This is a quick test to see how well a solar collector that uses a serpentine plumbing arrangement will drain in a drain-back style system.
Serpentine collectors use a piping arrangement that looks like this:
The collector uses a single pipe that winds its way down from the top to the bottom. So, there are a bunch of near horizontal runs with bends connecting the horizontal runs. A nice feature of the serpentine collector is that only a single run of pipe collects heat for the whole collector -- there are no top and bottom manifolds with many vertical risers connecting them.
I would like to use this arrangement on the PEX collector I'm working on, and would also like to use a drain back system. So, the question is, will a serpentine collector reliably drain back to the storage tank when the pump stops?
I've been running a prototype PEX serpentine collector for a couple months, and it has always drained back quickly and reliably, but I thought it would be worthwhile to look into this further with a test on a serpentine piping arrangement that uses clear tubing so that you can see the actual drain back process. It turns out to be an interesting process (at least to me).
The picture shows the test setup. The bucket has a small submersible pump that pumps water into the lower end of the serpentine arrangement of clear plastic tubing. The water returns to the collector via the pipe coming down from the upper left corner of the collector. The blocks of plywood just guide the clear poly tubing into the desired shape.
When the collector backing board is level, each run of the collector has a slight down slope (about half an inch of the length of each horizontal run).
To do each test, the pump is turned on until flow is well established. Then the pump is turned off to allow the drain back process to be observed.
Note that for drain back to occur, water must be pumped into the bottom of the serpentine run, and the return pipe must be kept above the water level in the bucket (this is true for any drainback system).
The tubing is mostly half inch ID, but the last section (slightly more white in color) is 3/8 inch ID. There are two tubing splices that introduce some flow resistance.
The initial test was done with clear tap water. After that, a tiny bit of dish washing detergent was introduced to see the effect of reduced surface tension. Some ink was also added to be able to see the fluid more easily.
Tests were done with front to back tilts ranging from 70 degrees (nearly vertical) down to about 20 degrees. In addition, in some tests the panel was tilted to the left or right to see what happens if the panels is not installed level.
After flow is well established, and the pump is turned off to start the drain back, here is what happens:
A bubble or air front begins at the bottom of the return pipe (this is why it must end above the water level).
This air front advances up the return pipe until it gets to the top of the collector.
The air front then advances down the collector at a fairly good pace -- about 1 ft per second for this collector.
Nearly all of the water in the tube drains during this phase, which lasts about 15 seconds on the test collector.
There is a little water left in the tubing after the air front passes, and this water drains in a small trickle along the bottom of the tube as long as the tube has a down slope.
This 2nd phase of the drain process takes longer -- typically about a minute.
This is a picture of the air front advancing along the tube:
The air front is near the right edge of this picture. Water to the right, air to the left.
This air front advances from the top of the collector to the bottom quite quickly and cleanly. It is even able to advance without any trouble when the collector is tilted so that it has to advance slightly uphill.
Most of the water drains during this phase. The water left behind after the air front passes drains in a slow trickle over time.
In this test, the collector is set up with a tilt of about 70 degrees (nearly vertical), and the collector backing board is level in the left to right sense, so that each "horizontal" run of the collector has a slight down slope.
In this configuration, when the pump is shut off, the collector drains in about 15 seconds. There is a very well defined air front that advances up the return tube, and through the collector. The air front advances in what appears to be a very reliable and repeatable manner.
The bit of water left in the tube after the air front passes drains down in a 2nd phase that goes on for a while (perhaps a minute or so). At the end of this, it appears that all the water has drained, although there may still be a very thin film on the pipe walls.
I tried this test initially with plain water and then with just a couple drops of dish washing detergent in the water. The idea was to see if the lower surface tension effected drainage. There did appear to be a little improvement. The detergent probably made more difference on the later tests, but I did not do these with plain water, so don't know for sure.
In all this appears to be a pretty robust drain back that I would feel pretty confident in working over the long haul.
This is the same as test 1, except that the collector tilt is reduced to about 35 degrees.
The results are pretty much the same, but the drain back time is greater -- about 20 seconds.
The secondary drain process after the air front passed left some very small pockets of water in the tube -- little beads of water, about 1/8 th inch in size.
So, not surprisingly, steeper tilts result in slightly faster drain backs, but the drain back process still seemed very repeatable. I am sure that if you kept reducing the tilt you would get to a point where the collector would not reliably drain back.
Note July 18, 2008: did one last test at a very low tilt of 22 degrees. This worked fine, no increase in time, and no hesitation even with this low tilt.
This is the same tilt as in test 1 (70 degrees), but the collector back board is not level in the left to right sense. The backing board was adjusted so half of the straight runs had a slight uphill tilt.
The first phase of the drain back appears unchanged from test 1. The air front advances along both the down hill and uphill runs at what appears to be the same pace as in test 1. The reliability of this phase of the drain back appeared to be unaffected by the adverse slope on half of the straight runs.
The 2nd phase of the drain back results in some pockets of water collecting at the start of each uphill straight run. None of these got to the point of filling the tube diameter. Most of the tube length was full drained, with these small pockets of water at the start of each uphill run.
This is the same as test 3, except that the left to right tilt was increased to about 5/8 inch over the length of the straight runs.
So, every other tubing run slopes up by an amount greater than the tube ID.
The first phase of the drain back appears unchanged from test 1. The air front advances along both the down hill and uphill runs at what appears to be the same pace as in test 1. The reliability of this phase of the drain back appeared to be unaffected by the even more severe adverse slope.
The 2nd phase of the drain back resulted in pockets of water at the beginning of each of the uphill runs that got large enough to block the full diameter of the tube. Here is a picture of one of these.
With 5/8 inch of adverse slope, the 2ndary phase of the drain back results in pockets of water that grow to the full diameter of the tube for a short distance at the start of each uphill leg.
While I see no reason to subject the collector to this much out of level abuse, I don't think that these little pockets of water freezing would be likely to cause any damage to a PEX collector, since there is plenty of expansion room for the freezing process. It might delay startup slightly on the next morning, since these pockets would need to melt out to allow flow.
The drain back process in this test appears to be pretty robust and reliable. This is reinforced by about 3 months operation of the prototype PEX collector without a hint of any drain back problems.
It appears that it is possible to use drain back in a serpentine collector as long as:
The collector has a significant
tilt angle. The shallower the collector tilt, the more risk that the
collector will not drain reliably.
All of the straight runs in the
collector have a slight down slope. While most of the water in the
collector will drain even if some of the straight runs are horizontal or
even have a slight up-slope, its seems prudent to have all legs slope
The supply and return plumbing
must all slope down toward the storage tank (as for any drain back system).
The pipe that returns fluid from the collector to the storage tank must terminate in the airspace above the water to allow the air bubble to advance up the pipe (as for any drain back system). And, the storage tank (drain back tank) must be vented to atmosphere.
I suppose that one caveat is that if you developed significant deposits inside the tube, it might effect drain back in the long term. To me, this seems unlikely in a well designed system, and would eventually lead to other kinds of problems (like blocked tubes).
Any thoughts, comments, or suggestions on this would be appreciated -- Gary ...
July 15, 2008, September 12, 2008