as the bottom of the slab heated up

Okay, but wouldn't the heat flow out the bottom of the slab increase, too, as it got warmer down there? How hot does it have to get down there in order for the surface to radiate enough to warm the interior of that big building to 58F when it's in the teens out?

Consequently, I'm still wondering about how leaky that building is. If it leaked even one air change in an hour, that might take up, what, a third the total output of the boiler?

Well, the floor isn't heating the building unless the floor surface temp becomes greater than the indoor air temp...

I wonder what the outdoor temp profile has been during this "experiment"...and if there has been significant solar heat gain that is heating the building...

Interesting question, but, it seems that the only place solar gain is coming through that heavily insulated thing in winter is through the "insulated" doors. I assumed there were no windows, and what's the chance they are on the South? And, since they got mentioned, I'm wondering if the 2" foundation insulation created a thermal break at those big doors. Or not.

Right, and just like that other post, I hope the OP doesn't freeze to death before we get answers to all our many questions... There is nothing better than a Christmas puzzle to share...LOL!

Ok.. let me try an answer alot of these questions. I have not had a heat loss calculation done. The building is 6 months old and its very first winter. I do not know the exact temp of the slab when I fired the system up.. but I can tell you it was very cold. Prob in the 30s. The temp gauge on the return (and feed for that matter) only starts out at 70 and it was practically an inch below that. It ran for 2 straight days getting the building up to temp and the return line temp to get to the 70 mark. I had the stat set at 70 degrees and the building got up to it in the 2 days.. but as I said it wasnt real cold out on those 2 days. It was a warm spell for a few days at 50 degrees outside.

My flow meters are at midway down the individual glass meter at 0.5 I can raise or lower each circuit to pretty much anywhere I want on the meters by manually turning them up or down. The black box is an expansion tank. The only box on the panel is the one the thermostat is wired into. It has both pumps wired into it as well. This site will only allow you to upload a pict if it is made very small.

Now.. with that being said. There is what looks to a mixing valve on this panel that has a "critical" sticker on it that reads to be in the open position when in use. Only closed when purging/filling the system. I am going to add a pict of this valve. I think it is directly what is causing this 50 degree difference on the return because it is resending half the 70 degree water directly back into the system to almost like recirculate. If this valve is closed it will only send the return back to the boiler.. not letting it halfway bleed into the feed side of the system.

To answer the question where is the pex at.. It is close to the top of the board. It is mounted/zip tied to #10 wire 5 ft x 5 ft sections for more concrete strength. Another case of a few people saying to manually pull up the pex and mesh up during the pour to get the pex more in the middle of the concrete. This was going to be about a fiasco and about impossible to try and do during the pour. I watched what seemed like 100 youtube videos of them pouring concrete over the tubing that is directly clipped/mounted to the foam board.

Side note.. the temp here is 40 degrees and cloudy.. the system hasn't ran once all day long and is actually riding at 61.5 with a stat setting of 60.

Pict of valve

well there goes my flow theory. thanks, man, thanks a lot!

So you have 4 thermometers here. Where are you reading 125 and 70, on the ones at the boiler? What are the two on the manifolds reading? You certainly are not pushing 5 gpm with a 50 degree Dt... that's 3x the output this boiler is capable of. Maybe you have a flow issue on the pump on the boiler...

That valve should remain with the handle open. this is primary/secondary piping. no problem. You're injecting heat from the boiler into the radiant loop, that's so the flow through the radiant is independant of flow through the boiler.

Other than the delta T question, it sounds like the issue you're dealing with is that the building was cold once when it was cold out (58 instead of 60 apparently). If that's the case that could just be normal thermostat swing?

Perhaps you would be willing to at least take a WAG at the building heat loss:

Borst Heat Loss Analysis Software

Yes, leave that “critical” valve open. BTW, your plumbing looks well done!

Yes, HR installation contractors just place the PEX using methods that they are comfortable using. Most HR contractors likely just staple it to the under-slab insulation…fast and low risk… However, as is often the case, the mob mentality isn’t always the best thing…at least not for the HR customer. The HR contractors who place the PEX higher up in the slab likely more often accomplish this by using chairs. Some will pull the mesh and PEX up during the poor. Obviously, placing the PEX on top of the finished slab would provide the maximum heat transfer into the room and would result in the maximum efficiency HR system. So anything lower than this is less heat transfer and less efficiency. Most HR contractors don’t worry about such things and that's neither here or there...just how it is.

