The pyranometer is mounted in a box which is aimed at the sun.
The glazing samples sit over the box.
The sides of the box are painted black to keep light reflections out of the pyranometer (which sees nearly 180 degrees).
Each glazing sample is placed over the pyranometer box, and the output of the pyranometer is measured with a millivolt meter. Between each sample, the pyranometer output for full sun is recorded -- this varied very little.
For the measurements of transmittance at 30 and 45 degrees incidence, the pyranometer and the box that surrounds the pyranometer were turned to the appropriate angle, and the pyranometer reading was taken with and with out the glazing in place. The transmittance I show is the ratio of these two readings. I was not able to find a good description of the "standard" way of measuring trnasmittance at non-normal incidence, so I don't know if this matches the usual technique or not.
For the twinwall and the corrugatted polycarbonate, the corrugations and the ribs in the twinwall refract the light in such a way as to produce a pattern of lighter and darker areas on the pyranometer. As you move the glazing sample left to right, the pyranometer readings vary. To get around this, I took 16 readings with the glazing moved a fixed small amount between each reading. So, for the corrugated glazing, the glazing was moved 0.2 inches between each reading, and the 16 readings covered 3 inches, which is the length of one full "wave" of the corrugated material. For the twinwall, the glazing was moved 0.1 inches between each reading, and the 16 readings cover 1.5 inches, which spans over about 4 of the twinwall ribs. Not a perfect method, but likely good enough.
Results
The sample glazings were:
Optix Acrylic 0.093 inches thick, new
Acrylic (brand unknown), 0.087 inches thick, new
Single strength window glass, 0.093 inches thick, new
Lexan poycarbonate, 0.093 inches thick, new
Suntuf Corrugated glazing, 0.03 inches thick, new
Suntuf Corrugated glazing, 0.03 inches thick, 9 years old
Suntuf Twinwall, 0.4 inches thick, new
Twinwall (brand unknown), 0.3 inches thick, old
Polycarbonate, 0.01 inches thick, new
Mylar film, 0.002 inches thick, new
The number 8 twinwall was very used and apparently retired from service. It came to me as packing material around an order of new twinwall -- so, to them it was scrap, and it looks like. It is discolored and has a lot of holes. I'd geuess it was from a greenhouse. Don't know how old it was.
The number 6 corrugated Suntuf is from my shop thermosyphon collector. It is still in good shape. If you hold it up next to a new peice of Suntuf, it looks slightly dimmer, but not much. It is still strong and flexible and crack resistant. I'm replacing the panel it comes from as part of another test -- not because it needs replacement.
I'd like to get some samples of typical window or sliding glass door glass just to see how much variation there is and how much things like low-e coatings matter. May talk to the local glass dealer about getting some samples from his scrap pile.
So, for the test, I went through all 8 samples twice. There was not much difference, so I'm only showing the first run. The full sun readings between each sample did not vary much (1047 w/sm to 1055 w/sm).
This plot shows the transmittance I measured for all of the new glazing samples.
The twinwall and corrugated glazings were only done for 90, 30 left, and 45 left because it was time consuming to do the 16 measurements associated with each angle.
One thing I've noticed is that generally the relationships between the different glazing materials are similar to the published values, but the transmittances I measured are consistently a bit lower than the published ones. I gather this is due to differences in the method used. The only material that is a bit puzzling is the Lexan sample, which seems low compared to the other polycarbonate samples and the generally accepted value for polycarbonate. I've tested this sample several times, and it consistently comes out as shown on the plot.
Sample
Brand
Material
Thick (inch)
Age
sun (mv)
Sun (w/sm)
Trans(mv)
Trans (w/sm)
Transmittance
1
Optix
Acrylic
0.091
new
13.2
1055.2
11.3
903.3
0.86
2
Acrylic
0.087
new
13.2
1055.2
11.3
903.3
0.86
3
Window glass
Glass
new
13.1
1047.2
11.4
911.3
0.87
4
Lexan
Polycarbonate
0.093
new
13.1
1047.2
10.5
839.3
0.80
5
Suntuf Corugated
Polycarbonate
0.03
new
13.1
1047.2
11.6
927.3
0.89
6
Suntuf Corugated
Polycarbonate
0.03
9 yr
13.1
1047.2
11.4
911.3
0.87
7
Suntuf Twinwall
Polycarbonate
0.4
new
13.2
1055.2
9.85
787.4
0.75
8
Twinwall
Polycarbonate
0.3
old
13.2
1055.2
9.95
795.4
0.75
Age -- the ones marked "new" were either just bought for the test, or they were sitting around my shop, but never used out in the sun.
Sun (mv) is the pyranometer reading in millivolts. This is multiplied by the pyranometer sensitivity of 12.51 micro volts per w/sm to get the sun value in w/sm.
The Trans (mv) column are the pyranometer readings in milli volts with a glazing sample in place.
The 9 yr old Suntuf sample was in service for 9 years on the collector on the south wall of my shop. It gets a lot of sun, and the winter sun intensity can be as high as 1200 w/sm with the reflection off the snow field in front of it that is there for 5 months of the year.
Items 1 through 4 (the flat sheet panels) were easy to get good readings on, and the readings were quite consistent. I think they are good.
The Suntuf corrugated and the twinwall (samples 5 through 8) are harder to get good readings on because the corrugations or ribs (twinwall) refract the light in such a way as to create patterns of light on the pyranometer. You can see them in these pictures:
Twinwall light patterns.
The pyranometer readings vary as you move the twinwall sheet from (say) right to left -- the readings range from about 9.2 mv up to 10.5 mv on new twinwall. The value I used in the table is the average of these two. This is clearly not exactly correct, but it can't really be off too far in that the min and max readings are not that far apart. It is probably at least good for comparing the new twinwall to the old twinwal.
Corrugated Suntuf light patterns.
The corrguations in the Suntuf sample also cause brighter and dimmer light patterns as in the picture above. The range in brightness is about the same as for the twinwall -- about 10%. For the table values, I used a sort eyeball average of the values I got as I move the glazing from right to left over one corrugation pattern. Again, this is not correct, but also not too far off as the range from min to max brightness is not very large.
The new Acrylic, Polycarbonate, and glass panes all have pretty close transmittances.
So, the idea that glass is much worse than the plastics does not seem to hold up for this glass sample. I bought this sample of glass from Lowes -- they sell various size panes for window repair. It is single weight, and I suspect that double weight would show lower transmittance due to the iron content in the glass.
It would be nice to get a variety of the sort of glass used in sliding doors that people like to use in collectors. I suspect that some of them are pretty bad. It would also be nice to know what kind of a hit to transmittance the various types of low-e coatings exact.
The transmittance for the Lexan sample seems on the low side -- wonder why? This was a brand new 8 by 10 inch piece of Lexan bought the same day as the test from Home Depot.
The 9 year old Suntuf corrugated glazing shows only a very small drop in transmittance compared to the new Suntuf -- just a couple percent drop. So, this is good news for people with Suntuf collectors.
One surprising result on the twinwall was that even though the old sample of twinwall that I had showed visible signs of degradation, the transmittance was the same as the new sample. These are two different thicknesses of twinwall and very probably different brands, and I don't really know the history of the old sample, but just looking at the two samples, I would have expected much lower transmittance on the old one.
