Grundfos Alpha circ

Mike T., Swampeast MO

Yes, the "open" I've been using refers to original gravity piping and valves.

But there's another "open" that comes into play. Gravity systems were originally open to the atmosphere--the only pressure in the system being that of the weight of water above.

In that case, gravity acts rather like a true pump--"lifting" the water, not merely circulating. But because this lifting force grows with elevation it had to be addressed by introducing intentional restriction the higher you went in the system.

Closing the system to the atmosphere is another step of the conversion process. While the open expansion tank (usually in the attic) was partly a safety device, I believe that also allowed gravity to have its' full "pumping" ability.

If "ghost flow" were sufficent to actually heat structures with modern piped systems, we'd still be using gravity for space heating--not just the fairly common use for DHW recirculation.

Yes, I know there are those weird gravity systems without an attic expansion tank. I've even worked on one. It was poorly balanced and extremely inefficient. Much of that probably came from age. There was however an overflow pipe that connected somewhere high in the system, went above the highest rad then looped back down to a basement floor drain. Still an open system even if there was no "real" expansion tank.

Has anyone ever tried a relatively low capacity circulator in a wide-open, multi-floor gravity conversion system of fairly good size? This as opposed to the much more typical B&G 100 or Taco 007 that moves 25-30 gpm or so under extremely low head conditions...

If so, did it work?

Mike T., Swampeast MO

You Might Have Forgotten Something Brad

IOW: Say a 2-inch main carrying 5 gpm. Velocity is less than 0.5 FPS (yawn) and head loss is about 0.074 feet per 100 feet of pipe (double-yawn). Branch one drops off a gallon per minute. Your next segment carries 4 GPM at a velocity of less tha 0.4 FPS. Head loss drops by the square to 0.047 feet per 100 feet of pipe.... etc.

As flow diverted through branch lines, the mains reduced in size proportionately--this helped keep that constant velocity that seems VITALLY important to producing a nicely balanced gravity system.

Even if the system was sized via mathematics (and that appears uncommon in residential systems), the dead men STILL considered the "cross-sectional" area of hand valve opening when sizing their mains--even if such was "buried" in the math.

Once enough valve cross section had been taken off such that it was greater than the cross-section of difference between the current main and the next smaller size, they used a smaller main. This was not about cost savings of smaller pipes! It was about maintaining relatively constant velocity throughout the system.

Seriously--they were using a gravity PUMP, not a gravity CIRCULATOR!

The exception to the "constant velocity" rule is long horizontals--be they branches or mains the dead men KNEW by experience that their gravity pump ran out of head if they had to move the water horizontally for any significant distance. It's that bas ackwards way of looking at things again. To compensate for horizontals they increased pipe size by at least one factor. With larger systems this was impractical if the mains had to travel any significant horizontal distance so they used PAIRS of mains to reduce cost by using readily available and reasonably threadable pipes.

Brad White_58

Flow versus Velocity

Let me see if I can grasp what you are saying here, Mike. I am not sure if we agree, disagree or are seeing the same thing differently, but I like the perspective and stretching things a bit.

We agree that the verticals had more "lift", the rise of supply water being negated by the fall of return water. And to some extent the mains and risers for that matter, do have a seemingly constant (yet variable due to available pipe increments) velocity.

And we agree that splitting of the mains to maintain reasonable use of available pipe was probably the dead men's rationale. Larger horizontals to compensate? Maybe. I see them where I do as being roughly proportional to the risers they serve, not overly large as compensation. They did pitch afterall. Although the cost of tees and reducers versus straight pipe does seem to contribute to a larger pipe size continuing to serve just one more branch set...

But I think the velocity is incidental to balancing, important in an oblique way, (for it affects pressure drop) but not vital. The velocity pressure or velocity head as it is sometimes expressed, is a minor factor, especially when the velocities are low.