Maybe I missed it, but what is the area of this 6” slab? Are you saying the slab is now at 61.5F or is the room at 61.5F?

You are saying that you have ten (10) 300’ circuits each being fed 0.5 GPM or a total of 5 GPM, correct? That's hard to believe for the reason Rob indicated...

Sorry to ruin your flow theory..lol

Yes, I have the 4 thermometers.. the out going one from the boiler reads just under the set point of the boiler. Its set to 125 and its pretty close to that. The in coming thermometer on the inlet side hot side manifold is a little over a hundred. You can feel the lines going in the floor are no where near the 125 temp the upper pipe is coming out of the boiler.. but this is after going past that mixing valve I refered ro earlier that I think is feeding alot of the cold water from the floor back in. The outlet temp on the manifold cold thermometer is 62 as is the thermometer going up to the inlet side of the boiler. The temp of the building as I said before also is set at 60 currently and it is riding about 2 degrees warmer but again its 40 outside.

Thank you for the compliment on the plumbing. I put the whole system together myself. I'm pretty picky about how I do things and I never cut any corners. I was the contractor on this radiant floor heat install job 100%.

I understand completely what you are saying about the pex height and the efficiency loss.. my thinking at the time was its all going to get heated/hot at some point and would balance itself out once the entire concrete slab was up to operating temp. The building is 64x48 with 6 inches minimum of concrete.

The thermostat on the wall is set to 60.. the temp its showing me is 61.5 Remember this system has been basically idle today because its so warm. Yes I have 10 circuits 300 ft each with the glass flow meters showing me 0.5 of flow. They drop about half way down the sight glass when the system is fired. Id like to get an infared point and shoot thermometer and see exactly what this floor temp surface really is..

It is always great to hear about a successful DIY HR design/installation project...congratulations! However, I am concerned that you didn’t first start with a good heat loss analysis before designing the HR system. While this isn't an indoor space where barefoot comfort is a real concern, one should always determine how much heat is required BEFORE constructing the heating system. Sometimes we get lucky...unless we conclude that your heat source is significantly under-sized...

If you are maintaining 62F indoor temp at 40F outdoor temp, your system is working...at least somewhat... It sounds like you may have solved your problem yourself and that your mixing valve setting may be the problem. I would guess that the required circuit supply temp should be somewhere between 100F - 110F and it sounds like you are providing nowhere near that temp right now…which would explain why the 5 GPM total is not unreasonable too.

We also have HR design software on our website should you want to accomplish a POST HR design which will tell you the required floor surface temp, supply temp, flow rates, etc.:

Borst Hydronic Radiant Floor Heating Design Software

Thank you. I had no idea of doing a heat loss analysis.. You talk to 10 people in this business and you will get 10 different answers on just about every aspect of a system it seems like. Ive never seen anything like it.

This mixing valve isnt really adjustable.. so I don't know what to do about that and you said to leave it in the on position as it is. So just for piece of mind. What should the actual floor surface temp be in relation to everything?

There should only be a 25 degree or so differance between the feed and the return line temp?

So are we ok and comfortable with the 40% glycol content? I am going to try the higher speed setting mentioned on the pumps as well when the temp drops back down and the system starts to work again.

I will try and do the post HR design and see what it tells me. I really appreciate all your guys input.

I turned the temp up to fire the system and decided to close that critical flow valve for a minute.. it almost immediately brought the temp of the inlet water manifold up to exactly what it's leaving the boiler at (125) I don't understand how leaving that valve open its ever going to let this floor heat up when it keeps reblending a big part of the water coming out of the floor right back into the floor. It does raise the system pressure about 2 psi when I close it as well to about 22 psi or so.

Oh, is this your primary/secondary bypass valve? If you have a manual bypass valve, then it should be somewhere between partially open and partially closed for exactly the effect you are seeing. It's purpose is to ensure you don't dead head the pump...which is hard on the pump.

