This page covers a set 6 of the copper pipe/aluminum fin collectors built by Chad for space heating. The collector design includes a unique hinged mount on the top of the collectors that allow the collector tilt to be adjusted by season. In addition, Chad shows an alternative way to build the collector frame.
Thanks very much to Chad for providing this material!
Chad's bank of six 4 ft by 8ft collectors
NEW: November 2011 update:
Chad reports that the system is running well and without any significant problems or maintenance.
Per his logging system, the
system has produced 31 million BTU, while using 570 KWH to run the
Gary’s copper – aluminum collector design was used as a baseline.
It was modified as follows:
The frame was made with the frame members cut to
eliminate the square absorber mounting strips. Square strips were used
under the glazing as per Gary’s design.
did not put the 3” aluminum strips under the copper tubes. My testing
showed the aluminum transferred heat to the copper very well without them.
They would help get the heat to the copper from the outside of the aluminum
panels, as the aluminum path is quite long. As copper has dropped to 50% of
what I paid for the first panel, the final 3 panels I assemble will use 9
vertical copper tubes with 5.5” wide absorber strips.
I double glazed the collectors. I calculated the R value from the aluminum to the front of the glazing to be 2.3. The second layer increases this by 1.6, a significant improvement. The payback is positive when the air to water delta is above 40-60 degrees, with it making a bigger improvement at lower power output levels.
Pictures of the collector frame components -- Click on pictures for full size -
The frame members can be cut easily and quickly if you have a table saw.
Save the waste from the 2x4’s for the side glazing
strips and top and bottom hold down strips.
Pre-drill the conduit holes, at 1/3 and 2/3 way, at the top edge ( the 1.5” wide part). Drill ¾ way through the wood. The holes are not detailed in the drawings.
To assemble the frame, I put the absorbers aluminum down and spaced them up 1 ½ inches. The 4 pieces were placed around the absorber so that the absorber was resting on the notch. The conduit was placed in the holes at this time. Next, the corners were drilled and screwed. After the frame was secure, the absorber was screwed to the frame at 5-7 spots for each frame member. Next, insulation was cut / drilled to fit and secured with expanding foam.
Click on pictures for full size
The collector frame with the copper pipe grid and
absorber fins ready for painting.
The two sheets of polycarbonate that form the inner layer of glazing layer were not siliconed together as I didn’t see much benefit in doing so. It was placed on the absorber after putting a silicone bead on the conduit. To separate the two layers, I cut ½” pieces of ¼” vinyl tube. I placed these about 4” from the top and bottom and every 10” across. They were secured to the inner glazing by a dab of silicone. I placed a piece every 12 inches on the sides, one rib from the edge. I then scattered them on the rest of the surface. In reality, I put down more separators than needed. The outside layer tends to bow out due to expansion and separate from the inside. They were secured in the same manner as Gary’s. When drilling, ensure wood particles do not make in between the 2 layers.
Vinyl tubes that separate the two layers of glazing.
The tubing runs parallel to the glazing ribs.
Drawings -- click on picture to view dimensioned drawings for the collector... (pdf files)
2X6 Top Rail
2X4 Bottom Rail
The ridged siding used on my shop and
the general design of a pole barn complicated mounting the collectors. There are
3 horizontal 2x8 pieces under the steel siding. There are also vertical 6 x 8”
posts every 10’. I preferred to be able to change the collector angle as needed.
I ended up with the mounting detailed here. I used stainless lag bolts to secure
door hinges to the collectors. Corrosion resistant fasteners are supposed to be
used with the new treated lumber. I mounted the hinges to the collectors,
positioning each to avoid the ridges in the steel. The attached 2x4s serve two
purposes: to space the collector off the side and to reach the shed 2x8.
Click on pictures for full size
For the hinge mount collector, I ran the copper lines out the top and bottom. To ensure the top line will not interfere with the wall at maximum tilt, I used a 45 degree fitting at the top of the collector to provide more than enough tilt.
