I think you misunderstood me. It's not a pressure buildup, it's a lack of
pressure. If you think of the collector up on the roof as part of a giant
inverted 'U', with a pump pushing water up one side of the 'U', and gravity
pulling it back down the other side you'll get the idea. Since the water
drains into a tank that is at atmospheric pressure, the water in the
'downward' side of the U is at a pressure *lower* than atmospheric. All the
way up to the top of the 'U'. So near the top of the 'U', if that is 20
feet above the tank, then the pressure there is about 10 psi *below*
atmospheric. At that pressure, the water will 'boil' at about 160 degF, not
the normal 212 degF.
In a closed, non-drain-back type of system, the tank is actually pressurized
slightly (say about 10 psi), so the pressure at the top of the 'U' is higher
by that same amount (say about 10 psi). So the boiling point is higher.
(not to mention that 50/50 glycol/water doesn't boil even at 212F).
Certainly a drain-back system could be designed where the pressure is
maintained even at the top of the 'U', but it would be a bit more
I understood that part. Perhaps what I don't understand is that when the
water flashes that pressure then builds up and that can be damaging. I
remember someone writing about exploding PEX.
It seems to me that what happens is that we have the water and the
air mixed (on start) and it is the atmospheric pressure that determines
what temp water boils at. You have a slug of water falling 20' or
20'/33.95' * 14.7 psi.
So unless I'm wrong, it's not a problem when the down tube is filled,
but when started. Is that correct? Or is this a steady state problem
near the bottom of the drain?
The solution to me seems to be to break the drop. Either with some kind
of resistance or with traps. Or not start it when the temp is greater
than 170F. If it's a steady state problem then the solutions would be
What kind of bit is that? Since my basement is 20' below my roof, it
seems like something I should find out!
I don't think that could happen in this type of drain-down system. From the
collector down to the tank, there is no obstruction and the water just
'falls' down. This is why the pressure at the top is lower than at the
But certainly in a closed-loop system (not a drain-down), if the flow is
stopped, the water in the collector can get hotter and hotter until it
starts to boil even at the pressure that's being maintained by the expansion
tank. This could increase pressure and burst a pipe.
It can be a steady-state type of problem. All depends on the absolute
pressure at the top of the inverted 'U' (i.e. collector outlet).
But it could be a problem too if the collector has been in the sun a while
with no water. Now you have cold water being pumped up to a very hot
collector. The first bit of water to reach the hot collector may be heated
above boiling, forming a 'surge' of steam/air that has to come out
somewhere. Best chance is to come out the return piping down in the tank.
As more water reaches the collector, (assuming the pressure surge hasn't
stopped flow completely), the collector cools and the 'event' is over. But
the formation/collapsing of steam can cause a lot of 'bangs' and 'clanks' in
the piping/collector that could damage things. Best is to keep the water
circulating while the sun is out.
Yes, that will solve the problem, but now you need more pumping power to
overcome the resistance/trap. From the standpoint of avoiding 'boiling',
you want the pressure at the top of the 'U' to be atmospheric or higher.
That means that on the downward leg, the pressure right before entering the
tank (presumably where the resistance/trap is located), the pressure would
be 20' of head *higher* than atmospheric. So the supply pump must have a
discharge pressure that is always <pressure at top of 'U'> + <rise of supply
pipe> + <some friction losses>.
In a closed system, the pump only has to overcome the <some friction losses>
because the pressurized expansion tank supplies the other terms.
A lot of drain-back systems do exactly this. The pump has enough head to
maintain pressure at the top of the 'U' at slightly over atmospheric
pressure, and a small float valve mounted at the high point lets air out of
the system until water is pumped up that high. The weight of the water in
the return leg isn't used to help the water circulate and the pump does more
Closed loop systems with a pressurized expansion tank are much simpler.
Just keep enough pressure in the tank so that water/glycol is always forced
up to the collector. Then the pump just has to overcome the friction losses
of the piping.
Open loop systems need a pump that will get water all the way up to the
collector. If the collector is vented by a float at the top, then the pump
must maintain this pressure all the time water is circulating. If the
collector is not vented, and you want to use the return line to help
'siphon' water through the system, then pumping requirements are lower, but
you have the problems I've been discussing.
But another complication is just how much electric power it takes to pump
the water with/without the siphon action. In theory it can be quite a lot
different, but in practice, some small pumps (such as used here) draw about
the same amount of power between zero flow and full flow. So the difference
in electric pumping power required between an open system that keeps
pressure above atmospheric at the collector and one that uses siphon action
may be trivial.
The vast majority of drain-down systems seem to be the type that maintain
pressure at the collector with an air vent at the high point. No problem
with boiling (as long as the pump is running) and although pumping power is
*theoretically* higher, practically I don't think there is much difference.
Any type of open loop drain-back system does require a pump that can develop
enough pressure to at least reach the high point with very low flow (at
least to flood the collector). Closed loop 'circulator' pumps don't have to
develop nearly that much pressure.