We use a differential pressure bypass valve in our designs (i.e., in lieu of a manual valve) to reduce the potential for hydronic fluid flow noise when only one or two zones are active and to prevent pump dead heading if all the zone circuits were somehow closed at the same time. This eliminates any need to “adjust” a valve. This also increases efficiency by not always bypassing, especially true for mod/con boilers that prefer low return temp. There was a similar post about this subject not too long ago:

Radiant Set-up Questions

Like I indicated in this other post, I would much prefer to work from a real schematic (i.e., prefer to actually first design the HR system, create the schematic and then accomplish installation in accordance with the schematic) rather than try to figure out the intended design from someone’s maze of plumbing…sort of like not wanting to use someone else’s tooth brush…LOL!

The required floor surface temp can vary significantly. For a well-insulated, well-sealed building with a well-designed, highly efficient slab-on-grade HR emitter it may be as low as 72F and only require a supply temp of 85F. For other less-insulated, less-sealed buildings and less-efficient HR emitter designs (e.g., under-floor, above-floor plate systems, hardwood/carpet finished, etc.), it could be as high as 85F and require a 110F+ supply temp. Again, it all comes down to how much upward heat flux you need to replace the building heat loss and the efficiency of your HR emitter. This is why one needs to accomplish a heat loss analysis BEFORE designing/installing the HR system.

The difference between the supply temp and return temp is called the allowed circuit temp drop or delta T. For a residence, we normally design for 15F delta T or less to provide barefoot comfort (i.e., in bare feet you can't sense any temp difference in the floor). We also don't exceed 12" for the PEX spacing for the same reason. However, for your building, you could certainly go more than 15F for your delta T. I think 25-30F might be a reasonable delta T target.

As I indicated previously, we don't recommend using any propylene glycol unless there is a significant freezing risk that must be addressed. And just so we are 100% clear, you should NEVER use ethanol glycol in a HR system! There are also concerns about using propylene glycol in gas fired HR systems because it is considered flammable should it ever leak in the presence of flame. If you elect to use propylene glycol, the general recommendation is to use at least 30% to inhibit the growth of bacteria in the system...and I wouldn't exceed 30% unless I knew my PEX would absolutely freeze tomorrow without doing so... But then again, my PEX is fully encased in good concrete and it is unlikely to fail even if it does freeze.

OK, here’s the promised slab heat up profile…

The quickest we could heat up this 6” x 64’ x 48’ slab with R22 under-slab insulation would be by applying the full heat source 81K BTU/H. Doing this while pumping at 5 GPM would result in 32.4F delta T (i.e., the difference between the circuit supply/return temps and delta T equals 81K BTU/H divided by 5 GPM divided by 500). We have to assume an initial slab temp and indoor room temp. Let’s assume 50F and 60F, respectively. I also assumed 150 PCF, 0.18 BTU/Deg F-Lb, and 1.56 BTU/Deg F-H-SF for the concrete density, specific heat capacity and convection heat transfer coefficient, respectively. Here’s the hour-by-hour slab temp heat up profile:

0: 50.0F
1: 52.0F
2: 53.9F
3: 55.9F
4: 57.8F
5: 59.8F
6: 61.6F (the room will start getting warmer)
7: 63.0F
8: 64.3F
9: 66.2F
10: 67.2F
11: 69.2F
12: 71.1F
13: 73.1F
14: 75.0F

So about a 1.8F rise in slab temp per hour would be the quickest you could heat up this slab with this heat source. However, it is highly unlikely that you will be able to apply the full heat source BTU/H, so your actual slab heat up time will be somewhat longer than this.

Wow thank you for all the info you provided! You supplied a wealth of it! I think my start up ran pretty close to your rise numbers only my concrete started out in the 30s.


Let me ask you this. I know my tubing is incased well in the concrete and also we know the insulation is def up to speed below it. Would you recommend pulling all the glycol contents out? Atleast lower it to 30% ratio maybe? Or do you feel I'm not losing enough heat transfer to justify messing with it? This sucks because I spent $270 bucks for 15 gallons of the glycol and according to you it wasn't needed and may be hurting my systems performance. Thank you again for all your time!

You are most welcome! With regard to the propylene glycol, I would just run with what you have for the time being and not worry about it. Yes, propylene glycol does reduce the heat transfer efficiency and also increases the pump load…but not to the extent that it is worth the effort/risk of dealing with this in the middle of the heating season. If you want to better understand or see how much higher the supply temp has to be with propylene glycol in order to provide the same heat transfer as without it, you can vary the % Propylene Glycol/Water input parameter in our HR floor heating design software. I would suggest that you just keep this info in the back of your mind for the time being and then consider if you want to do something different during your non-heating season.