The variable hinge mount requires flexible water lines. Multiple options were considered:
CSST (corrugated stainless steel tubing) natural gas tubing:
The tubing is reasonably priced. The fittings are $5 each. At 4 fittings per panel and the mating hardware, this got expensive quickly. The corrugated tubing will also decrease fluid flow. CSST is not supposed to be used for flexible gas installations. I suspect it will work harden after many flexes.
I did not know if this would prevent oxygen from entering or not. It is probably more expensive than CSST, should take the temperatures and definitely the pressure from the system.
This is the cheap and easy to install. The downside is possible thermal expansion, overheating, and sun exposure. The PEX is presently covered with pipe insulation to insulate and keep the sun away. The panel to panel PEX may be replaced with copper pipe to increase flow and decrease expansion movement.
The system is a closed system with
semi drain back capability, that is most of the coolant will drain back to the
house (200’ away). Some of the lines will probably retain coolant. I will have 2
tanks: one in the basement to collect solution and one in the shop ceiling to
hold the gas (argon / CO2) than fills the collectors during drainback. I plan to
use a small air compressor to charge the basement tank sufficiently to fill the
lines when heat is available and a air solenoid valve to discharge it.
I used a 45% solution of Noble Company’s NoBurst HD with 55% distilled water. This is a very concentrated propylene glycol fluid. It ran $140 / 5 gal bucket to my door. This will provide for fluid flow capability to below –15 F. Note, this is thicker than motor oil, so pre-mix it before putting it in the system.
After the system was detergent cleaned, rinsed, and emptied, I used an a/c vacuum pump to evacuate the air. I then connected a hose to the system and let it suck water and coolant into it. A small pump was used to pump the final bit into the system to pressurize it.
The pictures show the details of the hinged collector mount along the top of the collectors that allow the collector tilt to be adjusted by season.
Chad painted the absorber with a selective coating called Solkote. The advantage of this material is that it is a selective finish, meaning that it absorbers visible light very well, but also has a low emissivity in the infrared so that it does not radiate (lose) heat as much as a conventional paint. It is also a good, high temperature finish that will hold up to the high temperatures inside the collector.
SolKote's low emissivity depends on it being applied by a sprayer in a very thin coat.
This was the first time really using
a paint sprayer for anything that mattered. The paint is very thin, it is xylene
based, which seems to be a little less volatile than lacquer thinner. It seems
like they took powder carbon and mixed with the solvents ... probably not that
SolKote wanted it applied to bare aluminum as this helps the emissivity, (not practical) and they were concerned about if it would react with the paint. I tested the reaction by soaking a rag in xylene and sealing to the al for a few days. It mildly softened the paint and one could then scrape it easily, but once dry there was no damage.
I used an cheap automotive HVLP gravity feed gun from Harbor Freight at < 40 PSI. I did it in my 30 x 50 ft workshop in early January. Turned off the oil furnace for a few days then shot them. I did use a chemical respirator but with good ventilation don't think it would be necessary. We will see by end of next week on this theory.
I basically sprayed the stuff to cover the white AL. It was quite easy I used 0.019 white roofing coils. There were some runs, but it is good enough. It was mostly done in 1 pass. I did not measure the thickness. Don't have equipment to do at home, maybe machinists at work could ...
I shot 6 of the 4x8 panels. That took 1/4 to 1/3 of the gallon. There was no visible fog and minimal overspray. Fumes were not overpowering. All 6 took < 50 minutes with 1/2 of that probably refilling the gun and making space. I adjusted the gun to give a even coat over 4-7 inches and held it at less than 1 ft.
I closed up shop and came back a day later. There was little if any smell left. The shop is quite tight and well insulated so I was surprised.
