Down warmth metrics

**The goal:**To measure the heat loss from an object through different lofts, fill powers, densities, and overall fill weights of down in order to weigh the effectiveness of these different metrics. Yes, it's fairly crude, but I repeated the cycles many times to try and rule out inconsistencies and held constants as constant as I could.

**The test:**A cardboard box was lined with insulation. A half pint mason jar was filled with heated water and placed in the center of the box. Two panels were constructed similar to a down jacket and placed over the top opening of the box. The fill power rating, the loft depth, and the amount of fill in the panels was varied.

The cardboard box was lined with 2 layers of 5oz Climashield APEX insulation to ensure heat loss was convected up and not conducted down. This is an amount relatable to a -5f sleep system.

Two 12 x 16 panels built from Argon 67 nylon fabric were constructed with 3 sewn through baffles splitting the panel into chambers spaced 4" apart to mimic typical jacket construction. One panel was filled with 800 fill power down and the other with 950 fill power down. Throughout the test the amount, loft, and density of the fill was varied.

Fill power rating is a test that is performed on all bulk down that determines how much space one once of a given down will fill. A cylinder 241mm in diameter is filled with down and a 68.3g disc weight is placed on top. The height of the disc shows a height dimension and is used to find the cubic inch volume that the down is filling. 800 fill power means it filled 800ci. Fill power is determined by the down processing facility and is assumed to be accurate for this test. The method for determining the amount of fill in these panels was consistent with the method used for determining the amount of fill in down items such as jackets and sleeping bags. Length of the chamber X Width of chamber X Height of chamber / fill power rating. The length of each chamber was 12". The width of each chamber was 4". The height of a sewn through baffle starts at zero since it's a two dimensional piece. The height number plugged in here determines how much loft and density the chamber will have pushing the shells apart and it varied from 1.5" to 3". This is called "calculated loft" in this test. When multiplied together this gives the volume. When that volume is divided by the fill power rating it equals the amount of down weight needed to occupy that space. After the panels were filled "measured loft" was also taken. This is a measurement of the loft height between between baffle lines across the panel. It should be noted that in box baffle construction the height value is a measurements of the baffle height and the fill is calculated based first on how much is needed to fill the space, and then on how much "overstuff" is desired to achieve a level of density.

It was described above, but this distinction is very important to the topic. Measured loft is a measurement taken from a finished item that actually measures the depth of loft. Calculated loft can be interchangeable with "density" and it is the number, or numbers, plugged into the equation to determine the amount of fill put into each chamber. In sewn through construction, it is the "height" dimension in the volume calculation. Since sewn through construction starts out as a two dimensional object, this "height" figure represents the density pushed into the chamber, pushing the shells apart to make a three dimensional space for loft. In box baffle construction it starts out as a three dimensional "box". So the height dimension is already represented as the baffle height. That is plugged into the volume calculation to come up with only enough fill to fill the space. Extra density is added above that in the form of "overstuff". Both of these can be used to very accurately portray the amount of density stuffed into each chamber. Either as a number representing "calculated loft" in sewn through construction or as "loft, plus overstuff" in box baffle construction.

The panels were filled to 1.5" calculated loft, which translated to 1" of measured loft for both 800fp and 950fp, indicating that the manufacturer fill power ratings were accurate. The left is 800fp. The right 950fp. The 800fp panel took a total fill weight of 0.334oz. The 950fp panel took a total fill weight of 0.282oz.

**Measured loft vs Calculated loft:**It was described above, but this distinction is very important to the topic. Measured loft is a measurement taken from a finished item that actually measures the depth of loft. Calculated loft can be interchangeable with "density" and it is the number, or numbers, plugged into the equation to determine the amount of fill put into each chamber. In sewn through construction, it is the "height" dimension in the volume calculation. Since sewn through construction starts out as a two dimensional object, this "height" figure represents the density pushed into the chamber, pushing the shells apart to make a three dimensional space for loft. In box baffle construction it starts out as a three dimensional "box". So the height dimension is already represented as the baffle height. That is plugged into the volume calculation to come up with only enough fill to fill the space. Extra density is added above that in the form of "overstuff". Both of these can be used to very accurately portray the amount of density stuffed into each chamber. Either as a number representing "calculated loft" in sewn through construction or as "loft, plus overstuff" in box baffle construction.

**Round 1:**The panels were filled to 1.5" calculated loft, which translated to 1" of measured loft for both 800fp and 950fp, indicating that the manufacturer fill power ratings were accurate. The left is 800fp. The right 950fp. The 800fp panel took a total fill weight of 0.334oz. The 950fp panel took a total fill weight of 0.282oz.

Water was heated with an electric kettle that reliably heats to a consistent temperature. The water was transferred to the half pint mason jar and immediately set in the center of the insulated box. An initial temperature was taken with an infrared thermometer and a panel placed over the top of the box. The panel was partially lifted every ten minutes and a new temperature reading was taken. This process was repeated multiple times for each panel to try and rule out any possible inconsistencies and to note the overall average trends. The ambient temperature was kept consistent throughout. Each component of the test was returned to that temperature in between each test cycle.

