noisy slab sytem

brcook

i have some questions about excessive noise in my system.  i recently installed a  Grundfos up15-100.  this put the gpm at 3.7 which is well below the 5 i was shooting for.  I'm using a cheapo flow meter but i did test it thoroughly to establish the margin of error.  I'm relatively sure that number is accurate based on the delta value which is around 45 degrees to start with then after a couple hours drops to 30 (120 degree water going into the floor).  but then the noise makes me wonder if its moving much more than that.  however the house is slow to heat so this too confirms the lower gpm.  the circulator i was using before was an Armstrong astro 230 on the highest speed.  with that i was seeing just under 3 gpm but there was no noise.  and initially i had a Taco 007 which was under 2 gpm.   

BTW, I've got 5 loops with two of them at or just over 300' and the other three from 250-275'.  all are 1/2 pex but from manifold-heater-manifold is 3/4 pex and is approx. 6' total (with several 90s and a couple Ts for gauges, etc).

the noise is more of a growl than a hissing-type noise one would hear with a sprinkler running or the washer filling up.  i tried increasing the pressure to 20 psi (from 12) but this made no difference. 

i know now the system was not very well designed to begin with but I'm trying to get it working as well as possible (within budgetary reason). 

SWEI

Slab heat

really, really, really, really prefers constant circulation using ODR.

hot_rod

unusual pump selection

if you longest loop is 300 feet of 1/2" pex? that is a high head pump for high flow resistance applications, and drainback solar



What type of boiler are you trying to pump through? is it a tankless per chance?

brcook

re: slab heat

but what does that have to do w/ the noise? are you saying use the 'slower' pump and use continuous w/ ODP?

SWEI

The system was obvoiusly not properly designed

so before you start changing parts, be sure you fully understand what is needed.  A heat loss calculation together with the tubing layout (spacing, mostly) will tell you how many GPM at what temperature is required to heat the space.  That, combined with the head loss will tell you what pump(s) to use.  Your huge ∆T is the result of hot water hitting a cold slab.  If the water circulation is constant but the temperature is controlled by ODR, the space will be more comfortable and demand less of the boiler.  It can also reduce pumping requirements in many cases.

brcook

re: unusual pump selection

yes it is pretty 'hefty' for this application however i chose it based on input from Armstrong tech support. the 007 and the 230 were struggling putting the feet of head pretty high. and yes i am using a tankless but let's not turn this into a "those don't work" discussion 1)because what does that have to do w the noise and 2)the flow rate of the 230 was actually slightly higher w the tankless than through a 50g water heater. my question is why is there noise when the flowrate seems to be well below the noisy threshold.

SWEI

Making a tankless work with in-slab radiant

An ODR-controlled mixing valve, some hydraulic separation, and perhaps a buffer tank will do the trick.



What is the design day heat loss requirement?  How big is the tankless?

brcook

re: SWEI

yes it was poorly designed, I stated that in my original post.  again, though, my question is why the noise with under 4gpm?  and yes I have done a ton of research before making these changes. 

part of the 'problem' with this house is that once it gets warm it stays warm.  the floorplan is not conducive to the heat spreading/equalizing though (its a rancher with a great room in the middle and bedrooms on one end and bedroom and family on the other).  often the temp at bedtime (8pm) is low 70's, at least in the greatroom, and it isn't til 1am or so that it drops below 70.  with the 15-100 and new heater it'll be 68/69 when I leave for work at 430am (and the system just shut off).  the temp will rise some but when I get home its 70+ especially when my wife and kids have been home  awhile and the lights are on, the oven is on, etc but the system hasn't run all day.  so yes, by early the next morn the slab has cooled and there's the high delta.  none of this answers the question of why the noise at this flowrate and other than going back to the 230 is there anything that can be done about it?

SWEI

If you use ODR

The overheating will be reduced significantly, if not completely eliminated.  If you have a lot of solar gain, you can sometimes prevent that from overheating by careful location of the outdoor air sensor.  Running a high-mass emitter system without ODR is like trying to drive a car with the accelerator pedal bolted to the floor, using only the clutch to control your speed.  Now replace that car engine with one with one having 3-4 times the power and see how things go.  Adding ODR and hydraulic separation gives you a throttle and a transmission to work with.



If you don't know what your actual BTU requirements are, how can you know how much flow you need?  Sorry to sound like a broken record here, but you really do need to begin at the beginning before you start changing parts.

hot_rod

velocity noise?

does it sound like water rushing thru a pipe/ if so it caused by excessive velocity. ideally you want 2-4 FPS feet per second, flow.



By increasing the pump to a high head type, you are probably increasing the velocity, and noise.



In addition to noise the excessive velocity will cause wear in components and worse case pin holes in fittings.



