i just looked up "cavitation" and got this result:

The most common examples of this kind of wear are pump impellers and bends when a sudden change in the direction of liquid occurs.

interestingly, the way our system is designed, immediately above our circulating pumps, the supply line makes 2x 90˚ turns - a pretty dramatic change of direction! and also, when zone 1 "steals" from zones 2 and 3, the bends above the circulating pumps in zones 2 and 3 are RED HOT, but cold below.

i'm leaning toward this explanation - assuming that there is lower static pressure in the pumps than in the supply line from the Takagi. but not sure why it would only affect zones 2 and 3 but not zone 1. and i still want to know if raising the standing pressure might somehow reduce the cavitation? (just to connect the dots)

and also, if cavitation is the problem, but i still want to separate my domestic hot water from my radiant heat, will a heat exchanger work?

regretting not taking physics.

at your former 25 psi it's pretty hard to think you had a cavitation problem, that's twice as high as most closed systems run at. You'd need a lot of suction to cause cavitation there. But I guess with some confluence of factors it's POSSIBLE with 3 15-58 pumps running on high power with low flow rates... maybe. But you said the pumps are non-ferrous... can you provide more numbers on the pump model? You said UP15... UP15- what? 15-29? 15-58? are there multiple speeds on the pump and if so what speeds were selected?

I would normally say going primary/secondary would fix this problem, assuming it was just the pumps fighting each other for flow when the restriction through the takagi got too high. Since regardless of whether it's actually cavitating or not, going primary/secondary would have fixed it, as it would have removed the major restriction from the flow stream.

However you definitely should close off your heating system in either case. I would just be careful that you don't choose a restrictive heat exchanger.

The circulating pumps are Grundfos UP15-29su

Since I barely understand the concept of cavitation, it just seemed to make sense since there is a pressure change and a design with a radical bend immediately after the pumps. But I realize I could be grasping at straws. What I'd really like to know is why raising the standing pressure seemed to resolve the war between the circulating pumps, and whether switching to a heat exchanger system will work. We've been chasing windmills for two years trying to get this system to work right. Whatever we do next, I'd like to be relatively certain will finally fix it.

My biggest fear is that we will install the heat exchanger and then somehow the three zones will not work together. I'm not clear if a heat exchanger system would resolve a flow problem or a cavitation problem or whatever problem we have. If not, what else do we need to do? If yes, are there other considerations? ie how much should we pressurize the system? do we need a pump other than the 3x circulating pumps? anything else I'm not thinking of? I know that a heat exchanger system would require an air eliminator and expansion tank. Anything else?

Also, is there a specific heat exchanger you would recommend? (it seems to me that all heat exchangers are restrictive by design)

By the way, I did test the system and we didn't lose an ounce of pressure overnight so that seems to rule out a leak.

hmm. if they are 15-29s I don't think they could cavitate in a 25psi system. I have to admit I'm stumped on why static pressure would seem to help here.

can you take a picture?

i have lots of pictures! just can't post them.

think i figured out how to do it...

ah, you have a 3-way valve in there too for the radiant. looks like all 3 pumps are on it. between that and the takagi, that's a pretty massive amount of flow restriction. this should be HORRIBLE for your takagi as well from a short cycle perspective unless the radiant valve is set to the takagi's max temp or something.

this is really screwy piping for DHW though. where is the cold leg for the DHW out tempering valve, is that the valve next to your pressure gauge? why is it connected to the radiant circuit?

yep, we have a mixing valve to the radiant supply line. it is set to wide open (all the way HOT).

i think your next question is regarding the domestic mixing valve. the hot side of the valve is connected to the hot water outlet from the takagi. the cold side of the valve is connected to the cold water inlet that feeds the radiant system - it has the pressure gauge on it. that part of the system comes preassembled so i can't answer why it is designed like that! but i can definitely say that it is a problem because we have HOT water coming thru that side of the mixing valve!!! which is one of the reasons i want to separate the systems!!!!!!

i read up a bit on these systems. wow. my head hurts.

-I don't see how you can get hot to the DHW valve on the wrong leg for more than a very short period of time. is this a sustained issue when pulling hot water, or just an initial burst of hot water?

