Bull head tees

MikeDurigon

Argument in our office. Can you bull head tees on supply and return piping to hydronic radiant perimeter radiation. My argument is flow will take the path to lease resistance. Can it be done on return or supply or not at all. Thank you. 

hot_rod

You are correct, the flow doesn't always split evenly.
I thought the mechanical codes forbid them now? What code and version is used in your area.

MikeDurigon

Canadian Standards Association B214-07. Nothing on best piping practices 

Jamie Hall

In terms of the flow, it gets complicated. If you are feeding into the leg of the T, and out of the two ends and you have only one phase -- all liquid or all gas -- the split will be determined pretty much by the rest of the piping, not the T; the resistance to flow out either of the straight ends is pretty much the same (and pretty much the same as a 90 elbow, incidentally). Similarly, you can go the other way -- in at both ends and out at the leg, and the resistance is much the same. Where things get squirrely is if the flow comes in on one end and then out the other end and the leg, or in the leg and one end, and out the other end. There the leg is at a decided disadvantage, and all else being equal will have much greater resistance to flow than the straight through run. In large sizes -- such as ductwork for AC, for instance, it may be necessary to use turning vanes to get any appreciable flow out the leg if the supply is one end!

In mixed phase work -- such as saturated steam -- the situation is complicated by the fact that the liquid droplets will want to go straight, regardless. The situation is particularly bad if you are feeding both ends and going out the leg -- and the leg is pointed up. The water will accumulate below the leg and be forced up the leg in slugs. Bang... bang... bang.

I have no idea what codes say...

mattmia2

So this is mostly an issue with hydronic work where you need to balance flows more or less for equal output, it is less of an issue for plumbing where you are more interested in adequate flow to each branch rather than equal flow?

GroundUp

Bullheading is poor practice in any form of hydronics, IMO. Sometimes it's a necessity when we have to go left and right to radiation via a single riser without space in the enclosure to bury the crossover, and in most commercial cases the balancing is taken care of with balancing valves anyway, but the turbulence in the fitting always gives me the heebie jeebies. Low velocity is obviously less of an issue than higher velocity but even then, avoiding the bullhead will be better practice IMO.

hot_rod

ideally

EBEBRATT-Ed

Bullheading should be avoided. If the pipe and fittings are oversized it's less of a problem but should be avoided.

The only time I think it is ok is if you only flow through 1 run connection at a time (basically and elbow with no radius)

Jamie Hall

I certainly respect the field learned knowledge put forth here...

But, the interesting thing is that when actually tests the head loss through Ts in various configurations using single phase flow (all liquid or all vapour) it doesn't work out quite that way. Measuring coming in one run and out the other run and the leg, the head loss to the other run is about the same as a straight pipe; the head loss out the leg is about the same as a close elbow. Coming in the leg and out both runs, both runs show only slightly greater head loss than a close elbow would. Coming in both runs and out the leg the head loss is greater on either route than a close elbow, but not that much greater.

Assuming, of course, that velocities are reasonable -- 10 fps or less.

There are some situations -- again from testing -- where you can get some really bad instability, if you have a piping setup where the flow rejoins somewhere downstream (unusual, granted!) -- the flow split can under some conditions oscillate from a very large fraction going one way to having a very large fraction going the other, with the overall head loss going up and down at twice the frequency of the flow oscillation.