- The temperature of the glycol coming out of the hot water tank is 30 degrees C. The temperature of the return glycol is 21 degrees C.
- The hot water tank is set at 160 degrees F.

160F to 86F - the heat exchanger in the tank is fouled (likely) or undersized. You could move the tankless to a direct configuration (ie, direct to the radiant loops). Or try to clean the heat exchanger (acid + inhibitor in the tank).

I design around 20 dt standard in all residential systems. works great. no one ever gets upset. been doing it for years on many hundreds of systems. off grid I'll push 30 or higher to get pump energy down.

Reynolds numbers, laminar flow, and 2ft/sec thresholds are all moot considerations in residential hydronics that are completely ignored without consequence. They result in grossly oversized hydraulics if followed. see my earlier comments about PE designed residential hydronic systems. Now you see where such issues routinely come from... the rules are different here.

Jonr: he has a mixing valve reducing temps in line.

Have a look at the alpha pump curve chart.
You describe a 3,000 sq ft house, 2 zones , no details on loop length or count per manifold so lets make assumptions.
3,000 sq ft/2 1,500 sq ft zones 1/2 pipe on a 12 inch pattern 300 foot loops (agreed this is an assumption)
1,500 feet of pipe is 5 x 300 feet head of pipe only will be about 7.75 feet, add some water flow per loop, minimum say .7 gpm, ideal my be more or less based on load, but this is hypothetical, right.
so 7.75 head flow 5x.7 =3.5 gpm.
Now go to your handy Alpa pump chart http://media.blueridgecompany.com/documents/Grundfos_Alpha%20Install.pdf p. 15
You will see you are no where close on low. Proper speed is 2 or medium. on low you are barley moving water.
I also would suggest increasing delivered water temp to around 100f and look for a delta T as Rob suggested in the 15-20 degree spread.
Also not sure how raising the water heater to 160 degrees has an affect other than if you are capturing solar heat and this is the cut off. Seems excessive if you are using the flash water heater to reach this, and may be contributing to scaling, not sure about that, please correct me on this if wrong.
Short answer, turn up the pump a notch, make more electricity.
Dan

at 33 watts he's either on fixed speed 2 or constant pressure 1. constant pressure 1 would give him slightly less head than you're figuring but plenty of flow.

fixed speed 1 is a max 8 watt draw. 3 GPM/3Ft head or so. I use that setting all the time in off grid homes and I'd recommend trying to make it work here... 16 watts vs 66 watts is a big deal when you're off the grid. but you need steep dts and a higher supply temp to make that work.

Posted By NRT.Rob on 26 Dec 2012 12:40 PM
Reynolds numbers, laminar flow, and 2ft/sec thresholds are all moot considerations in residential hydronics that are completely ignored without consequence. They result in grossly oversized hydraulics if followed. see my earlier comments about PE designed residential hydronic systems. Now you see where such issues routinely come from... the rules are different here.

This is classic and timely as I can actually use this in an article I am currently writting...thank you!

Besides the fact that maintaining turbulent flow often involves UNDERSIZING the hydraulics (by using a reduced PEX diameter if necessary) it isn't very often that someone in the HVAC business actually publically states in writing that they believe fluid dynamic and thermal dynamic engineering principals are moot and can be ignored without consequence for residential hydronic heating design.

I bet you operate from a state that doesn't have any HVAC licensing requirements.  If so you are correct in stating that "the rules" are indeed different for you...caveat emptor!

My secret is the tiny little notes we tape to the inside of the pipes politely asking the air to please stay entrained. It must work, or I'm damn lucky. knock on wood, AM I RITE?

More seriously, have you ever even taken a field trip to a working radiant system? If so, have you ever seen one using a 1.2 GPM/loop flow minimum on 1/2" PEX? If they haven't all done so, please tell how all those systems continue operating under 2ft/sec in all those loops. Maybe you should call around a bit and see if you can find ANYONE designing to that minimum flow before you embarrass yourself in an article. 2ft/sec is a great MAXIMUM for hydraulic separation in common piping, but it's a meaningless threshold in a 1/2" pex loop field.

if you just bother to google a little (or even check a few major MFG's design manuals) you can find some support for 1.5ft/sec minimums. I have and will continue to go on record saying I think even that is a waste, and that laminar flow in residential radiant systems appears to be a bogeyman that no one I have ever spoken to has ever seen in person (which would manifest as underperformance and a low loop delta-T). the reasons why I could leave to an ASHRAE fellow to investigate. I would speculate it probably has something to do with the serpentine nature of most of our piping causing the boundary layer to be disrupted "regularly enough" that any performance hit is minimal. but that's speculation: what is fact is that I have tens of thousands of 0.2 GPM loops in operation today and performing well. a bunch right here in my own shop, heating with 95 degree water and cooling with 55 in the summer.

