backcalculating system psi/head using an air proxy

CC.Rob

Just wanted to thank you gentlemen again for your time and insight. I've learned a lot that helps me understand the problem and the solution.

CC.Rob

backcalculating system psi/head using an air proxy

Recent conversion from psi->head(ft) in another thread has me wondering. Backstory: My two zone hotwater system has significant air problems at "normal" pressures say 15-19 psig. A longish loop with lots of ups/downs and elbows on the second floor seems to be the culprit (TEL = ~325 ft). Air in this loop is generally always trapped in the top of the system, but some apparently makes it down to the vertical return manifold and gets trapped, causes cavitation in the circulator, degrades system performance, etc. I presume flow velocity is insufficient to get into/through the boiler and to the spirovent. (Yes yes I know we're not pumping away but way back when I didn't know that, and the system as piped presently is very efficiently done and repiping would amount to major surgery and if possible I don't want to go there, yet.)

However, bumping up the psig to around 24-26 appears to really quiet things down. Significantly reduced air noise upstairs, apparently no air accumulating in the vertical return manifold, etc.

Why is that? Possible reasons I'm thinking of include: 1) at higher psig, air is not allowed to come out of solution, 2) at higher psig, although system resistance (head)increases, the air is better entrained in what I presume to be sluggish flow on the long loop and makes it through the boiler to the spirovent.

Using these data and observations, is it possible to back-calculate the head of the long loop at "air free" psig? (e.g., psig(noair) - psig(air) = head but I'm pretty sure that's not the right way to do it). Standard calculations for this loop suggest head = 10.1 ft @ 4gpm, but I wonder if it's higher.

Thanks for any insight.

Brad White_9

Analysis Paralysis....

Just having fun here...

Part of what I see is that head pressure calculations and static pressure have little to do with one another in a closed system. Short of low pressure (cavitation) and high pressure (bursting of gaskets) I see no applicable effect of static pressure on pump head. Open systems? Sure, you may have a lift factor between when the pump is running and when it is not.

At higher pressures the free air (bubbles) will not readily re-dissolve especially if the water is heated. But under pressure the bubbles will be smaller, proportional in volume to the applied pressure. For the first 14.7 psig (OK 14.696 from and at sea level) added to a given atmospheric system, the volume of the air bubble will be reduced by half.

In other words say you have a system which is initially open at the top then closed with an air bubble of five cubic inches volume at 14.696 PSIA Absolute. Then the system is pressurized to 14.7 PSIG (29.4 PSIA). The bubble is then 2.5 inches in volume. Still a bubble. But it's ability to compress is now equal to it's ability to push back. It becomes less of an absorber of pressure because it is in balance. It has less place to go when more pressure is applied. For this reason you would have less bouncing.

Perhaps I over-simplified things, but that is the dynamic I see happening.

Now, reduce the pressure and the bubble returns to it's orginial volume at the original pressure. This is why, as a scuba diver, I am wary of the bends... healthy respect for tiny bubbles in odd places.

CC.Rob

excellent

Thanks for the explanation and examples. As a now-former technical diver, the last was very helpful.

If I may redirect the question somewhat: if the system has a lot of "sensible" air (i.e., makes a racket, degrades system) at ~16 psig, and we compress that air to 24 psig to make it less noisy (i.e. better entrained), will the air ultimately be eliminated or if not, does it pose a threat to the cast iron boiler?

And in a related vein (bends pun intended...), if calculations show 10.1 ft head @ 4 gpm, why is that not sufficient to entrain the air and push it past the elimination system? That's pretty close to what the circulator (Taco 007) should be doing. (i.e., resistance is quite close to the pump curve -- the crossing point is right around 3.95 gpm for this pump). Pump curves should be somewhat conservative, right?

Brad White_9

Glad the SCUBA reference was not too obscure!

Hi, Rob-
The feet of head atributable to a pump selection is merely the differential across the pump when referring to a closed system. Two identical systems, one at 12 psig and the other at 25 psig will need pumps of the same head pressure. One side cancels out the other, what remains is frictional and dynamic losses to be overcome. Your 10.1 feet and 4 GPM has nothing to do with velocity nor the static pressure in your system, simply. (See below on velocity.)

