Hot water system, rhythmically banging radiators -- Aka, my tenants want to kill me

mpls_apt_mgr

Hello everybody. I'm having a very frustrating issue with a hot water boiler heating system in a 4 story apartment building. Forgive me, I know some about the system but I am not an expert and do not know all of the proper terminology, but I'll do my best. The setup:



I don't know the specs of the boiler itself. It works great, though. The system has a single Wilo Stratos circulator pump, newly installed. The building is four stories. Building was built in the 1940s, radiators are largely originals. Each radiator has a Danfoss flow control valve.



I have exactly three radiators that are causing frequent, fairly loud repetitive banging noises. It's a constant knocking noise, sometimes more of a honking, in a perfect rhythm. Sometimes it's your standard 4/4 beat, and other times it goes all funky and does an interesting syncopated triplet rhythm or so. The noise can, without fail, be stopped by tweaking that Danfoss valve up or down, even just a little bit. Doesn't matter which way. Interestingly, it will also stop if the knob head of the valve is pressed in toward the actual valve assembly. It will resume upon release, though, and you can even vary the rhythm or tone by applying different pressures. The thing is, after anywhere from 20 minutes to 10 hours later, usually in the 30-90 minute range, it'll start right back up, even at the new valve position. Repeat. This wakes up the tenants living in the apartment the radiator is located in, or, even worse, if they're not home, it'll continue non-stop for hours and irritate half the building.



Two of the radiators are on the first above-ground floor, one is on the very top floor at the farthest position possible from the boiler, which is located in a sub level just below the parking lot, directly adjacent to the building. The building has something in the neighborhood of 56 radiators in it.



The system has had the air bled out of it very well. I don't get air out of any radiator, only water, both the noisy as well as noiseless radiators.



The pressure gauge at the fill valve for the radiator reads about 22 PSI. The head on the circulator is currently at 28, and can be changed at any time. This pump runs at a constant pressure and will alter its flow rate as necessary to maintain that pressure.



Sorry for the long post, I just wanted to provide as much information as possible, as my life may depend on it if I don't prevent a tenant revolt soon! Heh heh.

EricAune

How low can you set the Willo?

You said its a four story building, are the radiators fed off risers (vertical pipes originating at or near the boiler, supply and return)? Depending on actual measured pipe lengths, your head sounds a little high and that may be the cause of the noise as the flow is too high as well. 



Just doing some simple math (very simple as I have no idea the actual layout of your system) and the idea that your system consists of mostly vertical piping I come up with a head pressure under 10'.  Remember this is not gospel as I have no actual idea of how the system is built.  Do you have any type of schematic?

Paul Pollets

Radiator Banging

You say you have a Danfoss "flow control" valve on the radiators. Do you mean a thermostatic radiator valve?



If so, the radiators that are making noise have the valves on the return, rather than the supply.  The valve bodies are made to be placed on the supply of the radiator. 

mpls_apt_mgr

Making progress with your help!

Eric, I can set the Wilo at just about anything. It'll go down to 10. It's an electronically controlled pump with a little LCD screen and control dial. Unfortunately I don't have rapid access to the schematic. Basically, the actual setup is like this: Two directly adjacent buildings, one boiler is located centrally between the two and under them. The intake and output pipes are 2" copper. In this boiler room, the first branch off occurs which goes to each of the two buildings. There's approximately a 20 foot run to a central part of the building (it's longer than tall), and at that point there are risers that feed the four levels. One building is 3 stories, one is four. One of our boiler technicians seems to think it should be at 28 feet of head. I am absolutely no expert on this, just researching for the people that know how to fix it, but that seems REALLY high to me. Oh, something that may be important is that the actual system is, I believe, 3/4" pipe. Is that kind of the standard for a system like this? I'm just eyeballing it.



I set the pump to 18 because I was getting more noise at the higher setting, and it seems possibly dangerous. At 18, the people on the top floor are getting, the radiators are heating evenly, and when the bleed valve is opened, it *does* put water out. It's just kind of a dribble, but it's there. That's all I want, isn't it? Isn't the name of the game really creating just enough head pressure to get flow to the top?



Also, it appears that the pump is on the return side of the boiler. I assume this is normal.



It'd be nice to know what this pump should be set at.



I checked the radiators that are making the noise, and you are absolutely correct, Paul! I never really thought about the configuration in that way. Pretty much every radiator has the valve on the supply, but those few have it on the return side of the radiator. The contractor that maintains our system is not going to be happy to learn that he has to drain it down and reconfigure those, but could fixing that actually stop that noise?



