Hydronic Radiant In-floor Heating: Pump & air valve replacement

[Greenleaf]

In our recently purchased approx 20 yr old house, we have an in-floor radiant heating system, that seems to work quite poorly, as in very slow to heat (like, half a day to get rooms up to temp). I suspect the previous owners either did not use it, or only used it supplementally. (house also has gas fireplace & wood burning appliance). For some areas, at least, I think they simply ran the loops spaced too far apart to be very effective (between joist installation). In any case, on with the story...

After two months of this, the air bleed valve started to leak and drip water. I put a basin under it to catch drips and continued using the system. A couple of weeks later, I noticed a nasty sound coming from the circulation pump. I shut off the system for the day. The next day it seemed fine again, in fact, for a few days. But then the pump noise returned, and I noticed the temperature of the pump housing was too hot to hold my hand on it, but not sure if that is normal. Also during the period, the system seemed to be more noisy (air in the loops?). System has been shut-off since then.

I'm wondering if the bleeder valve was possibly bad enough to actually be causing air to be introduced into the system, eventually starving the pump itself of water, possibly now causing it to seize-up. Even if the pump was starved of water, perhaps a full flush after replacing the bleeder valve only, might get it back in order without needing to replace the pump cartridge.

Anyway, as of now I have ordered a replacement air bleed valve and pump cartridge, but would like some guidance on replacement sequence, as well as post-replacement air purge.

Here is a photo of the system, which I have tried to annotate as best I can (possibly incorrectly!). See notes below:

Here are my draft procedures to remediate these issues.

Circulation Pump cartridge replacement Sequence & Questions:
This one seems to be more straightforward since there are shut-offs immediate on either side of the pump:

1. shut off power to the system
2. close shut-off valves 2 & 3 (at least) above & below the pump (if you can make out the labels)
3. Replace cartridge
4. Reopen valves
5. Flush (see questions below on that)

Air bleed Valve replacement Questions:
This one I'm not so clear on which valves to shut-off as there is no shut-off between the boiler and the bleeder, but there is one on the other side of the bleeder. Perhaps shutting-off valve #1 above the boiler will be enough? Do I need to depressurize the system somehow? Perhaps I additionally need to shutoff the water from the main feed (far left blue valve below and to the left of the boiler (uncircled)).

Flush Procedure Sequence & Questions:
So as for flushing, here is my best guess as to procedure:

1. Close valves 3 & 4 on either side of the pressure tank to stop water returning back to the boiler
2. Attach a hose to the flush station on the bottom right (green circle) and open the valve there
3. Manually override the flow valves into each loop (one at a time) using the switches on each loop valve (upper-left in small blue circles)
4. If necessary, optionally open the pressure valve before the boiler (blue circle to the lower-left of the boiler) to increase the pressure as each loop is flushed (Is this always safe? Or should it only be special cases like 3rd floor loops, etc?)
5. Return each loop valve back to auto mode
6. Once all flushed, close the flush station and re-open the valves, etc.

So, as you can see, most of my questions are around the air bleed valve replacement sequence, basically because I don't have a nice pair of shut-offs on either side of the valve.

If anyone is willing to chime-in with some guidance, tips, or warnings, it would be appreciated.

Cheers!

 

[Dana]

It's sometimes cheaper to replace the pump than just the cartridge. The high temperature of pump motor could be an indication that the motor, not the cartridge is bad. There are drop-in ECM drive replacments for the Taco- 007(?) that should work just fine, and it would pay for itself over time in reduced power use (8-15 watts instead of 75-100 watts.) Shop around- they can be had for ~USD$125, sometimes less on the web.

Air vents mounted on ells like that aren't nearly as effective as air purging devices mounted in the middle of a straight section of pipe, and are more failure prone. This is due to the turbulence created by the ell, which tends to re-mix the air rather than letting the bubbles float up to the vent. The same turbulence also wears out the seals of the vent valving.

As long at the system pressure is at least 12-15psi the fact that the vent is spitting water doesn't much affect the operation of the system. If the pressure is allowed to go too low there will be cavitation at the pump, which will wear it out prematurely.

Almost all radiant heating systems have slow response times, making them less suitable for a night-time setback approach. "Set and forget" can work pretty well even with dumb thermostats, but some amount of SMALL setback can work with those using more sophisticated algorithms (but it's not usally "worth it".)

You definitely want to shut off the potable feed prior to replacing the vent. If you're really quick you can probably swap it out without adding too much air to the system.

When replacing the pump isolate it with the pair of ball valves, but don't fully tighten the top flange. Open the bottom ball valve first and let the pressure push most of the air out of the looser top flange, tightening it up slowly until the water stops dribbling, which will get rid of most of the air in that section.

 

[Sponsor]

 

[Jadnashua]

Depending on how the radiant piping is configured in the floor, it could be very normal to take as much as 2-3 DAYS or maybe even more (say in a slab, for example) to reach equilibrium. Then, it depends on the supply temperature, the makeup of the floor (under floor, in floor, heat diffusers or not, wood, carpeting, tile on the surface, and the list goes on.