The paint was 69 + 22 shipping for a gallon.
|Note that the SolKote may not be selective when sprayed over a painted surface, so bear this in mind if you want to have a selective coating on the absorber -- best to use it over bare aluminum.|
Chad did a copper heat exchanger using 2 parallel runs of 3/4 inch copper pipe hooked up in parallel.
click on pictures for full size
Chad did a test run with this heat exchanger -- click the thumbnail below for a full size plot.
|1) The water heater inlet was connected to the shop water supply.|
|2) the tank was filled with 135 degree water from a nat gas instant water heater.|
|3) The heater was turned off, the heater outlet was connected to the tank heat exchanger inlet.|
|4) The water was turned on at 0.04 hours to 2.4 gpm.|
|5) At 0.11 hours the water was turned on full, 4 GPM.|
|6) At 0.16 hours the water was set to 2.4 GPM|
|7) At 0.21hours the water was set to 0.875 GPM|
|8) At 0.21hours the water was set to 0.875 GPM|
|9) At 0.39hours the water was set to 1.2 GPM|
|10) At 0.39hours the water was measured at 1.325 GPM|
|11) At 1.25 hours the water was set to 2.4 GPM|
|12) At 1.45 hours the instant heater was turned on at 140 deg|
|13) At 1.55 hours the test was stopped.|
|Note: The top sensor was above the copper for at least 1/3 of the test and therefore not reading the water at the top of the heat exchanger|
|The water had a significant amount of stratification!!|
|Flow was +/- depending on well pressure.|
|datalogged with an Omega multi channel commercial logger|
|All using type K thermocouples|
|Flow measured with a marked bucket and watch|
|The tank water compartment is|
|neglecting liner loss and including insulation.|
|107 gallon gross|
|Filled to 43 inches.|
|Heat exchanger is 2 parallel rectangular loops of 3/4" copper, about 15 x 17|
|18 loops total ( 9 per) spaced 0.375"|
Refer the PDF plumbing diagram for the system layout. This may change somewhat as the house side is completed. This is drawn using electrical schematic software.
Plumbing Diagram... (pdf)
See the Excel sheet for a detailed time log. I averaged about 9 hours per panel. I wasted a bit of time early remembering how use a table saw and figuring out than a second fence was very useful for running multiple boards through with the same cuts.
Detailed time long... (Excel Spreadsheet)
I attempted to make like parts for
all panels with only one setup process per part. This saved considerable setup
I used power tools as much as possible.
Air caulk gun
Circular Saw for cutting all OSB in one pass. Electric stapler
Grinding wheel to debur 5.8” copper pieces
Lift for keeping panels at an appropriate height while working
Power saw for cutting copper
Drill mounted copper debur / deoxidize brushes
Home Depot could not get the SunTuf or so they claimed. Lowes happily ordered a similar product for $21 per sheet.
I found the 0.019” thick aluminum in a 2 x 50’ coil. I cut this into 7.5” x 2’ strips using a shear. This took about 20 minutes per roll, or about 8 minutes per panel. This time is not in the log.
A Tankless water heater has replaced
a tank heater. This will peak the solar tank output. Unfortunately, it is 25’
from the rest of the system due to exhaust complications. Additional insulation
is needed on the copper pipes to reduce heat loss.
2 tanks were designed and built based off Gary’s design. The hot water tank is 102 gallons, the dimensions are 29x30x48”. The house storage tank is 292 gallons, the dimensions are 3.5’ x 8’ x28” tall.
The hot water tank was a compromise of fitting into an area between the furnace and a sump pump. I would not suggest the height to width ratio of the hot water tank for 3 reasons:
1) Waste of liner material. This tank used nearly the same amount of material as the big tank.
2) Difficulty in installing liner. The liner was formed around a box then inserting this into the tank. Otherwise the liner would not seat well.
3) Excess surface area leads to increased losses.
The house tank doubles as a dryer stand. The washer is also on the tank, but this rests on a steel frame than is bolted to the wall and has it’s own feet. The waster / dryer are front loaders. The added height elevates them to what is ideal for moving clothes between the two and removal from the dryer.