I guess somehow I've stumbled on some magic way to make laminar, airbound loops heat properly. whatever it is, I can be pretty sure that my low temperature, low energy systems would not be improved by multiplying their flow rates by 3 or 4, typical.

Our overreaching friend does make an excellent argument for using experienced designers instead of "professional" engineers whom lack the specific training and experience in the scale we work with every day. It is basic thermal dynamics but science can turn in to "junk" in a nano-second. Frankly my money is on Rob, whom has clearly forgotten more about radiant panels than you can Google. We share a minimalist view of hydronic radiant systems, often shunning the use of multiple pump for one. This is particularly noteworthy since Rob sells pumps! I too have thousands, maybe millions, of feet of PB and PEX running in Reynolds numbers well below the "minimum" 2000. In fact several cast iron gravity conversions run so low the flow meters do not move!

Anyone who has experienced unintended gravity flow knows that flow is flow. The only time that turbulence is a factor is in engineered heat transfer devices such as a plate heat exchanger of low-mass water-tube boiler, where Reynolds number and laminar flow will result in lower-than-published outputs. As Rob quite correctly points out, following minimum design flow recommendations touting 2 fps is a waste of time and parasitic pump wattage. No one wants to burn more electricity, to pump more water, to keep those nasty bubbles entrained...and they don't have to, since temperature and pressure out-weight flow, in most systems, in this minor scale. Since radiant floor panels are typically over-sized for the area they must serve, the very idea of turbulent flow is foreign to the experienced radiant floor heating designer.

So, please mention me in your article.

PS. I work in a State and a City "Minneapolis" where you need a license for nearly everything including; radiant floor heating. They call it, Master of Steam and Hot Water, and I hold it.

To quote BlueLagoon;
Statistics
- With the pumps running at 15W each, there is about 1.5 - 2 gpm flow. With the pumps at 33W each, the flow rises to about 2 - 2.5 gpm. The temperature of the glycol coming out of the hot water tank does not seem to change whether the pumps are at 15W or 33W.
- The temperature of the glycol coming out of the hot water tank is 30 degrees C. The temperature of the return glycol is 21 degrees C.
- The hot water tank is set at 160 degrees F. The hot water tank does get up to temperature: I have three different readouts for temperature and recently the bottom one read 50C, the middle one read 65C and the top one read 68C. Just as I took the reading, the tankless hot water heater came on to heat the tank a further 6C before turning off.
- The radiant floor closed loop is at the top of the tank.

He is pumping at 15 watt, not enough,
At .5 gpm head on 300 feet of pipe with 30% glycol is 4.23 read head 4.23
Alpha head at 15 watt tops at 3 feet of head 1.5 gpm and declining 1.5 gpm divided by probably 5 300' loops = .03 gpm per loop, aint much, just say-in...

flip it to II position head goes to 7.75, flip it to auto adapt and you land right in the sweet spot on Alpha pump ann your building gets warmer!

on low speed let the debate continue... . . . .
And work with the assisted ghosting/ convection system.

Cheers
Dan

Why are you using glycol in that system? Is that necessary in the radiant floor loop?

I also don't understand why the hot water tank is so stratified. When the tankless heater runs, don't all the readings equilibrate fairly well?

dan, 15 watt is not fixed speed 1. fixed speed 1 tops out at 8 watts. both published and observed maximums. you have to be on fixed 2 , constant 1 or 2, or autoadapt to hit 15w. and I NEVER recommend autoadapt... horrible setting. very hard to predict what it's going to do and especially if you use reset water temps it's unlikely to do what you want. constant pressure for typical systems, fixed 1 for ultra low energy systems.

you got me on the temps though, I spaced it on the tank temp (160 F) not being 30 C. herp derp. thanks for that. silly conversions. I will go sit in the corner on that one...