Yes, there is some pretty expensive propylene glycol on the market. There are people who will tell you that you absolutely must use this expensive and magical propylene glycol for HR systems and for solar water heaters, etc. I have a friend who is a PhD chemical engineer who works for Dow Chemical who convinced us several years ago to just use the cheaper propylene glycol that you can purchase at NAPA in mass quantity. He basically said that the only difference between the cheap stuff and the expensive stuff is that the expensive stuff has undergone extensive certification testing to allow it to be used in the food processing industry. Chemically, it is identical and our experience has been that it performs well in HR systems. Propylene glycol is NOT toxic like ethanol glycol. So if your dog drinks it because your absent-minded son left a can open on accident, your dog will just end up having bad gas for a day or two. Don’t ask me how I know this…

Unless your PEX is pretty much in the middle of the slab, I don’t know if I would consider it to be well-encased in concrete... PEX is pretty tough stuff. If you fill a piece of PEX up with water, seal the ends closed, and then freeze it, it will expand without failing. If the same sample is well-encased in good concrete and you then freeze it, it won’t expand because the volume can’t change (a different type of ice is formed than what we are normally accustomed to seeing), and the PEX won’t fail. If the same sample is at the bottom of the slab (or there are air voids around the PEX) and you then freeze it, the lower section of the PEX will expand toward the insulation (or the PEX will expand into the concrete voids), and the PEX will likely fail at these locations because more expansion occurs over a smaller area than the PEX can handle. Anyhow, that’s what our testing has shown and take this for what it is worth... We still recommend using propylene glycol, but only if there is a significant freezing risk. Freezing risk can also be mitigated by having a temp alarm and redundant circulation/heating capability.

Posted By Limited87 on 22 Dec 2014 03:40 PM
Hello everybody, I am new to radiant floor heat systems. I am in NE Ohio where winters can get pretty brutal. Sometimes many days on end of below zero temps. I spent months researching and talking to different companies about what I wanted to build. Seems like everyone had a different idea of what I should do. Here is a full break down of what I ending up going with and I have some questions about my system. I had a 64x48 shop built with 14ft ceilings. 2 12x12 2 inch thick insulated doors. I installed 3 inch foam board (R22 R value) under the slab. I installed 3000 ft of 1/2 oxygenated pex tubing (300 ft runs) with 10 circuit feed/return manifolds. The concrete is 6 inches thick. The shop has Owens R30 insulation in the walls and R50 blown in the ceiling. Building is full metal finished inside. I basically could not have went with any better or higher insulation in this shop to make it as energy efficient as possible. I didn't have an option of natural gas and I did not want a huge propane tank outside my shop so I went with electric. I went with hydro-smart out of Minnesota for my system. They built a panel with 2 pumps with all the required valves, etc. I went with an 81k btu hydro-shark instant boiler. The building seems to heat fine but it hasn't been real cold yet. (high teens) I have the temp set at 60 and the other night it just ran and ran at 58 degrees and didn't want to hold 60 like it has been on the warmer nights. There are several questions I wanted real world installers, users like yourselves opinions on. There seems to be a HUGE difference of opinion on glycol and what % should be used. I was told I needed 50% ratio.. some say that's way to much and will kill the heat transfer and make the system pumps work much harder, etc. I wound up putting in 15 gallons of cryo-tek glycol and 23 gallons of distilled water in and pumping this into my system. Which left me about 4 gallons left in my fill tub when finished. That ration puts me in at just over 39% of glycol in the system. Is this to much, not enough? I cant seem to get more than 70-72 degrees of water return temp no matter what temp I heat the water to or how long it runs. I currently am sending it at 125 degrees to the system. Does this sound normal to you? Also the pumps have 3 speed settings.. should the water/gylcol be sent through at max speed? Medium? Some say slower some say faster.. Any insight is appreciated. I'm sure Ill have some more questions. Thanks in advance guys!