Remember too that pressure drop is an exponential function of flow. (Some say by the square -I do-; Siggy says to the power of 1.75, something I was meaning to ask him about). Point being that higher velocities yield exponentially higher dynamic losses at fittngs, turns and tees. Cut the flow rate in half, the PD drops to a quarter of what it was. Double the flow, quadruple the pressure drop.


If I have 7.0 GPM in a one-inch Type L new copper pipe, the velocity is 2.72 FPS. Friction per 100 feet is 3.60 feet.


The same 7.0 GPM flow rate in a 1-1/4" pipe of same material, velocity is 1.79 FPS. Friction per 100 feet is 1.32 feet. One foot per second less and less than 37% of the pressure drop.

In true gravity systems the velocity was typically higher than when using a circulator as we do today. The larger the pipes the higher the velocity because it could! So on this score I can see how velocity was to be restricted somewhat, especially on the upper floors where the pressure gradient was higher.

By inserting in a restrictor/orifice plate or an OS fitting but especially the restrictor, velocity was slowed, hence flow... so it really is not about velocity but that it has to be controlled in a given pipe size as a function of flow. (Circular logic but this is all about circuits! Chicken, egg, cart, horse...)

It all boils down to the key to balancing being pressure, approximate equal pressure all to assure a given flow to a given point. How fast the flow gets there is incidental although to kill the extreme, we want it there sometime soon!

Back to today: A converted gravity system with a fixed circulator probably moves less water than the former "gravity engine", hence at a lower velocity. It still remains that pressure differences in each circuit to each radiator for a given flow rate have to be equal (to confuse the water into not caring to find the path of least resistance for they would appear to be all the same). And the flow rates have to be sufficient to serve the radiators output to match the heat loss.

So it still all boils down to flow rate (and temperature of course). Velocity really remains incidental. When I balance a system today, I do not look at velocity, just pressure across a known orifice (balancing valve at a given position) and cross check that with temperature.

I may have a line, say 2-inch, carrying 35 GPM to a coil, and another line, say 2.5 inch, carrying the same flow to another coil. Both will see the same pressure drop once I close down the balancing valve to the 2.5-inch fed coil. Velocity does not enter into it for me. Maybe I am missing something in what you are saying? Anyway, I enjoyed the brain stretch!

Best Always,

Brad

Mike T., Swampeast MO

In true gravity systems the velocity was typically higher than when using a circulator as we do today

Sorry, but by no means!

Velocity was the root of evil to the dead men. Increasing velocity meant increasing friction--increasing friction rapidly consumed the meager forces of gravity water flow. The theoretical velocity was almost always MUCH higher than the actual. For example--at 30° DT with 20' of elevation the theoretical velocity at the TOP of the system is about 3.7 fps. One floor down at 10' elevation it is about 2.7 fps.

"The actual velocities of flow as shown by test results seldom approach the theoretical and are usually very much less, ranging from 15 to 35 percent of these theoretical values, depending on the friction of pipe and fittings. It is therefore impossible to make any practical use of these theoretical velocities in designing a hot-water piping system, since the actual velocity in any given gravity circulating system under steady flow conditions will always be such as to make the friction head just balance the head available. If the friction head is small, in a system, with a given head available, the velocity will be high, and, conversely, if the friction head is large in a system, with the same head available as before, the velocity will be low. The range of velocities may be, relatively, very great, depending upon size and arrangement of piping between heater and boiler."

That said, the range of acceptable velocities in gravity systems was generally from 0.3 fps to 1 fps.

As said above, designing for a constant system velocity was impossible, BUT nearly all of their piping methods had the effect of keeping velocity as consistent as possible. Just like us the dead men designed around a theoretical, constant delta-t in the radiation--20°-30° for typical open systems. Say there were two identical rooms with identical rads and identical heat loss on two different floors of the same house. Each is supplied by dedicated branches. The higher rad will have smaller branches. Why? Because more gravity force was available--more force meant more velocity--more velocity meant more flow--more flow meant lower delta-t through the rad--lower delta-t through the rad meant higher rad output. They did not want this, so instead they used a smaller pipe to keep the velocity down by intentionally adding friction. If for some reason (like a number of fittings or long horizontals) the next pipe size down had a bit too much friction, they would instead use the "too large" pipe and install one of those restrictor plates in at the radiator--reducing velocity by adding restriction but little friction.