30C outgoing to the radiant seems way too low for those floor coverings. turn up mixing valve and IF the water temp does in fact rise with the mixing valve reset that should fix any underperformance issues. 30C is an appropriate number for uncovered slabs, or maybe wood floors in very, very low load conditions, but not carpet and not typically in a home with wood floors over a thin pour of concrete. note the use of the word "typically".

ICF: he's off grid. glycol is a must in that situation if his climate sees freezing temperatures.

I know, but the need for glycol is one thing if he's up at Whistler, and another if he's down on Lasqueti. The glycol in there represents a substantial hit in pumping energy and that seems to be what he is after.

What he needs to do first of all is to drain that tank and get an inspection camera on the heat exchange coils to see what they look like. At this point that will yield the most definitive evidence.

I disagree. turning up the mixing valve and seeing if the water temp rises should be plenty of evidence

Hi Rob,
My point was some where in the post Blue Lagoon was saying that he only wanted to use speed 1, low because of his home made power issues.
He had run on higher speed with the benefit of heat transfer as well but was committed to low (1).
In my mind reading the pump curves there is the issue especially given the low water temps sent out to the slab. It aint a gona work,
Increase delivered h20 to 100 degrees, pump to 2 and problem solved.
Dan

Yes, I agree he should attempt to use speed fixed 1. if he raises his supply temp further, and is willing to run a high delta-T, it can work. no reason not to try.

A number of the suggestions made seem to belie the fact that the system used to perform brilliantly. For example, someone noted that the heat exchangers may be inadequately sized, but, if that's the case, why would they have worked previously?

Presumably, the mixing valve is still set the same as it was or he (heh-heh) "may" have mentioned the change. And, he already noted that increasing flow rate does not appear to change the temperature of the glycol exiting the tank coil at 30C. If the mixing valve was munged up and delivering low temp to the floors, wouldn't we see an abnormally low return temp? 21C seems in the range of reasonable.

I agree, if there actually are two mixing valves (one on the glycol loop and another on domestic hot water), then turn up (or fix) the former.

Hi everyone,

Wow! Thanks for all the responses. I must admit I feel confused by many of them. Let me try to go through and address them one by one.

Extra Data
I forgot to mention a few things.

1) I mentioned a mixing valve was on the glycol radiant floor loop just after the closed loop that is inside the hot water tank. This is true. But I failed to mention that the mixing valve is set all the way to MAX. In other words, the glycol is as hot as it can get with the current environment.

2) I mentioned there are two zones. Each zone has one Grundfos Alpha pump. So if we trace the system from the hot water tank…inside the hot water tank there is an enclosed loop which has glycol in it. The glycol absorbs heat from the hot water tank. It exits the tank, goes through the mixing valve (which is set at Max), then splits into two pipes, one for each zone. Each pipe has a Grundfos Alpha on it to push the glycol forward. The two pipes then goes into a manifold which splits the glycol out to loops of pex piping. I know that one zone is slightly smaller than the other and therefore has a smaller manifold but other than that I do not know how long the runs of pex are unfortunately.

3) The hot water tank and the Grundfos Alpha pumps are in a plumbing room that is below my house. The manifolds and pex piping are at ground level. So the hot water tank and pmps are about 10 feet BELOW the manifolds.

4) The temperatures that I have given for radiant floor glycol temperature - supply and return - are estimates taken via strap-on thermometers. The radiant floor system was not designed/supplied with any temperature gauges unfortunately.

Responses to Comments

@sailawayrb: Thanks for all the comments about what a good designer would do. I'm pretty clear that I did not have good designers and also did not have good contractors. Unfortunately I do not have photos of the pex tubing, nor do I have lengths of tubing. I also don't have any original temperature measurements; I only just put the strap on thermometers to estimate glycol temperature last month.

It sounds like based on your explanation that the return glycol temperature represents the maximum temperature that my house could get heated to. Is that right? Already that is more information than I knew before.

@ICFHybrid:

Yes - hard water is what I meant, not heavy water!! haha! We have well water.

The radiant floors are on one tank loop. The other loop is the solar hot water panels. The tankless heater is a Takagi TK-3-Pro and is set up basically a pipe goes from the hot water tank to the takagi and back. Another Grundfos Alpha pump pumps water from the hot water tank to the tankless heater and back into the hot water tank when an aqua stat calls for the pump to turn on. Very simple.