The initial temperature reading started at 170f. Surprisingly, the 950fp panel consistently lost a degree or two more in the first 10 minutes than the 800fp panel. From that point on both panels mirrored each other with almost exactly 6f degrees of loss every ten minutes. After 30min the 800fp panel drops to an average 151f. The 950fp panel to an avergae 149f. From there they both continued to drop 6f every 10min. It would appear that if loft and fill density is constant, fill power had little effect on the R-value of a panel. If there was any effect seen it would point to lower fill power having a slight advantage. I would assume this could be due to the lower fill power down's increased density and that it showed up in the first check due to a larger temperature gradient. Also of note, this round showed two panels with identical lofts, very similar R-values, but much different overall fill weights. This indicates that a person shopping for a down item could not use total garment fill as a comparison of warmth between items with different fill power rating, but could use chamber loft and density.

In this round the calculated loft was doubled to 3", which brought the measured loft up to 1.25". This is a 100% increase in fill weight and a 25% increase in measured loft. The left is 800fp. The right 950fp. The 800fp panel took a total fill weight of 0.67oz. The 950fp panel took a total fill weight of 0.562oz.

**Round 1 results:**The initial temperature reading started at 170f. Surprisingly, the 950fp panel consistently lost a degree or two more in the first 10 minutes than the 800fp panel. From that point on both panels mirrored each other with almost exactly 6f degrees of loss every ten minutes. After 30min the 800fp panel drops to an average 151f. The 950fp panel to an avergae 149f. From there they both continued to drop 6f every 10min. It would appear that if loft and fill density is constant, fill power had little effect on the R-value of a panel. If there was any effect seen it would point to lower fill power having a slight advantage. I would assume this could be due to the lower fill power down's increased density and that it showed up in the first check due to a larger temperature gradient. Also of note, this round showed two panels with identical lofts, very similar R-values, but much different overall fill weights. This indicates that a person shopping for a down item could not use total garment fill as a comparison of warmth between items with different fill power rating, but could use chamber loft and density.

**Round 2:**In this round the calculated loft was doubled to 3", which brought the measured loft up to 1.25". This is a 100% increase in fill weight and a 25% increase in measured loft. The left is 800fp. The right 950fp. The 800fp panel took a total fill weight of 0.67oz. The 950fp panel took a total fill weight of 0.562oz.

**Round 2 results:**During this round the initial temperatures were 172f. The two fill powers were even closer matched, but where there was a difference, even though incredibly small, it went to the lower fill power again. The initial drop was less than round 1 with each losing 6-7f. From there the drops were slightly less at 5-6f. After 30min the two consistently dropped to an average of around 155.5f. After this they each dropped about a 5-6f every 10min. Again this round showed, if loft and fill density is a constant, then fill power has very little effect on R-value. Compared to round 1, fill weight and calculated loft was increased by 100%, measured loft was increased by 25%, and the heat lost was decreased by 17% over 30min. The improvement in R-value was much more consistent with the increase in measured loft than total fill weight. Again, with the 800fp panel taking 0.67oz of fill and the 950fp panel taking 0.562oz of fill and the R-value testing similar, we can rule out total fill weight as being an effective metric of warmth between items using different fill powers. Measured loft and calculated loft (density) prove to be very accurate metrics of warmth between items with different fill powers.

**Round 3:**This round did not involve any further testing. Just research and calculations dealing with the issue of how overall surface area effects how the overall fill weight is spread out over an item. After researching 6 different brands I found that given jacket dimensions can vary as much as 15% without changing size. For instance, the chest circumference in a medium in one brand could be 15% smaller in a medium of another brand. I found up to 10% differences in arm and waist measurements between same size/different brands. If a customer were to compare two jackets and one of them was smaller by 15% in the chest, 10% smaller in the arm length, and 10% smaller in the waist circumference, it would add up to a significant reduction in the overall surface area of that jacket. Meaning the stated fill weight of the larger jacket is spread over that much more space and any given section has less R-value than a jacket that spread a stated fill weight over much less space. I can't quantify how much less since these aren't all the dimensions of a jacket, but it is significant.

**Conclusions:**While measured loft alone is questionable as a metric of warmth since it doesn't factor in the density within that loft, total garment fill weight also proves to be mostly irrelevant due to so many variables. The idea behind stating fill weight is solid. We want to know the density of fill. The problem is that overall fill weight of an item doesn't provide an accurate picture of that density. One cannot compare jackets of different fill powers as described above. Two items with different fill powers can be the same warmth with drastically different fill weights. One cannot compare jackets with hoods to jackets with collars. A hood adds a varying amount to the overall fill weight without necessarily changing the loft/density/warmth in the jacket. One cannot even reliably compare items with the same fill power, hood/collar, or stated size because the final dimensions could be spreading that total fill weight over considerably different surface areas. One cannot compare items that might use different levels of fill throughout an item. For instance, if a jacket used a higher loft and density in the torso than in the arms. Knowing calculated loft (or loft, plus overstuff if it's a box baffle) will tell a consumer the density in each chamber of the item regardless of the fill power, size, dimension variation from manufacturer to manufacturer, and the presence or absence of features like hoods or collars or insulated pockets, etc. Measured loft seems like a useful partial metric to also have, since it does contribute greatly to overall warmth. Knowing measured loft along with calculated loft (density) can then give the consumer a clear picture of the density vs loft in any section of any item without the variables.

Whether I am using the best terms to portray chamber density is probably debatable, but my point is that, in lieu of a manikin clo test that would be prohibitive in a cottage industry, I propose that we offer some measurement of the density in each of the chambers of an item, along with the final loft height rather than just fill weight, which is a sum of many variables. We all know it already without any extra calculation. It's just a matter of customers asking for it. If nothing else, this is an exhaustive explanation for why I provide these metrics before total fill weight....and it was some mildly educational fun.