Flow Distribution Test -- Copper/Aluminum Collector

The test is intended to check the evenness of the flow distribution through a collector made from copper pipes in which both the risers and the manifolds are built from half inch diameter pipe.

The copper/aluminum collector uses half inch copper pipe top and bottom manifolds. This allows the whole collector array (risers and manifolds) to be built from half inch copper pipe, and to use inexpensive standard half inch Tee fittings for easy and cheap assembly.

Commercial collectors usually use 3/4 inch or 1 inch manifolds, so the question is, do the half inch manifolds on this homemade version provide a sufficiently even distribution of flow in the collector, or is there a tendency for the risers near the supply end of the collector to get more flow at the expense of the risers at the far end of the collector?

An additional question is how wide can this kind of collector be made without the half inch manifolds causing an uneven flow distribution?

To get an idea how even the flow is, I removed the glazing on the prototype collector, and then set up a flow of water through the collector and let it go until things stabilized. I then measured the surface temperature at the top and bottom of each riser using a non-contact, IR thermometer.

The idea is that areas getting lower flow will be hotter because there is less flow to remove the heat. If the risers at the supply end of the collector are getting more flow, they should be running cooler, and this should show up in the surface temperature measurements.

Collector with glazing
removed

Measuring temperature of
absorber surface at riser tube

Absorber grid before
installing aluminum fins.

Note: If you can think of a better way to check the flow distribution through the risers, please let me know...

The Test:

February 7, 2009

Clear and sunny with a light and variable breeze.

Tambient = 41F

Tcollector surface with no flow = 124F to 127F (depending on breeze)

Flow rate = 1.5 gpm or 0.047 gpm/sqft

Top Manifold
79F	85	85	88	87.5	104.5	82.5
Riser 7 +9.5F	+15.5	+13.5	+14	+12.5	+28	Riser 1 +13
69.5F	69.5	71.5	74	75	76.5	69.5
Bottom Manifold

The table above is arranged like the copper tube array for the collector -- each column represents a riser. There 7 risers numbered 1 to 7 from right to left.

The manifolds that deliver water to the risers are marked "Top Manifold" and "Bottom Manifold".

The supply water from the tank enters the bottom manifold at the lower right corner, then along the bottom manifold with some water flowing up each riser. The hope is that the flow divides evenly, and a roughly equal amount of flow goes up each riser.

The flow from each riser is collected by the top manifold, and exits the collector at the top left corner.

The top row of numbers are surface temperatures taken near the top of each riser.

The bottom row of numbers are temperature taken near the bottom of each riser.

The row of numbers across the center of the table show the temperature rise from the bottom of riser temperature to the top of riser temperature.

So, for example, riser 1 has a temperature near the bottom of 69.5F, increasing to 82.5F at the top for a 13F rise.

We attempted to get all the measurements within a short period of time, but there is probably some variation over the about 1 minute it takes to get the readings. There are also variations due to small changes in collector construction at the spots measured.

There are other little inconsistencies coming from slight variations in breeze, sunlight, the IR thermometer itself etc. -- but the data appears to be consistent enough to reach some conclusions.

Conclusions:

1. Overall, the top and bottom of riser temperatures, and the temperature rises are pretty even. This indicates to me that the flow is fairly even through the risers. So, from a flow distribution point of view, the half inch manifold appears to be working OK.

The exceptions to this even flow are:

Riser 7 (the left most riser) appears to have a consistently lower temperature rise, indicating that it is getting somewhat more flow than the other risers. This may be due to the fact that the manifold to riser connection for the last riser is an elbow, while the other manifold to riser connections use Tees -- the rounded flow path of the elbow may result in more flow.

The riser 2 temperature rise is consistently greater than the others. This may indicate that there is some kind of partial blockage -- perhaps a solder blob from my sloppy soldering? The fact that it is still substantially lower in temperature at the top of riser 2 than the surface temperature with no flow indicates that it is still getting enough flow to work OK .

2. It appears to me that the collector could be wider than it is, and a half inch manifold will still given even enough flow to work OK.

I'm not sure what the limit is, but I would think that a 6 ft wide collector would be OK.

A caution:

Commercial collectors are sometimes connected in banks by coupling the manifolds together end to end. With the large, 1 inch manifolds, as many as 5 collectors can be connected in this way. This should not be done for collectors with half inch manifolds.

One fairly compact arrangement that would work with half inch manifold collectors and give a large collector area is to:

Make two large area copper pipe collector arrays -- perhaps 6 ft wide by 10 ft high each for a total area of 120 sqft.

Place the two arrays side by side -- they could even share a common housing.

Arrange plumbing so that the supply for the left collector is at its lower right corner, the supply for the right collector is at its lower left corner (so the two supplies are right next to each other at the center of the housing).

The return for the left collector is at its upper left corner, and the return for the right collector is at its upper right corner.

So, the two supply lines are right next to each other, and can share a common supply line. The two return lines could be run to the center of the collector, and be Tee'd into a shared return line.

The shared supply and return lines should be at least 3/4 inch diameter. For a drain back collector, these lines could be PEX.

If you need more collector area, than this arrangement could be replicated multiple times.

Remember that if this is a drain back system, the copper pipe arrays must be "clocked" a bit in the housing so that the lower manifolds slope down toward the supply line. So, the left collector is rotated a bit clockwise, and the right collector a bit counter clockwise. This allows the collector to fully drain for freeze protection. The supply and return lines back to the storage tank must also be sloped toward the tank.

Gary February 7, 2009