Sizing System Side Header Pipes for Hydraulic Separation of Circulators

StewBrewer

I am sizing the headers for the system side of a 4-pipe buffer tank with 5 circulators drawing from the header. Sources say the velocity of the water in the header resulting from the combined gpm of the 5 circulators should be: “less than 2 fps”. At 2 fps, a 2 inch copper header carries a 20 gpm max and a 2.5 inch a 30 gpm max.
Is this “less than 2fps” basis for header pipe size selection for hydraulic separation overly conservative at pipe sizes of 2 inch and above?
If it is overly conservative, at what gpm level above 20 would it be necessary to switch from 2 inch to 2.5 inch to maintain hydraulic separation of the circulators?

Jamie Hall

The reason for a velocity restriction is not to maintain hydraulic separation -- the velocity is quite irrelevant to that -- but to keep headloos, noise, and pipe erosion under control. 2 feet per second as a maximum is a handy number to use as a standard for that, that's all.

hot_rod

2- 4 fps is acceptable, lower is better., So if you are on the borderline, accept a bit higher velocity compared to a doubling in price of pipe, insulation and fittings, maybe.

Same applies to sizing a sep, we show suggested sizing based on velocity inside the sep.
There is some wiggle room. We test all our designs on a bench with velocities well beyond standard best practice recommendations.

As for the generous header size reccomendation. Two things we want are adequate flow capacity to all zones circulators running, but the "generous" size also provides separation between one pump to another on that header. There is a separation function of sorts happening between the connections.

Notice on this header different size/ capacity circulators. Could be a blend of high head and low head circs getting along.

Be aware of Sep knockoffs that don't seem to be following the 3-1 rule on that barrel dimension. Without that critical dimension you will get flow out to the secondary. Just as you do with closely spaced tees with too much dimension between the tees. It leaves the installer trying to deal with ghost flow challenges.

StewBrewer

Okay, so lower header fps is better, but do you think the 5 circs will still function correctly if they are ever all running at the same time as long as we are in the lower half (2-3 fps, 20-30 gpm) of the acceptable range on a 2 inch header.

Would be nice to avoid building headers with those 2.5 inch pipes and fittings!

hot_rod

these tables are more accurate as they are for different types of copper. Some of the tables out there are conservative, rule of thumb. These are based on the engineered flowrates.

2" copper L, should be fine flowing 30 gpm, assuming it not 100' long🤔

Do you know the actual flow rate, when all circs are operating at the OP? Or is 30 a guesstimate?

3952E822-5B5F-4CB1-9A03-A19F19A15560.png

hot_rod

what is your heat load? Do you need the full boiler output?

Looks like you model could handle up to 20 gpm. I don't see a pressure drop chart in the manual. I doubt that 15-58 can move 20 gpm?

Here is the sizing chart for the Sep 4. You might go with the 1-1/4 if you think you need 15 gpm or more.

the low loss header kit they sell looks to be 1" pipe size, so?

1C9CCA0D-DDAD-43F1-A82D-B0A3E3254D30.jpeg

42105A9C-292F-4AB3-9F90-FF1B6B2A988B.jpeg

3E3922C8-FEB2-4DF1-A509-BBBC8C7E511E.png

StewBrewer

Thanks for the copper tables - want to make sure I get these headers sized right. The table’s Type M copper numbers seem to agree with the 20 gpm @ 2 fps and 30 gpm @ 3fps for 2 inch that I was working with. I modeled the house in the loopcad software and it seems to indicate that the design day scenario worst case (5 degrees outdoors, 70 indoors, 10 delta t for floors) would be 23 gpm total for all 5 circs with individual circ heads ranging from 5 to 13 ft. The headers will only extend 5’ out of the side of the 4 pipe buffer tank.

Do you think the 2 inch header will work for these 5 circs at the same time in this case?

hot_rod

https://forum.heatinghelp.com/discussion/comment/1743715#Comment_1743715

23 gpm in 2" type M would be under 3 fps, that should be fine.

I would go with an 1-1/4" sep if the 23 gpm is accurate.

Screenshot 2023-03-21 at 2.06.19 PM.png

Screenshot 2023-03-21 at 2.05.22 PM.png

Zman

StewBrewer said:

Thanks for the copper tables - want to make sure I get these headers sized right. The table’s Type M copper numbers seem to agree with the 20 gpm @ 2 fps and 30 gpm @ 3fps for 2 inch that I was working with. I modeled the house in the loopcad software and it seems to indicate that the design day scenario worst case (5 degrees outdoors, 70 indoors, 10 delta t for floors) would be 23 gpm total for all 5 circs with individual circ heads ranging from 5 to 13 ft. The headers will only extend 5’ out of the side of the 4 pipe buffer tank.

Do you think the 2 inch header will work for these 5 circs at the same time in this case?

If you are pushing the 4 fps number with a simple header for circs, you might consider piping them reverse /return. I have seen circs have issues and compete for flow in a header but only when the header was grossly undersized.