Their mains were similar--again as more and more flow was diverted to the rads, the pipes were sized down thus keeping velocity in the mains relatively constant. This made it MUCH easier for the dead men to estimate head loss in the mains because they didn't have to keep re-computing friction for different velocities in each main section.

The pipes weren't big because they had high flow--they were big because they had to keep friction VERY low.

The GREAT benefit of forced circulation is that it allows much smaller pipes to carry the same amount of flow. Same amount of flow in a smaller pipe means higher velocity in forced systems--not lower!

The only way the dead men could increase velocity significantly in a gravity system was by increasing the force of circulation. They could do this by using significantly hotter water in a pressurized system because pressure allowed them to go near or even above the boiling point of water--aka the Honeywell "Heat Generator" and similar. Some British systems used EXTREMELY hot water, VERY high pressure and VERY small pipes--after spectacular failures these systems fell from use and were converted to forced flow--the ONLY way they could get them to work with those small pipes.

You said: Back to today: A converted gravity system with a fixed circulator probably moves less water than the former "gravity engine", hence at a lower velocity.

Again, by no means!

Take a gravity system of 1,100 sq.ft. EDR and a boiler with a net output of about 190 mbh operating at 25° DT. Flow would be about 15 gpm. Convert to forced flow and use the typical (and almost universally recommended) B&G 100 or Taco 007. Those pumps will easily move 30 gpm through the piping that would have been used for a system of that size. Reasonable flow balance is achieved because of INCREASED velocity that increases friction in the ENTIRE system to help negate that bas-ackwards gravity sizing that had high, far rads with the greatest friction and low close rads with the least. Use a pump rated for 15 gpm at the original design friction losses and velocity and you'll be quite lucky if the system has even reasonable balance.

BUT, there's still another problem. Then as now, heating systems were designed quite conservatively. Even without insulation/weatherization measures, a well-sized replacement boiler will likely be significantly smaller. Add reasonable insulation/weatherization and the replacement boiler will be MUCH smaller--say 80 mbh.

Keeping with the 25°DT, it now only takes about 6 gpm to supply the system. If you vary the speed of the pump to keep a constant DT, I'll state with 99.9% assurance that flow balance will be completely destroyed. Go back to the 30 gpm or so of the typical circulator and DT plummets to 5°! Such conditions make it quite difficult to get reasonable efficiency from the boiler--particularly if it does not modulate. Even if it does modulate, it's still difficult to maximize efficiency unless you also take flow control measures, e.g. TRVs. Why? Because you're not going to move 30 gpm through any properly sized mod-con so you must use primary/secondary or a low-loss header. Both measures loose efficiency when multiple times more water are moving through the radiation circuit than through the boiler circuit. Of course you could always re-pipe the ENTIRE system for forced flow but you'll spend a LOT more money than by using TRVs--and the TRVs will produce a system superior to re-piping without them anyway...

If this circulator becomes available in the US, it would be ideal for gravity conversion systems--IF they have TRVs! Such would be WONDERFUL with mod-cons using primary-secondary or LLH. Of course there is at least one boiler that usually doesn't need primary-secondary or LLH on such a system to begin with.

To quote something VERY IMPORTANT again, "...the actual velocity in any given gravity circulating system under steady flow conditions will always be such as to make the friction head just balance the head available..." [emphasis added]

This is EXACTLY what TRVs on a gravity conversion system driven by a condensing/modulating boiler and variable-speed circulator achieve--the ORIGINAL design conditions! To my knowledge there is no other way to do this other than proportional flow control and TRVs are the least expensive, most reliable and simplest method!

