Hoffman Differential Loop -- a commentary

Jamie Hall

The other day I answered a comment from one of our best Weallies regarding the Hoffman Differential Loop and why he usually removed them, and on thinking about it I realised that first, my asnwer had been only partly correct but more important, that this is a mysterious contraption which is likely misunderstood.

It isn't really that mysterious -- in fact, it could hardly be simpler. But what it does is really quite remarkable: it ensures that the differential in pressure between the steam mains and the dry returns in a two pipe steam system can never exceed approximately 8 ounces per square inch, and it does this with no moving parts (at least in the Loop itself; there is a vent or vents involved...) and without diminishing the ability of the system to keep delivering heat. What's not to like?

The error in my answer the other day was to agree that the action of the Loop caused delivery of steam to the radiation to stop. This is incorrect. The delivery of steam to any radiator which already has steam in it when the Loop activates will continue at exactly the same rate as before. To see how this is so, compare it with a one pipe radiator with a vent. On a one pipe radiator with a vent, steam delivery doesn't stop when the vent closes. Rather, steam will continue to be delivered at the same rate at which it condenses. Indeed, the same thing happens in a two pipe radiator with a trap once the trap closes. And it also happens if the dry return is pressurised.

We often say that if steam passes a trap in a two pipe system and gets into the dry return that heat from other radiators will be affected -- which is quite true, but only if not all the air has had a chance to leave those other radiators first. If the other radiator is full of steam -- and its trap, then, closed -- it makes no difference what the pressure in the dry return is.

So that is one concern out of the way. As an extension on this, in a closed system with a boiler and radiators, the only difference the absolute pressure inside the system makes is in the temperature at which it operates. If our little demon in the boiler really gets carried away and the pressure rises to -- oh, say 10 psi, other evils may occur, but the system will still make steam in the boiler and still condense that steam in the radiators.

Now the differential pressures in the system are a concern -- hence the invention of the Loop.

There are two main concerns if the differential pressure becomes too large. First -- and the one the Loop was invented to address -- the condensate will not be able to get back to the boiler. It may flood the dry returns (bad) or it may push the water level in the boiler down, perhaps even to the crown sheet (much worse). This is because the driving force to return the condensate is gravity, and as we've often quoted, it takes a water column 28 inches high to overcome 1 psi of differential pressure. Second, the greater the pressure differential across a trap is, the more stress there is on the trap and the shorter and more miserable its life will be.

Therefore we resort to two approaches to keep the differential reasonable. Nowadays, with oil or gas firing, the simplest way to do this is to turn off the burner when the differential gets too high. For better or worse, the way we do that is with a vapourstat, which responds to the gauge pressure in the boiler (the gauge pressure is the differential between inside the boiler and the atmosphere). This is a decent analogue of the system differential pressure, since ideally the dry returns are open to the atmosphere. This didn't work with coal firing: you can't just turn off a coal fire. The Differrential Loop, however, manages the neat trick of responding directly to the differential pressure between the steam mains and the dry returns, regardless of what the gauge pressure in either one is. This is exactly what we want to happen.

In very brief, it accomplishes this neat trick by simultaneously closing the dry returns off from the atmosphere and bleeding just enough steam into the dry returns to raise their pressure to the desired 8 ounces per square inch less than the steam mains.

The Loop itself has no moving parts. Envision a glass of water (or soda... or whatever) with a straw in it. If no one is blowing on the straw, the water inside the straw is standing at the same level as the water in the glass. Now blow very gently on the straw. The water level in the straw drops. Blow a little harder, and you can blow bubbles out of the bottom of the straw (if you didn't do this as a kid, you had a deprived childhood. Sorry...). The Lopp is simply a rather tall glass with a rather long straw in it (14 inches long, in fact). The geometry of the various loop designs varies, but that's the principle. Now connect the straw to the boiler (usually directly to the header) and cap the glass and connect that to the dry returns. If the pressures are equal, nothing happens. As the pressure in the boiler rises relative to the dry returns, the water level in the straw drops. At some point, the pressure in the boiler pushes all the water out of the straw and the steam bubbles up. Now what? Well, the second part of the game is that the main vents -- which normally never close -- are located right above the Loop, and they are the only vents in the system. As the steam bubbles up into the dry return area, it also gets to the main vents -- and they close. Now the system is closed, and the steam bubbling up will pressurise the dry returns -- but only enough so that their pressure plus the water column in our glass is equal to the boiler pressure. Then the bubbling stops until the boiler pressure rises again -- but the boiler pressure can never get more than 8 ounces per square inch greater than the dry return pressure.

There are more considerations, but that's the basics. Next time you encounter a Hoffman Differential Loop, admire the genius which created it. Make sure it's piped correctly, and then leave it alone. It will do what it's meant to do for the next century or so, and it's one less thing for you to worry about.

EBEBRATT-Ed

Great explanation Jamie

JUGHNE

Thanks, Jamie.
If I ever come across one of these, I will refer back to your explanation.

DanHolohan

I wish I had written that. Thanks, Jamie. 

ethicalpaul

And for the steam bubbles to actually reach the top before condensing, all that water in the trap will have to be right at 212 and stay there, so it seems like it would only happen on a real long call for heat if ever. Do you think so?

guzzinerd

Really appreciate the prose. What a great forum.

Jamie Hall

ethicalpaul said:

And for the steam bubbles to actually reach the top before condensing, all that water in the trap will have to be right at 212 and stay there, so it seems like it would only happen on a real long call for heat if ever. Do you think so?

I think so. If the thing starts bubbling fairly soon after a start, there are likely other problems -- like insufficient main venting (but remember, all the venting has to be at the Loop! Use an antler if need be). Even after a long run, it won't be instantaneous -- as you say, it may take a minute or two.

mattmia2

So you could make one of these with some pipe and fittings.

Jamie Hall

mattmia2 said:

So you could make one of these with some pipe and fittings.

Yes you could. Interesting little project.

Steamhead

DanHolohan said:

I wish I had written that. Thanks, Jamie. 

@Jamie Hall had some great examples, @DanHolohan .

ayetchvacker

Isn’t she beautiful?