As Dana said...these systems tend to work best at set and forget...there usually isn't enough 'extra' radiation to quickly heat the floor. If there were, you'd put a lot of stress on things, as right at the tubing, it would be really hot, while still cold in between. You need to slowly bring the floor up to temp, and leave it there for best operation.

 

[Greenleaf]

Thank-you both for your replies. True I think we are putting the system through some unusual stress by being out-of-town 2 or 3 days each week and I would just shut the heat completely off. Upon return it would take basically a full day to bring things back up to temp, as it has been as low as 10C/50F after a few days away. But even overnights the temperature can go from our normal 21C/70F down to 14C/57F by morning, if we turn the heat right off. Perhaps we just can't do that anymore.

We have yet to do a full review of the pipe layout but will pull some insulation in the basement ceiling soon to have a closer look. I spot checked one area and in an approx 3ft section of joist, I only saw one run of pipe, so the density seems a little sparse. There are heat spreaders, but they are just hanging in the air -- not in contact with the flooring above. But there is insulation beneath, at least.

It is indeed the Taco 007-F5 pump. The replacement cartridge (007-042RP) seemed to be a good deal at $87CAD/$69USD. The best price I could find on the entire pump was nearly double that. So far as I am aware all the moving parts are in the cartridge, so crossing my fingers!

 

[Dana]

Unless you KNOW it's the cartridge and not the motor, the replacement cartridge is a lousy deal at any price. I've seen both fail. The blistering hot motor isn't the usual symptom of a failed cartridge. Bad bearings on the motor and/or shorted windings would be my first guesses.

Replacing it with an ECM drive pump may be 2x the price of a cartridge, but it pays for itself.

 

[Greenleaf]

There are drop-in ECM drive replacments for the Taco- 007(?) that should work just fine, and it would pay for itself over time in reduced power use (8-15 watts instead of 75-100 watts.)

The 007e does look like a nice option... especially if I can just drop it into the existing flanges, and wire it up.
https://player.vimeo.com/video/170631626?autoplay=1
http://www.tacocomfort.com/products/variable_speed_products/007e/index.html

 

[Greenleaf]

Just thought I'd post a quick update:

I replaced the air bleed valve the other day. Went pretty smoothly after I realized I need to bleed off the excess pressure after shutting-off the feed.

After that, did a full system flush one loop at a time. Lots of brown water poured from each, and some appeared to contain lots of air, others seemed pretty air-free, but still brown. I found that there was very little pressure to flush without lifting the pressure regulator bypass lever to boost it.

Once flush complete, powered the system back up to see how the pump sounded now. Seemed ok.

Once the system had been on for a while, went back to check on it -- pressure way too high at approx 30 PSI and water was dripping out the expansion tank seam.

Bled some pressure off at one of the purge stations, and now pressure varies between 15 or so when hot and 0 when cold.

Speaking of pressure variation, I now suspect the expansion tank is dead. It's a Watts ET-15. The bottom makes a dull thud sound when tapped and the top makes a bright "ping" -- shouldn't it be the other way around? Sounds to me like there's water in the bottom of the tank where the air should be.

Anyway, in general the system seems much quieter overall now, with less gurgling under the floorboards etc. I think there is some knocking/hammering when the zone valves open and close. Up until recently we were sure it was animals in the attic, but now realize the sound always corresponds to a thermostat clicking on or off.

In other news, I did go ahead and order the 007e-F4 pump now (with the universal flanges) so hopefully that will improve our electricity consumption, and perhaps run a little more quietly. Should arrive in the coming days.

p.s. I have reduced the temperature drop overnight on our thermostats so the climb back in the morning is lessened. Seems to be fairly happy with that setting in most of the house. One room, however, is still struggling, and in fact, I am noticing now it can't even maintain the lower overnight temperature, which I think means the system is basically on all night trying to heat up this one room! I have placed a space heater there to try to help and give it a morning boost. Once up to temp it seems to be able to maintain it, which is somewhat surprising since overnight it loses the heat it had and more. Maybe just heat from the sun on the roof during the day, radiating down in the afternoon.

 

[Jadnashua]

The end of the ET with the air valve should be without any water in it. Depending on the system pressure, and whether you've precharged the ET properly, when cold, the ET should be mostly empty. You cannot test the precharge while there is pressure on the other end. YOu have to relieve the pressure and leave a valve open, then you can use an air pressure (tire) gauge to check the precharge. It should be set to about the normal system pressure (typically 15psi or so).

IN a typical hydronic heating system, it will have some iron components. When you add fresh water, it will have some dissolved gasses in it (including oxygen). That oxygen will cause some rust to occur, but once all of the oxygen is used up, the water becomes fairly inert and it stops. The more fresh water you add, the more things rust. Ideally, it never leaks, and never needs to be topped off, and then, things can last a long time. Some rust is normal when draining the system unless it is unusual and was done in all copper/plastic/brass/bronze/SS. Most circulators and autofill valves are iron, and your boiler may be as well. Some use a lot of iron pipe.

 

[Dana]

Since it's an electric boiler it might be hard to "see" the reduced power consumption of the ECM drive pump, but the pump won't ever be as hot as the gunked-up 007 was, and the boiler room won't be as hot, lowering the parasitic heat load of that room.