I sincerely appreciate your help in this matter.

Gordan

Valves oriented the wrong way

The problem is not necessarily which side of the radiator the valves are on, it's which way they are facing. (Practically, though, because of the connections on the valve if it is on the "wrong" side it will be facing the wrong way.)



If you look at the cross-section of the valve, it is made to flow in a particular direction. Reversing the flow can cause it to cavitate.

Ironman

Setting The Head Too High...

Can actually cause the pump to cavitate. It can also lower the pressure on the return side because you're causing a higher pressure differential between supply and return. The fact that the pump's on the return complicates this more. It should be pumping away from the expansion tank, not toward it.



The purpose of the pump in a closed loop is not to create pressure in the system: that's the job of the fill valve. The pump's job is to create a pressure differential between the supply and return that is sufficient to induce the necessary flow.



Try setting the head lower as Eric suggested and see if you are still getting good heat at the radiators that are farthest away.

mpls_apt_mgr

System information clarifications

So... a little information as I've been able to figure it out by observation.



The system is heated by an old Kewanee Type M fire tube boiler. It's quite well maintained and in great shape, inside and out. For lack of a better description, it's somewhere around 5 and tall, maybe 6 feet long, rounded top, flat sides and ends.



The circulator pump is positioned such that the output of the pump is feeding into the input on the bottom back side of the boiler. The output on the boiler is dead center on the top and feeds out to the buildings.



So from what you're saying, the position of the pump is unusual?



I'll have the Danfoss RA-2000 valves placed in the proper orientation. Does anybody have much experience with this valve? Are there other quirks of note? The noise that is being made is like a ...honking noise. It can sound like anything from a ticking to a very loud honking. Quite odd! The property manager seems to think it's from air in the system, of the type that can be bled, but from my experience that's just not the case. You're saying it's cavitation at the valve? As I understand it, these valves do control the amount of flow at the extremes (all the way up or down), but in between have an internal sensor that opens or closes the valve to keep the room a a temperature approximated by the valve position. Is this right?



I think I can probably get the owner to stop telling me to have the head set so high if I let him know that it can cause pump cavitation that could ruin his new pump.



Again, I deeply appreciate all your help. I know I probably seem very undereducated regarding these things, and I sincerely appreciate your patience and explanations. You're helping answer some questions I've been working at finding answers for for a full year. Even the plumbing supply as well as heating system installation and maintenance guys I've been asking haven't had answers, and they've physically seen the system. Thanks again. I'm really glad to have found this site!

Paul Pollets

TRV flow

Re-piping the TRV's to the correct orientation will stop the noise. The arrow on the valve body should point towards the radiator inlet, on the supply. The higher the pump setting, the more noise the errant valves will make, until corrected.

EricAune

Paul has got it

I missed that the first time around and he's spot on.  I still think the head setting is too high.

Ironman

Pump Position

Having the pump on the return is not "unusual" at all, it's simply not the ideal way to do it. The pump has to work a lot harder when it's pumping towards the expansion tank, and not away from it. It's also making it more proned to cavitation if other issues exist in the system.



A crude illustration would be driving your car at 65mph into a 25mph head wind. The engine has to work harder and therefore would use more fuel. If the car turned around and had a 25mph tail wind at 65mph, the engine would work less and be more efficient. Putting the pump down stream of the expansion tank connection is like having the wind at your back.



Get Dan's book: "Pumping Away" under the "Shop" tab above.

mpls_apt_mgr

Book

I saw that book in the shop, definitely going to pick it up! May get some of the other stuff too. Looks fascinating and really well done! I have high hopes for getting these issues resolved that have plagued this building for the last year!

Unknown

Stratos Set Up

I agree that the Stratos might be set too high as well.



What I recommend on start-up is to leave the pump at it's factory default (differential pressure - variant, head set mid range) and run the system.



If all zones are getting heat, there are no complaints and your delta T is approximately 20 deg F (regardless of one or all zones calling and regardless if the zones are different sizes) that it is set properly.  I can say that we are finding about 80% of the pumps require the head setpoint be turned down believe it or not (that's how oversized our pumps are in the USA).



I noticed you have the pump set on constant pressure (delta PC).  Was there a reason to change this from delta PV (where the differential pressure increases as flow demand increases)?  This is a more common setting.



Drop me an E Mail of you have any questions on the set-up - I will be happy to help (steve.thompson@wilo-usa.com).