Brad White_9

Short response on Velocity in Gravity

To quote Dan Holohan himself from the Q&A section:

"Q: What determines how fast the water moves?
A: Several things. First, there's the height of the system. The taller the building, the quicker the flow. Within reason, of course, because if the building is too tall, the water will cool and slow circulation to the upper floors. A three-story house is the practical limit for gravity hot-water heating.

And then there's the size of the pipes. The larger the pipes, the faster the water will flow. This is because large pipes offer less resistance to flow than small pipes. It's also the reason why the old-timers used two supply and two return tappings on their boilers."

More later but I wanted you to know one source. Counter-intuitive but it also shows that velocity was less of an issue to create, "it just was".

I wish you would show me how to do italics, Mike. Is it an HTML trick? I am ignorant on that.

Thanks

Brad

Mike T., Swampeast MO

Bold and italics are extremely simple HTML. I'll send you the instruction via a text e-mail because if I do it here, it happens.

Guess I didn't go back far enough in the gravity sizing scheme...

Dan's statement isn't counter-intuitive at all. Bigger pipe means less friction which means velocity is closer to the theoretical limit.

But that's only PART of the story. Move water at a higher velocity through a larger pipe and you also get more water! More water means more heat!

The FIRST part of sizing gravity piping (be it the branches or the mains) is determining how many pounds per hour of water you want to move--through each branch and through each section of each main. It's this pounds per hour that determines actual output of the system--remember, the dead men made the same sort of delta-t assumptions that we do.

Once you know the pounds per hour you want to deliver to each and every radiator, the balancing act begins. Since the forces involved are so slight, this is MUCH more complex than designing for forced flow. That's why a well-balanced gravity system is a true work of art that came not only from math, but from experience AND observation.

To achieve the desired pounds per hour, two main factors come into play--the available force (determined by elevation & delta-t) and the friction of the piping all the way from the boiler, through the radiator and back to the boiler.

Velocity and friction here are inextricably linked. To determine one, you have to know the other so the dead men made a guess (usually velocity within some acceptable range) and then computed friction at this velocity. If friction was too high, they had to recompute with a larger pipe size; if friction too low, recompute with a smaller pipe size. Then they had to compute pounds per hour delivered to see if it was still close to their desired. Risers serving rads on multiple floors were especially complex. Then, after doing this for ALL of the radiators, they had to make sure that their theoretical pounds per hour delivered through all of the rads was a good match for their original system pounds per hour requirement and main sizing!

Again, think of two identical rooms with identical loss and identical rads on different floors of the same house. Both have quite direct routes back to the mains. The upper rad had approximately twice the available force as the lower, yet both need to receive the same number of pounds per hour of water. Use the same size branch piping for both and even though there's more vertical pipe on the way to the second floor you'll get significantly more pounds per hour because the gravity force increased faster than the friction. SO--they used a smaller pipe to intentionally increase friction to "eat up" the increased force moving the water! Now, since this pipe is smaller but moving the same number of pounds per hour its' velocity IS higher than the one serving the rad on the lower floor. This is the COMPLETE opposite of the way forced-flow systems are designed. It really is bass-acwards and rather counter-intuitive.

I probably confused you by saying the dead men designed for constant velocity. They did do this (in a sometimes coincidental way) but there were different "constant" velocities in the same system depending on elevation and piping layout. Again, they tried to keep velocity in their mains constant as this GREATLY simplified calculations--it also helped reduce material cost. If these mains were long and horizontal, they usually increased pipe size above the theoretical requirement thus reducing velocity and friction but not increasing pounds per hour of flow--the same is true for long, horizontal branches.

While pipe sizes were limited, velocity to directly and mainly vertically supplied rads on a given floor would be rather similar. If their calculations showed that velocity in a given branch was too high but a smaller pipe had too much friction, they would add intentional restriction via those little plates with holes at the rads. That's why you'll only find these restrictors on UPPER rads where gravity force was considerably higher. For rads on the lowest floor (particularly those close to the boiler) they had to resort to "tricks" to get their required pounds per hour and those tricks did NOT involve adding any sort of restriction or friction...