the radiant. Open the return valves back up, the supply pipes got cold again, kind of proving my theory above. Left restricted, the floors very gradually warmed up.

I’ve diagrammed the piping and attached below. Referencing that picture, here are my questions:

1. My only hunch as to why this thing started behaving differently regarding water flow through the radiant: the expansion tank. Everywhere I can see references for these tanks, they are always shown on the return side, usually off the make-up water line, never on the supply side, where ours seems to be at point A. Is this a problem? My service company says no.
2. From every example I could find of radiant systems, the supply and returns sides generally never directly connect below the heat loop connection (point B), except maybe for a differential pressure bypass, but instead we have an open line (point C). Is this a problem? Should the connection at point B be eliminated, essentially extending the heat loop through to point C? Again, my service company assures me how it is now is how it should be piped. Note that the same company recently repiped the heat loop side when the previous boiler setup had failed, prior to my purchasing the house. I'm wondering if even when it worked it might not have been working as efficiently as it could have.
3. There is a manual purge valve upstairs but no air bleeds/eliminators. Given that this is the topmost point in the system, seems like a natural place for air to want to accumulate, should there not be some mechanism to trap/release air up there?
4. The tstats in the house are all standard. Given my setup, is lag a problem and should I consider a different thermostat setup downstairs?

Thanks in advance for your help!

No pump between B and C? Possibly a variable speed injection pump?

#1 your expansion tank is on the suction side of both pumps, it's fine for a location.
other points are probably not relevant other than the question about piping. FBBP hits on one point, absent a valve AT the "B" location you can close you could not get any flow from boiler circuit to radiant circuit without a pump. at least, no significant flow. if there IS a valve at "B" and your piping is accurate, you should be able to safely close it entirely so all flow from the on demand must route through your mark "C".

that should end your problems with heat delivery. this is a rather odd way to pipe a system but with a valve at B it should be functional now that you have the expansion tank.

You shouldn't have any more air problems once you get this round of air out, unless there is a leak in the system.

Not sure a valve between B (closely spaced T ) would help other than purging, it should remain open, other wise the primary loop for the air handler is obstructed.
I would suggest 2 things, a valve between C Closely spaced T's kept closed, or if mixing is required a Valve between C kept shut and a Honeywell 3 way valve on the legs between B and C ,
Simple solution is Valve between C Closely spaced T's kept closed.
A proper air scrubber on the primary loop may help as well, Taco 4900 or similar.
An injector on the "A" leg between B and C would work as well for mixing down.
Dan

Closing a valve at C is exactly the same as closing a valve at B, you're just changing which end of that pipe has closely spaced tees.

the problem is that we have an extra point of hydraulic separation here and no pump between. closing either end will have the same effect of matching the number of hydraulic separation points to the number of pumps involved.

I suppose you are right Rob, I would normally leave my primary loop un obstructed, but yes looking at it again, OK, Still prefer valve between C, or a mix injector.

Posted By NRT.Rob on 14 Jan 2012 02:42 PM
Closing a valve at C is exactly the same as closing a valve at B, you're just changing which end of that pipe has closely spaced tees.

the problem is that we have an extra point of hydraulic separation here and no pump between. closing either end will have the same effect of matching the number of hydraulic separation points to the number of pumps involved.

I must be missing something. Would a valve at either B or C forcing the flow through the secondary not run the risk of over heating the secondary when there is a higher heat demand on the air handler?
I'm assuming the air handler is wanting plus 110 degree heat so the primary loops output would have to meet that. With B closed and only one zone calling on secondary, I think that zone might easily do plus 90 degree. With C closed the 100 plus water would have to flow through that zone.
But hey maybe I'm missing something.
Bob

this system does not appear to do mixing at all. that circuit could be made into an injection mixing circuit if mixing is actually needed, as Dan noted, but as drawn it simply gets whatever temp the boiler is spitting out, perhaps minus a small drop if the air handler is running.

injection mixing would actually be a superior solution, IF NEEDED. but you'd have to know what water temps the air handler needs (likely it's set up to get something like 140 right now, but hey, maybe not) and the radiant needs. Ideally you could turn the air handler down to a temp the radiant needs but that's a bundle of snakes too... gets into port locations for air and such.

Thanks all for the input, you all are very helpful! So a few things:

• regarding a closed valve at B versus C: the up/down piping shown at C is a kit manifold (right word?), so not as easily modified. At point B it's standard 1/2" copper pipe so cutting here would be far more easily accomplished.
• regarding temperatures, the Munchkin is currently set to hold temps between 150-160 as it runs, and (at least right now given that we are seemingly not pumping all the cold water back in) it seems to be able to hold in that range easily. I don't know what the air handler needs or is set for. Is this temp too high for the wooden (laminate) floor on the radiant? I've read a bit about radiant overheating floors and damaging wood, particularly laminates.
• is a differential pressure bypass between the pumped supply side and the radiant return warranted? I've read that can extend the life of fittings through the system and reduce flow noise when only a single radiant zone is calling for heat. Flow noise is not noticeable under a single-zone condition today. And I don't see where implementing a closed valve at B changes this dynamic from what it is today... but the up/down manifolds in my pic have an extra connection I didn't show on both sides today that is unused.