-your pressure booster is on that cold line, right? at least now I see how it might help flow in the zones in some cases... if you use any hot water, it blows water through your loops on the way to the takagi. I wonder if you did have an air bubble issue and this gave you the power to overcome it... I wonder if you leave the zones off for several days and then run them all, with NO hot water demand, whether you might see the original issue reappear temporarily?

You may want to install a few more pressure gauges - pressure drives flow.

Hey Rob,

I actually have not spent very much time troubleshooting the DHW valve issue so I'm not sure when it happens and if it is sustained or not. I only know for sure that in winter when the system was consistently in use, the pipe leading to the cold side of the DHW mixing valve was VERY hot to touch. The only troubleshooting I did re that valve was after we installed the booster pump and check valve and the system was not working right. I was trying different things and at one point I turned the Radiant mixing valve all the way to cold (feeding from the radiant return line) and the DHW mixing valve all the way to hot (feeding from the Takagi) and I was surprised to discover the DHW temperature plummet! I sent an email to the Radiant Floor Company describing the situation but never received a response. Since then I've left the radiant mixing valve on all the way hot (feeding from Takagi) and DHW mixing valve also all the way hot. Otherwise I've given up trying to figure it out. But it is definitely one of many reasons why we decided to separate the systems and be done with it!

The booster pump is on the cold line. I've never tested leaving the zones off for several days with no DHW water demand so I'm not sure what would happen. But after we installed the booster pump and later the check valve I purged air from the system multiple times because the system was acting so strangely. For example, after we installed the booster pump, the Takagi would not start in heating mode unless a DHW tap was opened. When the DHW tap was closed, the Takagi would eventually shut off despite pumps circulating and the return water temps at 80˚ and below. If I turned the booster pump off the Takagi would start when any / all pumps were circulating like it always had before. Oddly that problem seemed to resolve itself once we put the check valve in but I have no idea why. It is one of the many confounding problems we've encountered.

So, what now?

Jon - That's what I thought, too. But Rob has convinced me that "the PSI only helps flow in DOMESTIC USAGE mode."

i made a sketch of what i think our system would look like if we convert to heat exchanger system.

questions:
- does this look right?
- will it work? (given our previous issues with flow)
- does the expansion tank / air eliminator need to be on supply or can it be on return? (expansion tank is currently on return)
- anything missing?

the pipe to the DHW mixing valve may heat up from thermosiphoning or recirculation within the pipe when there is no flow. However, this would be overhwelmed by cold in in a short period of time, so at most you might get a very short slug of hot water followed by properly tempered water. I can't see any way you'd get sustained, radiant temperature water into the DHW out leg without seriously malfunctioning valves.

you can leave the zones off with no SPACE HEAT demand for several days, and then to test, simply don't run any DHW at a particular time and go turn up all 3 heating zones. If you don't run DHW your booster won't run. I wonder if you'd see the problem reappear... if so, it would be an air bubble problem (combined with your massive pressure drops) and your booster pump just gave you the oomph to clear them out regularly.

Your sketch looks pretty good. you'll need makeup water in the closed system as well, usually a backflow preventer and fill valve assembly is used. expansion tank on suction side of the pumps.

I'm inclined to not worry too much about DHW mixing valve - feel like it will be resolved by separating the heating from the DHW. If not, I'm hoping my plumber can figure it out! (but he is not familiar with radiant systems)

An air bubble problem sounds bad. I think you are suggesting that there might constantly be new air bubbles forming and that sounds like something that could be impossible to fix. Can you explain what causes it? In terms of testing, my heating is off all the time now. I turned on all three zones w/o DHW. Zone 1 and 2 heat up immediately, but Zone 3 (the smallest, upstairs) is lagging (only lukewarm). If air bubble problem, I assume switching to heat exchanger system won't help. So, once again, I'm at a loss.

Just to be clear, have we ruled out the possibility that raising the standing pressure corrected the flow and/or cavitation problem?

air comes in with fresh water... that would be resolved by closing the system. my theory at the moment (and i'm not solid on this... it's little shaky... but it's the best theory I've got so far) is:

1. Fresh water brings in fresh air, which may over time come out of solution and collect in the loops. when your zones operate independently, the bubbles may restrict flow somewhat but your small pumps can still circulate flow as resistance is low and no other pumps are really fighting for flow. Prior system pressure (25psi) may not have been adequate to clear these bubbles fully on DHW calls.