@sailawayrb:

My floor coverings have not changed at all in 6 years.

@jonr:

Yes, 160F to 86F temperature drop. This is why I have wondered if the heat exchanger is coated with hard scale or something. It's really the only thing that possibly changed besides the change of pumps.

@Blueridge company & NRT.Rob:

I think the 33 watt setting is a constant pressure setting and the 15 watt is a fixed speed setting. I haven't seen how to get down to 8 watts. One of my main goals is to minimize power draw. Would love to figure out how to adjust the system so that I can use 15 watts each…and 8 would be even better!

Also, does the fact that I forgot to mention earlier about the hot water tank and pumps being 10 feet below the pex pipes - does that change things?

@ICFHybrid:

Not sure why the installers elected to use glycol. We do get occasional freezing and I always assumed that was why.

Would switching to water from glycol help me at all?

You mentioned draining the hot water tank and getting an inspection camera on the heat exchanger coils. yay! Finally someone who seems to know something about this. No plumber i have talked to has ever mentioned doing this or knowing how to do this. So…any recommendations in the Victoria area?


Questions

1) Would raising the temperature of the hot water tank even higher than 160F make a difference?

2) Is having the hot water tank at 160F vs. 120F costing me more propane to heat it? I would honestly rather have it at 120F but only because I imagine it is costing me more propane to heat…

3) How could I set the system up to use the least amount of power for pumping? Some of you have suggested I could run the Grundfos Alpha at 8W each which would be like a dream…

4) How likely is it that scale in the hot water tank is causing me loss of efficiency? And if so, how to de-scale it? (I provided a link to the tank specs earlier.)


Thank you

I have spoken to so many plumbers, spent tons of money on plumbers coming out to this remote island, even posted on other green building forums but never gotten so much help from people who obviously know what they are talking about. Thanks so much!!

1) slightly but not recommended
2) more tank losses in the summer and reduced tankless efficiency - neither is big $.
3) higher supply temps allow lower flows. So does mild weather.
4) almost certainly, I'd check with the tank manufacturer for cleaning recommendations

I would wonder if maybe the mixing valve is installed improperly? this could perform better at cooler tank temperatures and get worse as you turn up the tank temp, so if you were ok, and then turned it up it would stop performing properly. would be kind of a rookie mistake for a pro to overlook, but also would not be the first time that occurred. '
you want: hot in leg from tank, cold in from radiant return, and you want to pull out of the mix leg. those should all be marked on the valve.

As for the pump settings, the alpha has a setting denoted with a single vertical line. this is fixed speed 1, which varies from 5 to 8 watts of power. system height doesn't matter... this is a closed system. falling water pulls up rising water. I would use both pumps on this setting and raise supply temps until I either got the results I wanted or found that the floor temps were just too uneven for my taste... or until I hit about 130 supply temp max. of course you have to figure out how to get this supply temp to rise!

you do not want to keep the tank at 160. drop it to 120 (50C) and see what happens to your radiant supply temps. if the valve is incorrectly installed you might even see the supply temps rise to the tank temp.

Ha Ha. Are you on Salt Springs or Lasquite?

You may get freezing temps out there occasionally, but unless your home goes unattended for long periods in the winter, during which time you could run out of electrical power and/or fuel, it might be a good idea to reconsider the glycol. You would still use it in your solar loop, but I think you might be able to dispense with it in your floor heating loop. It would have to get really cold for long periods to allow the slabs to freeze. Glycol is about 10% LESS efficient in transferring heat and more viscous, so at low temps, your pumps work maybe 25% harder to pump through long loops. Put it together, and you might be able to do with as much as 1/3 less pumping power to get the same job done. That is something you should consider whatever happens with the other issues.

Inspecting the tank will consist of shutting down the heating system and disconnecting one of the feeds or outlets, preferably at about the midpoint of the tank. You, or your plumber, should be able to use a borescope style inspection camera to look at each set of coils, top and bottom for scaling. You will also want to look and see if debris is building up in the bottom of the tank. If you feel comfortable undoing the plumbing, Costco has been selling a 9mm "Whistler borescope" inspection camera at about $150 that would suffice for a DIY job. You can also get it from Amazon for $130 - $190.