The house tank is heated by circulation glycol through about 120’ of ½” copper pipe. The hot water tank is heated by circulating its water through a flat plate heat exchanger in the glycol loop.
A 190’ trench was dug between the collectors and the house for the heat transfer tubes. It was put for the most part in the shortest distance between the two. PEX was run through the middle of 8 x 9.6” extruded styrene blocks. These were made from 2” sheet.
During the week of 11/20/09, the solar system was installed. A controller was programmed to control the entire heat collection system along with enabling heat extraction alone or in parallel with the gas furnace. The basic control elements for the system were programmed. Others will be added as time permits and desired.
Additional controller items desired:
data logging to a SD card
error/ problem detection
summer / winter profiles
efficiency based upon output and calculated sun position.
Two additional panels were installed
the last week of 2009 for a total of 240 sq ft of active collector area.
One added panel was the prototype panel with updates. The original vinyl glazing was dead after a year.
The framing was made per Gary’s design rather than the one I detailed. As I have concluded that conductivity across the aluminum is a limiting factor, I decided to experiment with this. I added secondary absorbers ( on top of original ones). The base aluminum panels were 3.75” wide prior to forming. These had the typical RTV bead down the center then were stapled on top of the original absorbers.
It is noteworthy to mention that many of the 1 yr old stainless staples could be pulled from the OSB with a fingernail. This panel stagnated most of the summer though the vinyl deformed at the top and had a big air leak.
The second added panel was fashioned after Tom’s: plywood, insulation, then absorber. The rest of the panel is per my 6 panel construction methods.
When sunny, the system starts
collection at 9-9:30 AM and makes useful heat for the house until 3-3:30 PM. It
sometimes makes some useful heat for the workshop after 3PM but this depends on
how much heat the house needs during the day. The system does not measure heat
added to the shop, this is done solely on collector temperature. The shop floor
is assumed to be 50 degrees.
The system seems to product 1/4 to ½ of sunny output for mildly overcast days.
The log (just below) was from a 30 degree day. Measured heat collection is about 155kbtu. This is to the house and is net of any losses from the collectors to the house.
Performance with below zero temperatures is at least 135kbtu / day. There have been very few sunny days since the switch was turned on.
Temperature losses due to the trench seem to be mainly due to night time cooling. Quick measurements of trench in and out temperatures do not show measurable losses.
Click on graph for full size
The delta between tank temperatures
and outgoing water is excessive. At 15 degrees, this is costing near 10 btu / sq
ft hr. I intend to make the flat plate heat exchanger also service the house
tank. I will make a tight coil of PEX or HDPE and drop this in the hot water
tank. The pump will then circulate the water in through rather than directly
from the tank. A second 10 watt pump will circulate water from the house tank
directly through the heat exchanger. This will drastically reduce the glycol to
Additional flow through the collectors would be beneficial. Presently, a rate of 2.6 GPM is calculated based upon delta T and tank temperature rise. A high head Grundfos UPS29-99 pump is used. The plumbing to the collectors contains many PEX to brass T’s and such that probably have considerable flow restrictions. These will be investigated as to the likely pressure losses. They may be replaced with ¾” copper T’s which would also remove 30’ of travel but require flow balancing the collectors. A second pump would get the flow to about 4 GPM at a cost of 170 additional watts.
While good intentioned, this seems more effort than it is worth. The system is designed to allow for closed loop drain back operation. The intent was to pneumatically pressurize the house expansion tank and depressurize the shop expansion tank. The drain back would add 5-10kbtu / day. This is quite low on the to do list. Time would be better spent adding insulation to the house.
Complete and install remaining 1 + 2
x ½ collectors for the final 60 sq ft.
Update controller program.
Investigate performance enhancement.
April/May 2009, and updated January 2010
Chad will answer email questions here: Clendenc AT execpc DOT com (replace AT with @, and DOT with a period)
Gary April 26, 2009, May 29, 2009