2. As more zones turn on though, flow increases through the mixing valve and takagi, which increases resistance of those items very quickly.

3. with all zones on, restriction in the valve/takagi from the increased flow makes the pumps "fight" each other, that plus air bubble resistance effectively knocks the two small pumps out of commission.

4. adding the boost pump means that whenever you call DHW you now get a power boost through the lines that clears the bubbles if any collect and maybe makes it harder for them to collect in the first place... that may make my "test" moot unless you could run the DHW system at 25 psi for awhile again to replicate the initial conditions.


I do think all your problems will be solved going closed, but I have to admit your existing problems are a bit flustering.

STANDING pressure could not have fixed your problem. but pressurized flow through the lines could have helped. just to be clear on that.

more than a bit flustering for us! ...it is totally aggravating.

but your insights have been incredibly helpful. you've offered much better explanations than the tech at the Radiant Floor Company and for the first time i have some confidence that the system will work if we go with a heat exchanger.

that leaves me with just a couple more questions:

- embarrassed to ask but when you say the expansion tank goes on the suction side, is that the supply line or return line?
- how to choose a heat exchanger that isn't restrictive?
- do we need to pressurize the radiant system? if so how much?
- assuming the back flow preventer / fill valve assembly goes on the cold water makeup and Ts into the radiant supply near the heat exchanger?

i'm going to have my plumber do the work as my husband has a new job and travels for work. i def can't sweat pipes!!!! while my plumber is not familiar with radiant systems, he is a licensed pro and will know the basics. i'll purchase all the components and sketch out a diagram for him. please let me know if there's anything else i'm missing?

(added revised sketch)

suction is usually the supply. the "bulb" on the bottom of the pump is the suction, and the pump body has an arrow on it pointing in the direction of flow.

-closed systems are typically pressurized to 12-15psi.

-yes on the fill valve.

-heat exchangers are not USUALLY that bad. but it should be set up to allow flushing (drains on both sides) in case of clogging.

(oops! didn't see your message - that all sounds good to me)

are these the right components?

air elminator:
http://www.pexsupply.com/Bell-Gossett-112105-1-NPT-Enhanced-Air-Separator-8607000-p
fill valve / backflow preventer:
http://www.pexsupply.com/Watts-0386461-1-2-9-11S-M3-Combination-1156-9D-Sweat-3688000-p
heat exchanger:
http://www.pexsupply.com/FlatPlate-DW10X20-14-14-Plate-10-x-20-Double-Wall-Vented-Heat-Exchanger

It takes quite a bit to make Rob's head hurt. What everyone should glean from this teachable moment is that "open" heating systems are not worth the trouble and tankless water heaters are not space heating boilers.

As to pressure, pumps and mixing valves. I have seen the original and the new drawing. Pressure, temperature and flow effect the amount and location of entrained air and where it comes out of solution. At 50lbs. static pressure, it is highly unlikely that air is coming out of solution at any condition you have described. I use tank-type water heaters as dual-purpose heat sources all the time in my local plumbing and heating business here in Prior Lake, MN and across N. America in my radiant floor designs.

The difference is one of isolation. A plate heat exchanger is a must in order to properly and legally separate space heating from potable water. It makes everything much simpler to install and to diagnose should you have a regular maintenance problem. The reason I asked you for a heat load is because this is where every space heating design starts. In your particular case, the plate heat exchanger must be sized for the design water temperature (determined by a proper heat load) and the load itself (how many btu per hour you need to transfer from water heater to space heating through the heat exchanger).

What have we learned? Don't buy radiant floor heating products from online vendors unless you have a proper heating design from which to order. If, in the rare case, the online radiant floor vendor does provide the heat load (rare) proceed with caution. In no case should those who advocate code-averse system be trusted to design a heating system for your dream home.

As for your piping diagram; size matters. First the HX, then the near-piping and finally the pumps. Look at the piping diagram that came with the Takagi and then design the radiant side using excepted practice. Finally, put the two together so that one does not effect the other hydraulically i.e. a plate HX. Air separation is only needed on the radiant side along with an expansion tank. If you don't have the potable water directly connected to the radiant system, you don't need a BFP.

And that is all I have to say about that...