What is key to removing the air (what my too lengthy answer somehow completely avoided) is twofold (one, the other or both to an extent): Venting the high points under a pressure to allow about 4 psig at the top of the system. That is one. The other is to have sufficient velocity if not under pump pressure then under hose pressure as in a purge. Flow at sufficient velocity to force air with it. Generally this is about 2 FPS.

You want to have this velocity coupled with an air separator such as a Sprirovent. Quite the bubble scrubber. Velocity has to be sufficient to grab the air down to the scrubber (or whereever the scrubber is) where it is then released.

This is why pipe sizes are generally designed for flow rates between 2 and 4 fps. The slower is a minimum to entrain air, the upper is to keep noise to a minimum. Water being incompressible tends to transmit noise over some distance.

Ask any whale you walk by on the beach!
Uh, pardon me... is a good way to begin that conversation

CC.Rob

Becoming clearer still! I was using 4 gpm to represent a flow rate in 3/4" copper, which from my reading says 2 ft/s = 3.2 gpm; 4 ft/s = 6.5 gpm (but obviously I'm not telling you anything new). Since 4 gpm equates to >2 ft/s in 3/4" pipe, why isn't the air being moved along and past the spirovent?

My observations suggest that a) there is air "stuck" on the 2nd floor, and b) air is also trapped in the vertical return manifold (I think this is 1" or maybe 1.25" going into the circulator, about a 5 ft run with all 3 zone returns coming into it at different heights [3rd zone is an indirect]). The system has been purged several times by different people using different methods, yet all seem to end up over the course of ~3-7 days with the air in the same places. And none of it seems to leave. Puzzled.

Brad White_9

Clearer still? Cool!

I love teaching and learning. Hand in hand.

You do have a Spirovent I just read and I missed that in reading your earlier post until just now. Anyway, the reason the air may not be drawn down is that the velocity is not constant.

When it hits the larger pipe at the same flow it "loses it". (Kind of like thinking about Janet Reno during a, ah, well, passionate moment.

Speaking of pumping away, Back to Basics: I see that you are not pumping away from the expansion tank connection, and that you know this already. Nothing I can do about it. Big one, part of Hydronics 101 but you have atoned.

By pumping toward the expansion tank you could be inducing air into your system (less than atmospheric pressure indeed) on the suction side of your circulator. Not likely as an absolute, but possible. Make sure you are pumping away from the ET. Also make sure you have a diaphragm tank as opposed to the conventional air cushion tank.

The other part of this is, even a well-purged system, with all obvious air removed, will outgas air when heated. Once a system is heated past oh, 180 or to limit a few times, sufficient air is released. It will only go back into solution when the water cools. (Henry's Law; Henry was Boyle's more pedantic cousin, apparently.) By and large most of the air will remain free though. It just has to be vented or coaxed down to the Spirovent.

Anyway, the fact that the air re-appears after a few days leads me to that.

Another thought posted by the Amazing ME (Mark Eatherton) was to put in an ounce or so of dishwashing detergent to break the big bubbles into smaller bubbles and carry them better in solution to a release point. Just remembered that.

But remember too, because you are not pumping away from the ET you could be drawing in air into fittings, vents, gaskets and from places unknown. Less likely given the higher static pressure but still possible.

You know what, Rob? Come to think of it, next system I design, I will specify a top-level Spirovent or their nice air vent device as a second choice. Why not but for short cost?

Random thoughts-

Brad

CC.Rob

Wonderful conversation, this. Thanks for taking the time. I am indeed learning. Outgassing of freshly introduced water is exactly what I've been thinking is the air source. How about adding a wrinkle? System worked great for two years, then had between ~20-35 ft TEL added to 2nd floor zone. That is when the problems started.

Could that have sufficiently changed the hydraulics of the system such that we have the present problem? For example, if we weren't drawing in air previously due to not pumping away, why would we now?

Would have thought adding <~10% to overall loop would be within the margin of error for head, flow, etc. Unless of course it was already near a practical limit (but calcs say it should be fine).

Related observation is that the problem doesn't appear to get terribly (maybe even not measurably) worse over time -- air seems to behave the same way, pressure holds constant. Although the system really only gets about a week to run between being tweaked in some way in an effort to fix it.

To address the comments: ET is a State Thermoflex. All appears well with it. Followed the thread on the dish detergent. Sounds like a neat idea. However, I can't seem to find the "squirt dish detergent in here" fitting. Maybe the label fell off?