With the orientation of your expansion tank the bottom half should be the one with air in it. If you press the Schrader air valve on that end and nothing but water comes out, it's toast and needs to be replaced. If air comes out it needs to be pre-charged with air to the intended system pressure or a little bit higher to have sufficient expansion room. This has to be done with the system at zero pressure, boiler & circulation pump off, ideally with a flush valve cracked open a hair to allow water to escape as the tank gets pumped up.

 

[Greenleaf]

UPDATE:
Ok, have now completed the system maintenance I've been working on for awhile. Here's some updates and a few remaining questions at the bottom:

1. air bleed valve replaced (as above)
2. expansion tank replaced
3. circ pump replaced
4. full system flush (multiple)

Out with the old expansion tank. Perhaps original to the house, and therefore 20 yrs old. In any case, once I removed it it was clear it was fully and completely shot. 100% full of water!!

Out with the old Taco 007-F5 pump. I think it was still working "ok" actually. I think the noise previously noticed may have been due to huge air pockets in the system (just a theory, but it has been running fairly quietly since the bleeder valve was replaced and earlier flush). Note that black sludge that was dripping out of it...

In with the new expansion tank, EX-15. Same capacity as the old. Went with an Amtrol this time. Leaving the precharge at 12psi.

Removing the old tank was a bit of a bear and I ended-up cracking the lower solder joint in the process. So off to the big box store I went to buy a torch and supplies to clean up the old lower joint and resolder it. The result could be prettier, but it is leak free. The new tank has been hand-tightened only -- lots of leverage gripping just the outside of the tank and tightening moderately.

In with the replacement circulator pump: Taco 007e-2F4. At first I thought I had received the wrong part since what I ordered was an F4, but apparently it is the same thing? Taco's website doesn't even mention a 2F4.

Anyway, this is the one with the universal flanges, which work fine. Pretty straightforward swap. It arrived with the housing rotated the wrong way for my installation, so had to turn it 180 by removing the 4 bolts and spinning it.

I did not use the optional IFC valve that ships with it, after reading various opinions that it is usually unnecessary for systems that use zones with zone valves (as opposed to zoning with circ pumps), and occasionally may cause hard to diagnose issues down the road.

So it's been a few days now, and I have some remaining questions about the system pressure:

• Should the boiler pressure remain at the same steady pressure both when on and when off? I have gently purged the hot system to a boiler pressure of approx 12psi. But when the system cools it drops to 0 at the boiler gauge. I might have thought that the expansion tank precharged to 12psi was supposed to maintain a 12psi system pressure, but maybe that's not how it works.
• If the pressure really drops to 0 when off, shouldn't the regulator automatically top it up again to the preset pressure?

Cheers.

 

[Jadnashua]

A properly sized ET that is properly pre-charged (you must set the pressure before installation or with all water pressure relieved from its inlet side), and the pressure should never get lower than the 12psi you pre-charged the system to. It may rise a couple of pounds during heating operations, but not anywhere near enough to cause problems. IF the pressure actually drops to zero, the boiler shouldn't actually turn on! If you have an autofill valve, and have its isolation valve open, it should never get lower than the set point on the autofill valve. If it drops, you have a leak somewhere or the autofill valve is defective.

How high is the highest point the water must be pushed from the boiler? You need pressure in the system that is high enough to do two things:
- close the low-pressure safety interlock (often in the 8-10# range - you'd have to read the boilers specs)
- keep the pressure at the highest point a positive value (the pressure will drop at 0.43#/foot elevation rise). This is to prevent a vacuum and to ensure the water doesn't boil.

 

[Greenleaf]

Thanks, Jim. I did check the pressure of the tank before installation. It was right on 12psi.

I don't think there is a leak simply because when the system heats up again it returns now to approx 12psi, rather than the 18-20psi it was reaching right after the initial fill and before I flushed the excess. Since flushing the excess it has remained under the red mark on the boiler gauge when hot, but still drops to zero when cold.

Likewise, I don't think the regulator is topping-up the system when cold, for the same reason. If it was, the pressure would be going too high again.

Perhaps it is just an oddity of this boiler and its gauge that it drops to zero when cooled? I will dig a little more on that question.

I guess I should perhaps just get a pressure gauge that I can use to test the pressure independent of the boiler, threaded on at one of the flush stations.

Boiler is in the basement of a two-storey, and this is all radiant flooring, so the max height above boiler is only about 4'.

 

[Jadnashua]

If there's radiant on the second floor, it's more likely around 12-14', which would require a minimum of 0.43*12 or 14, plus add a couple for safety. Most boilers have a low-pressure cutout switch, often set to around 10psi. The exact value is up to the manufacturer, but that interlock MUST be closed due to sufficient pressure for the boiler to turn on. So, if your gauge reads zero, and the boiler does turn on, one of two things is happening:
- the switch is defective and shorted closed
- the gauge is not reading properly and the switch is working (or the control board is defective)

I suppose that a UL approved boiler might not have a low-pressure cutout, but if so, they are not common.