Problem 1 = LACK OF tubing 64x48 = 3072 Sqf

that means you need to take 3072 * 2 to get the amount of PEX for 1/2 and that would be 6144 FEET of tubing needed to cover the area. Or go with 3/4 pex and than 3072 * .75 = 2304 FEET

Problem 2 = NOT ENOUGH BTU take 3072 * 50btu/h = 153600 divide by .95 boiler efficiency and your looking at a 161k BTU BOILER need for that area

Problem 3 = poor calculation take 161000 divide by 10k for a 20degree drop or 7500 for 15deg drop, lets do 20 - 161000 / 10000 = 16.1 GPM needed

now u ran 300ft loops 1/2 times 10 16.1 GPM divide by 10 loops = 1.61GPM each loop needed

u need 24.6 HEAD LOSS PUMP give or take (24-26)

U need a pump to push 16GPM with 24-26 HEAD

and thats if you place the pump at the manifold from and too 300 ft this is NOT including the piping for your system which will add MORE HEAD LOSS

lets do 25btu/h than u need 81k boiler with 8GPM and 9-10 head loss but you have LACK/NOT enough tubing so you need to go with higher btu output

Problem 1 = LACK OF tubing 64x48 = 3072 Sqf

that means you need to take 3072 * 2 to get the amount of PEX for 1/2 and that would be 6144 FEET of tubing needed to cover the area. Or go with 3/4 pex and than 3072 * .75 = 2304 FEET

Problem 2 = NOT ENOUGH BTU take 3072 * 50btu/h = 153600 divide by .95 boiler efficiency and your looking at a 161k BTU BOILER need for that area

Problem 3 = poor calculation take 161000 divide by 10k for a 20degree drop or 7500 for 15deg drop, lets do 20 - 161000 / 10000 = 16.1 GPM needed

now u ran 300ft loops 1/2 times 10 16.1 GPM divide by 10 loops = 1.61GPM each loop needed

u need 24.6 HEAD LOSS PUMP give or take (24-26)

U need a pump to push 16GPM with 24-26 HEAD

and thats if you place the pump at the manifold from and too 300 ft this is NOT including the piping for your system which will add MORE HEAD LOSS

lets do 25btu/h than u need 81k boiler with 8GPM and 9-10 head loss but you have LACK/NOT enough tubing so you need to go with higher btu output

Maybe you guys understand what I mean when I said 10 different people have 10 different answers..lol

I hope you are joking my friend.. I had a hard enough time installing 3000 feet of pex in this shop 1 foot apart everywhere and making the bends and runs.. I cant even imagine tryng to install 6k..lol That would be absolutely impossible. It would kink trying to make a 6 inch end to end bend. Not 1 person ever recommended 3/4 inch pex for this job.

Now the boiler.. 161k btu??! I would need a jet engine to produce that kind of BTU practically and a small block Chevy to pump it..lol I appreciate the feedback but these numbers and recommendations are out of this world. If this was the case I would have walked away from radiant floor heat 5 minutes after thinking about it because the cost would have been off the reservation.

That being said the temp was in the mid 30's here all day long and very windy.. the system sat idle/off all day long with the t stat at 60 and the shop temp sat at 61.4 I cut back that mixing valve about half way and it has really increased the return line temp. Almost 10 degrees and it hasn't even ran that much since doing it. I will repost when the temps get real cold and let you know how the performance of it is acting.

Believe me, I know exactly what you mean...LOL.

You will need to experiment with your manual bypass valve to find the "sweet spot" which is typically between half closed and almost fully closed. A 12" spacing for 3072 SF area is 3072' of PEX. You probably could have done 18" spacing because I don't think you will spend much time walking around in bare feet in this building.

Merry Christmas!

Posted By sailawayrb on 25 Dec 2014 09:57 PM
Believe me, I know exactly what you mean...LOL.

You will need to experiment with your manual bypass valve to find the "sweet spot" which is typically between half closed and almost fully closed. A 12" spacing for 3072 SF area is 3071' of PEX. You probably could have done 18" spacing because I don't think you will spend much time walking around in bare feet in this building.

Merry Christmas!

I will do some experimenting over the next several weeks and report back with my results with the valve. That's what several said I could probably get away with but I went ahead and did the 1 foot and made up the 70 feet towards the very center of the floor as several advised to do. Wasn't hardly even noticeable in the lay out honestly. Yeah this isn't a bare foot type of shop..lol Thanks again Sail! Merry Christmas!