Thanks again.

Brad White_44

New Wrinkles....

Rob, Hi-

Back again...

It is interesting that a modest addition of piping (and do I gather more height hence more static pressure?) precipitated the problem. I cannot gather why that might be, frankly. The only inkling is that when you are pumping TO the point of no pressure change (the expansion tank connection point) you by default HAVE to impose all of that head as a negative head on the suction side. Perhaps the added resistance gave "rise" to greater negative pressure on the suction side. Just a guess. I agree it is not much and on paper it works... Conversely, bumblebees cannot fly which can be amply demonstrated on paper. Yet they buzz along in blissful ignorance...

Point being, and I never mind saying this: I do not know.

Dish detergent entry: I am fortunate to have a Neptune filter type shot feeder in my system. It filters the water with each pass, I inserted some rare-earth magnets to harvest any loose iron and it gives me a portal to add chemicals. In your case you can do a couple of things, both of which I have done before my beloved Neptune FF:

I had a port consisting of two check valves in series with a ball valve at the open end and hose connection. The two check valves were cheap insurance. I would hook up a hose with a garden type sprayer, (cheap plastic type used to spray Agent Orange, DDT, Phosgene and a host of nerve agents upon the unwitting pest population). I would place in said container what I wanted to inject into the system. Cleaner, inhibitor and yes, I admit once, leak stopper

Soap was not on my list, I just learned that.

The other method which is at once really tedious and stupid looking, is to drain down the system below the level of one of your highest floor radiators. Use the smallest funnel and plastic tube you can find to dribble the chemicals into the radiator hence into the system. I used this method to inject a dye to detect leaks. It showed up on ball valve stems by the way. FYI.

So try the soap thing- I would be curious how it works in your case. Leaves piping sparkling clean!

Brad

Rob_32

No height gain. Addition was done laterally, not up. Everything on 2nd floor at same level before and after, just more pipe. Increasing the system pressure is to help suppress the loudly gurgling baseboard and get some sleep.

If I understand correctly, what you're saying is that in effect we are now sucking through a longer straw, inducing greater negative pressure than there used to be. In which case the valves on the returns could now be admitting air when they were not previously.

If that's correct, then it would appear there are two potential courses of action: 1) pump away from the point of no pressure change (the expansion tank), and/or 2) make the straw shorter. The former would require some creative piping, as the existing system is really quite compact and efficiently laid out (albeit with circ on return). But certainly doable given what I've seen in pictures posted on this forum. The latter (splitting this zone into two smaller zones) would be fairly straightforward, although quite invasive as it would require opening a stud bay on the 1st floor as well as a piece of ceiling. The former seems like it would be the logical thing to do.

Am I correct about the straw length? Thanks once again.

p.s. the mention of DDT always reminds me of warm summer evenings long ago, when all us neighborhood kids would ride our bikes through the mist of the mosquito control fogging truck, sometimes for miles....

p.p.s. a syringe loaded with an appropriate quantity of Palmolive could probably be injected through a manual bleeder on the 2nd floor with minimal need to drain down. I'll think about that.

Brad White_44

Yeah, Rob , you have to Pump Away

just about as sure as I can be. You did not add much of a straw it would seem, no.

Syringe- excellent idea! Wish I had thought of that.

Mosquito spray- smelled like kerosene, but we never did get Dengue, Yellow Fever or Malaria come to think of it...

Good Luck-

Brad

Christian Egli_2

Glub glub glub

OK Brad, so, I'm wearing my tight fitting scuba diving outfit, I'm making bubbles along the sea shore, and while I'm looking for a whopper of a whale I keep thinking about Janet Reno. Oh, and should I see one, armed with my garden sprayer filled with phosgene (no less, wow), a little skirt will get me out of harm. Cough.

Well you know what? playing with pipes seems like more fun than going through all that trouble.

Rob, why don't you just run the system at high pressure since it works that way.

For weeks once, I was hunting for a problem on a water system. The air leak into the system was at the pump only during pumping. It was the paper type gasket at the pump that was porous enough for air to scuttle through. Everything was tight though. The ugly quick fix: a little tar painted all around the seal got my leak asphyxiated.

DDT might have been stronger of course, but the ultimate is the sterilization with... steam, don't you think?

Neat thread, fun numbers.
Thanks.

[Deleted User]

What he said...