There are ways to make a tankless behave, but it would require addition components like a separator or buffer tank.



it's tough to get the rated output from most tankless with a pumped flow.



Connected to a potable water system you have somewhere around 60 psi to push through the tankless. with a small or medium circ, maybe 20 psi.

brcook

re: hot rod

like I said it is not the same hissing type noise like when a sprinkler is running but definitely related to the water moving through the pipes.  at under 4 gpm split into 5 loops aren't I well below the 4'/second velocity? 

the system pressure is at 20 psi and at the current gpm the heater has no problem heating the water.  its set at 120 and will maintain that even when input temp is only low 70's.  I cranked it up all the way and the output temp went to just under 140 (advertised max temp). 

I read that maybe its cavitation but the noise is in the floor and the pump itself is quiet.  that's also when I increased the pressure to 20 from 12.

SWEI

Air bubbles?

Can you post a picture of the tankless, pumps, and pressure tank? 

brcook

re: SWEI

I've not read much about constant circ/ODR but it does make sense.  probably should start a new thread under ODR instead of going off on a tangent in this one. 

i came up with 50k btu which led to the 5gpm which led to going for more pump once i saw how little flow the 007 was providing and the lack of performance of the system in general (which obviously has some issues that cannot be changed now).  

with ODR and tankless what happens when the system calls for less or no heat?  does the water circulate back into the slab ahead of the heater (which would cause it to shut off) kind of like excess fuel pressure being diverted back to the gas tank in a car?   can i use the 230 or have to use a special circulator?   wouldn't i still need to get more gpm to get enough heat when it was calling for it?  which then leads us back to my current prob of not enough gpm and too much noise.  or do you think it could maintain even with going back to the 230 and an ODR

SWEI

50,000 BTU/hr

could most likely be run at a 15º ∆T, which would require 6.7 GPM.  Running a concrete floor at a 20ºF ∆T is a recipe for uneven heat, especially with long loops.



Heat loss calculations already include several safety factors.  If I had to guess, most of them come out 20-30% higher than the actual demand (assuming the correct outdoor design temp was used.)  Combine this with the fact that ODR with constant circ removes the need for a pickup factor and you can often get by with a fair amount less than they predict.



ODR with a non-modulating heat source (even though yours modulates its input, it puts out a fixed water temp) involves mixing of cool return water from the emitters with hot water from the heat source.  The mixing valve or injection pump is adjusted based on the outdoor air temp to deliver the right amount of BTUs to the slab.  Because it adjusts before the indoor air temp changes, it offsets the flywheel effect of thermal mass in the slab.  And yes, you may still need more GPM in order to move those BTUs into the slab.



If you can't afford to implement ODR, at least get a thermostat with PWM and a slab sensor.  Uponor and Tekmar come to mind here, there may be others.

brcook

noisy system

the line on the bottom is the return and the vertical on top (going out of the pic) comes from the water supply (valve is closed)

SWEI

Expansion tank

is on the inlet side of the pump?  And the outlet feeds all five loops in parallel?



300' of SDR9 PEX flowing 1 GPM with 80ºF water presents 10.8 feet of friction head.  At 120ºF the number drops to about 10 feet.  If your loop lengths are accurate and the embedded PEX did not get crushed or kinked during installation (or during the slab pour) it's probably the tankless HX.  What make and model is the tankless?

Snowmelt

Tankless

Have you tried getting al the air out of the system.

First thing I noticed was no isolation valves and pump is on wrong side of the water heater. On a small job like that you really don't need primery secondary, but I recommend doing that with a 009 pump and maybe go up to one inch on the primary. Then go with a delta t on the secondary. Use the tankless at 140 degree you may want to put a extra air separator on the secondary loop. That way you run between 3 - 5 gallon through the heat exchanger and the heating you set the pump for a 10 or 12 degree delta tee.

And please get the flex hose out of there and hard pipe everything. Make it look more professional.

brcook

re: expansion tank

yes on pump/expansion tank. i have experimented with closing the valves on each of the loops and combos of loops. with only the two longest loops open the gpm was 2. i am a little foggy on how the number of loops etc affects the head.



i came up w the same head figure you did and that's why i bought the Armstrong. but then its flow suggests a much higher head value. and like i said earlier, its gpm was actually through the tankless than through the big water heater.



tomorrow after work i'm going to bypass the tankless and see what happens gpm wise.

Harvey Ramer

Tankless

Looks like an EcoSmart electric tankless. It will have fairly large passageways in the heat exchanger to allow the insertion of electric elements. Large enough to entrap air and low enough water velocity that the circulator doesn't have a chance to purge out the air. Even worse, both tappings come out the bottom.



You'll have to give the tankless a good purge with DW pressure while disconnected from the radiant system. Easiest way would be to install a purge valve on the tankless return line.



Harvey