Where is the best location of the circulating pump in the heated loop of an in-floor radiant system, that is, should it be placed between the heat source and the heat exchanger on the hot side or on the return side.  I have a tankless heater that will provide hot water to a heat exchanger, and both sides of the loop are equal.  Is it best to place the circulator near to or far from the heat source pumping toward the heat exchanger or pumping from the exchanger back to the heat source, or does it matter at all?  The total loop length is about 12 feet.

What temperature are you running in the loop? What is the temp limit of the pump? In very high temp systems, sometimes they are placed in the return side so the pump operates at a lower temp and lasts longer.

If you are not at an extreme temp, it probably dosn't really matter as long as the pump can remain free of air...

OK, thanks.  Running 125 deg. F; pump rated at 220 deg F.

More important is to have the expansion tank just upstream of the pump so it always has a reservoir to pull from. And as ronmar mentions, making sure air gets out rather than into the pump.

No expansion tank as it is on a well system with a large expansion tank that manages all the water loops in two dwellings.  A small expansion tank on the floor side of the Hx, however, along with a pressure regulator, second pump, etc.  Just wondering if the high pressure/low pressure gradient was important to consider on the short loop, heated side of the system.

So at some point this loop has a T where the supply from the expansion tank connects. I'd put that just before the pump. It probably doesn't really matter, but that seems like good practice.

Circulators can be located on either the return or supply side of the boiler. They were historically placed on the return side because it was believed the lower temp would prolong their life. However, most circulators these days are rated for temps well above the boiler supply temp.

It is more important to locate the circulator so the inlet is very close to the connection point of the expansion tank. Doing so results in the pressure increasing in nearly all sections of the circuit when the circulator is operating. This is desirable for air separator operation and to avoid risk of circulator cavitation. Every hydronic radiant system should be closed loop which requires having an expansion tank and a pressure relief valve.

OK, thanks.  As a point of clarification, are you suggesting that radiant systems need their own heat source?  I have a system that uses a Bock water heater as the heat source, and this new one will use a tankless water heater as the heat source, but in both cases the heat source also provides DHW to their respective dwellings.  The floor systems are in separate closed loops isolated from the primary heat sources by heat exchangers.  The heat exchangers have a floor side closed loop, but they also have a very short open loop on the heater side of the heat exchanger.  For the most part, water that is in the heated side of each heat exchanger, through the use of check valves, doesn't mix with DHW, but the water is to some small degree co-mingled.  There are only about 2 gallons of water in the heated loop that takes in the heat exchanger and the piping from the heat source to it, but still, every time hot water is demanded in the house, some part of the water in the heat exchanger loop gets taken with it.  In both cases, that heated loop has no expansion tank because the entire domestic water system has its source in a well, pressure pump and a large pressure tank.  The heated side of the exchanger works under the same pressure gradient (30-50 psi) as the entire domestic water system.  My original question was to clarify whether or not there was a preferred place for the circulator in the short loop that goes from the heat source to the neat exchanger and back to the heat source.  I imagined there would be a pressure differential between the outlet and inlet of the circulating pump, and I wondered if that would create a preference for placement of the pump in the circuit.  Descriptions like these are full of flaws for which I apologize.

The heat exchanger makes the system closed. The issue is the components' exposure to potable water and everything in it, and the potential for increasing the amount of biofilm that can grow nasty things. In your case the component exposure is a pump, heat exchanger, and maybe a couple of valves, and the biofilm you've created is very small (presumably... hopefully the heat exchanger isn't 100' away from the water heater).

A short length of pipe to an air handler is another common and acceptable open system situation for similar reasons.

But radiant is not that, at all. You're good as long as you have the heat exchanger though.

As sailaway was saying the pump should be pulling from the expansion tank and pushing through everything else, if possible. The expansion tank is the "point of no pressure change" so you're negative from suction to the tank, and positive from discharge to the tank.

On the potable side of HE the question is unimportant in every regard unless you for some reason had a pump that was way too big for a given heat exchanger. That's pretty unlikely though.

"For the most part, water that is in the heated side of each heat exchanger, through the use of check valves, doesn't mix with DHW, but the water is to some small degree co-mingled. "

It's not clear to me where the check valves are. If you are not careful, there's the potential that when the water downstream of the check valve heats up, it builds up pressure with nowhere to go--no expansion tank and no way out the check valve.

Got it...I wrote a couple inaccurate things in the last post.  First, in the Bock system, the water in the heated loop of the heat exchanger is completely co-mingled with the DHW because the return line goes right into the water heater tank.  The check valves in both cases are (will be) located at the split between the DHW and the heat exchanger loop.  If water is drawn into DHW, water sitting in the heat exchanger loop won't be drawn into it and vice versa.  In the tankless system, a flow switch will turn off the circulators when there is a demand for DHW, but check valves at the wye will still prevent one from being drawn into the other.  At the return there is open flow back into the pressurized system.  The entire loop to and from the heat exchanger is about ten feet; the heat exchanger is about five feet from the heat source. 

Thank you, chrs, for your comments.  I reviewed my plans and had, in fact, planned a check valve in a location that would have created a problem.  Point well made. mm