Brad hit the home run on the head out of the park...

Either move the circulator so that it is pumping away from the PONPC, or alternatively, move the PONPC so the pump is pumping away from the PONPC, but towards the heat source and your problems will go away. Guaranteed.

It's hard to visualize in your minds eye, but even with the low head 007 circulator, under your circumstances it produces around 10 feet of pressure differential, which after elevation changes etc leaves you with very little pressure in the top of the system. Air LOVES to come out and play when the pressure is low. If you have ANY type of air vent on the top of the system, it WILL act as a vacuum breaker and ALLOW air to come into the system on a regular basis.

Can't tell you how many 3 story apartment complexes I see where a first year hydronics technician went through and installed a box full of air vents on the upper floor trying to address the air in the upper floor issue, only to get called back because the pressure relief valve is dripping. The system had the old non captive expansion tank installed, and the air vents lets the cushion go away, then pressure rises, and relief valve does its thang. There is usualy a pile of relief valves in one corner of the mechancial room, and a pump pumping towards the PONPC... Guess they don't make'em like they used to eh...

BTW, FWIW, this has been a VERY interesting conversation.

Thanks for the entertainment:=)

Now, quit hanging out on the internet and go fix your system so your wife can sleep at night:-)

ME

CC.Rob

figure interpretation

Glad to know this has been entertaining and/or interesting. When I can detach myself from the aggravation this problem has caused, I find it interesting, too.

BUT, Mark opened a door with that figure. This has been a heck of an education; I've got a bit more to go. Would you mind adding some interpretation? Just a couple numbers I don't understand (one relevant, one not but interesting). Let's see if I can interpret this, and if you would, correct me where I'm wrong. It looks like you've diagrammed my system and plugged in my calculated gpm (3.95). (Or did you use my actual TEL?) From that, the head added (10.05 ft) is derived from the selected circulator (007), and the differential pressure across the pump (A->B) is calculated (4.24 psi). No balancing valves (DPBV off). D is the heat source and begins the supply part of the circuit at 180F. No heat emitter is specified, so the 2250 BTU/hr is the heat loss in bare 3/4" copper at 3.95 gpm. I assume that you did specify a TEL somewhere to get that loss value. The heat loss is obviously the irrelevant number, but it is interesting to know.

The number I don't get: E appears to be about halfway through the loop, i.e., at the top of the system. The EHR at the ~midpoint of the circuit is .90874. Questions: 1) what are the units of EHR, 2) what does that value tell you, 3) what should that number be for the circuit to function properly, and 3) how simple is it to model what happens when the circulator is moved to pump away from the expansion tank? (What I'm asking is Mark, if you've got a spare minute, can you plug this in and show me?)

Christian, the air problem is suppressed at higher pressure, but not eliminated (e.g., air in return manifold, circ still cavitates).

Thanks again for all your insight. Consider me appropriately educated. Will be taking the necessary steps to fix the problem and report back.

Jerry Boulanger_2

When branches that

are upfed get 'airbound' they may remain airbound when the pump comes on if there is not enough differential across the branch to 'lift' the water over the top and get it moving at a sufficient velocity to carry the air out of the branch and to the air separator. See the attached schematic.

Increasing the static pressure can help by pushing the water up further in both the supply and the return, thereby reducing the differential necessary to get the water over the top. It also helps to flow balance the system and create the required differential that way.

[Deleted User]

What you're looking at....

Actually, I plugged in your TEL number, the fluid, fluid temp and the pump. Siegenthalers software did the rest. I did only use the TEL, and used 3/4" finned tube pipe for that value.

If you can tell me how many linear feet of baseboard you have on that circuit, and its related output per foot, I can fine tune that number and the software will calculate the actual delivered btuH.

The feet of head added and pressure differential are one in the same expressed as different values.

The EHR value is explained more in depth at

http://www.pmengineer.com/CDA/Archives/b8975647d7298010VgnVCM100000f932a8c0____

Moving the PONPC will not have any affect on these values, only on air being and staying in suspension and not causing issues with circulation.

I've attached an Power Point demonstration that I put together to help my students understand the importance of pumping away from the PONPC.

Let me know if it helps your minds eye see it better.

EDIT: The wall won't let me attach the PPP file. Email me your address and I will send it to your directly. markeatherton@hotmail.com


ME