If anyone has comments on my last question regarding standard thermostats given the flooring setup, happy to take those too.

The piping of your system is far from perfect. Yes a differential pressure bypass would be beneficial.The expansion tank and fill station should be repiped. Check valves are needed at the pumps.What is the piping sizes between B and C and also D.The pump located at D should be back at point B. The installation of a hydraulic seperator would also be in order.

The piping between B and C, and the heat loop off the boiler, is all 1/2 inch copper.

The piping for the kind of U shaped thing (which C and D are a part of) looks like 1 inch, though I have not measured it and I'm away from the home right now... it might be 3/4" - it all looks like part of a pre setup kit.  The connections towards the boiler were all redone post the original install when the first boiler setup was redone.

The pump in the heat loop has IFC in it's part number, so there is a check valve built into that pump, but the 2nd pump does not.

Found a picture, not the best, but might help. The piping in the foreground is for connection to/from the potable hot water and can be disregarded except where you can see the make-up line tee off.

Coming down from center are all the connections from the boiler, the straight run all the way down is for the PRV. The two pipes with the untaped and somewhat disformed insulation are the connection between B and C. Slightly cut off on the right, the primary heat loop is just below the outlet. The radiant valves and supply lines running out are cut off on the left.

Posted By wahwahwah on 14 Jan 2012 04:13 PM
Thanks all for the input, you all are very helpful! So a few things:

• regarding a closed valve at B versus C: the up/down piping shown at C is a kit manifold (right word?), so not as easily modified. At point B it's standard 1/2" copper pipe so cutting here would be far more easily accomplished.
• regarding temperatures, the Munchkin is currently set to hold temps between 150-160 as it runs, and (at least right now given that we are seemingly not pumping all the cold water back in) it seems to be able to hold in that range easily. I don't know what the air handler needs or is set for. Is this temp too high for the wooden (laminate) floor on the radiant? I've read a bit about radiant overheating floors and damaging wood, particularly laminates.
• is a differential pressure bypass between the pumped supply side and the radiant return warranted? I've read that can extend the life of fittings through the system and reduce flow noise when only a single radiant zone is calling for heat. Flow noise is not noticeable under a single-zone condition today. And I don't see where implementing a closed valve at B changes this dynamic from what it is today... but the up/down manifolds in my pic have an extra connection I didn't show on both sides today that is unused.

If anyone has comments on my last question regarding standard thermostats given the flooring setup, happy to take those too.

Think of your setup as two separate loops. In the primary loop the pump circulates the boiler water. The boiler is probably set to maintain a certain temp in this loop when there is a heat demand. When there is a heat demand from the air handler upstairs, the pump in the air handler starts to draw the 140 to 160º water from the primary loop and forces it through the coil in the handler and back to the primary loop to be reheated.

The secondary loop is the loop that the radiant comes off. When there is a heat demand from one or all the zone valves the pump in this loop starts to circulate the water in this loop through the zones. The reason for the second loop is that this temperature should usually not exceed 90º, usually kept in the mid 80's. Most hardwood and laminate warrantees are voided after this. Plus you would not want to walk on a floor that is much hotter than body temp. On the other hand the air coming from the air handler has to be considerably above body temps. or it will feel cold. So you need to maintain two different temperature zones. The only thing missing in you setup is some way of keeping the temperature in the secondary at that 80 to 90º temp. If you flow water constantly from the primary side you would run the risk of over heating the secondary.

As it is, the is no differential pressure between the two loops to cause a flow from one to the other, so you normally have no flow into the secondary. This is what Rob meant by "no flow". You have been able to create a slight difference by choking the valves on the secondary and that caused some flow between the two, hence the heat you managed to get into the floor loops.

What is needed is some way of running just enough hot water into the secondary to maintain the secondary loop at its target temperature. One way to do this is to use a Tekmar 356 controller and a variable speed pump that speeds up or slows down to maintain that target temp in the secondary. The pump would be located between the two loops.

I'm not sure this is the best one to use in your application but at least if you look at the literature at http://tekmarcontrols.com/hvacsystems/356.html you will get a better understanding of the process than what I am able to give you. The 356 does have an out door reset that will allow for lower temps in the secondary loop when the weather is warmer and will run the temps a bit hotter when the weather gets colder (like today -28ºC).
Also please understand that I am not a heating expert so when some of the true professionals on this forum correct me, weigh their words carefully ;-)> Bob

FBBP has nailed it.

Revelation!

FBBP, Rob, and everyone else who contributed, I'm very grateful for your time. FBBP's last post made all the bits & pieces I had gathered regarding injection and secondary systems (and design temps, and lag, and on and on..) fall into place and helped me relate them to my situation.

Now that I understand things a bit more, it looks almost as if someone had in mind to setup the kit on the left side as a secondary loop, but didn't do this properly when the portion on the right was redone.

I have to say, the more I learn about hydronic heat the more fascinated I become. I could easily foresee myself becoming addicted to tuning and the topic at large.