Down Fill Power, Testing, Baffle Volume

[R]

First, a stupid question:

Wikipedia has a description of the US and European down fill power test protocols here: https://en.wikipedia.org/wiki/Fill_power

In a nutshell, one ounce of down is placed in a cylinder and a weighted piston is applied; after one minute the volume is measured.  What I haven’t seen stated explicitly is whether the fill power is the compressed volume measured, or if there’s some additional calculation or conversion involved?

That out of the way…

In several videos of the test procedure, the sample volume is reduced by roughly half during the test.  This implies that using exactly the calculated mass of fill for a given quilt – zero over fill – will actually provide around 100% over fill.  The pressure of the shell fabric (~0.6osy) is only ~1/60 that of the compressive piston, so that shouldn’t make much of a difference.  But baffles with zero over fill generally don’t look any where close to 100% over fill – they hardly bulge at all.

I did some number crunching around baffle dimensions, bulging of the upper and lower shell layers and maximum volume.  My calculations say that a 5″ wide baffle w/ a wall height of 1.8″ will bulge in height by about 2.6″ (that’s counting both top and bottom) and shrink in width by an inch.  This will increase the volume by 63%.  It will look like an air mattress at that point – because air mattress inflate to maximum volume with similar geometry.

Quilts and bags with continuous baffles do not appear inflated like an air mattress.  The baffle tops are fairly flat.  And that’s even with claimed 30% over fill.

So what’s going on with all this?  Down should loft to nearly twice it’s rated volume, baffles with zero over fill should balloon to double their wall height, and quilts that claim 30% over fill don’t look anywhere near their maximum baffle volume.  Sure, I could have gotten my math wrong, but you can’t argue the shape of an air mattress, and just watching a fill power test will have you wondering.

I guess what I’m really looking for is the reason why baffles never inflate to their maximum volume, and where that leaves the calculation for the ‘full’ volume of down for a given baffle dimensions.

[James Marco]

“Fill Power” IS the calculation for cubic inches of volume. Example: 1 ounce of 850 fill down will fill 850 cubic inches.

That said, it is an “ideal” measurement. Freshly washed and dried down is used. They use several tricks to increase the testing numbers.
Air injection is used to insure maximum separation of each down cluster and fibers within the cluster. During the test, this is removed, hence is “legal.”

I suspect that down is washed to degradation, stiffening the fibers somewhat. Again, there is no residue on the down, hence is “legal.”

Static electricity, though rarely mentioned, will also impart a degree of structural rigidity and support by expanding fibers.

Basically, for economic reasons, they are “gaming” the test system. So, for a quilt, I would normally buy some overfill to compensate for:
1) dirt, dust, bodily oils, all have the effect of reducing down’s fill power.
2) static electricity, it will normally ground out in use.
3) Random overlap of fibers and clusters.
OR
Simply buy a quilt/bag about 10F warmer than your intended target temp.

Now, we are talking theoretical vs “in-use” measurements. I tend to keep a very clean bag by washing with water every trip out (rinsing it out) then drying it. I may wash it with down wash every 14 nights of use, but only use a couple tablespoons to a quarter cup. I have a washer with no agitator, so no harm is done. When it is thoroughly dried in our dryer, it will spark, snap and crackle a high static charge. Measuring the loft ALWAYS exceeds the “new” spec, and all baffles are actually bulging, and, full. This is how the bag was “designed” to look. But in the field, it rarely looks this way. I have purchased five bags and a quilt. All perform pretty much as expected, and, all perform this way at home. The only time they do indeed look all puffed up is cold and no humidity (~25F/0% humidity.) In damp, warm conditions, it is limp, and fairly flat with the baffles all looking very anemic…

Loft, is only one measurement for insulating value. Indeed, if it was the only important measurement, we would simply use inflatable bags/quilts with no fill. But air only has about a R1/inch. This is far less than pad or down filling. Soo, what else is going on?
We also use this notion called air entrapment or borrowing from the concrete industry, entrainment. By breaking up the air space smaller and smaller we reduce convection. By using thermally efficient materials we further increase this. So, it is not hard to see that an R19 in 3″ of down is possible. This is a far outline of the discussion: https://backpackinglight.com/forums/topic/6520/

Anyway, I hope this will help answer your questions. There is no one set formula. The easiest to use is fill power and convert directly to cubic inches. But, loft is only part of the numbers and should only be used to compare other bags made from the same materials (down or synthetic.) It was quoted somewhere in the reading that a bag that lost around 27% of it’s loft still retained about 93% of it’s warmth. So, loft, fill, and insulating value are all somewhat different in meaning and have only approximate relationships.

[Jerry Adams]

good question, good answer by James

it’s more than just the weight of the fabric directly over the down.  The weight of the entire sleeping bag pulls down on the fabric.  And if you’re tossing and turning you can put more pressure on it.  Or wind pressure.

I sort of disagree with the gaming the system opinion.  It’s a standard so you can compare different down and calculate how much down to put in each baffle.  As long as its consistent, then it has value as a standard.

The U.S. method produces a higher fill power number than European.  I assume the European standard came first?  That’s sort of gaming the system if U.S. manufacturers sell identical gear and claim a higher fill power.

And the fact that 25% overfill is good means they’re sort of gaming the system.  They should use a heavier weight in their test, or just multiply the final result by 25%.  But now that this is a standard they should just leave as is and we can multiply by 25%.

https://www.scribd.com/document/156159764/Mammut-Sleep-Well-Pt1-E has some good info

[Jerry Adams]

or http://4sport.ua/_upl/2/1410/Mammut_Sleep_well_pt1_E.pdf

weird – not on mammut web site anymore, and they don’t appear to sell sleeping bags anymore

[Greg Mihalik]

“I guess what I’m really looking for is the reason why baffles never inflate to their maximum volume,”

Part 1:

Verbage – in a quilt/bag let’s call the thing containing some of the down a “tube” and the fabric between the tubes the “baffle”.

The baffles limit the travel of the fabric forming the top of the tube.  If I gently fluff and float a flat quilt I will see the top surface “bulging” between the baffles.

And that is only half of the “bulge. The other half would be on the bottom, except it is squashed by the weight of the overlying down and fabric –

If I tension the quilt in any way, the bulges will disappear.

“Constraining tubes” are used instead of “maximun volume” tubes is to prevent down shift/migration during real-world use. [Edit, Redacted: Mr. Nisley pointed out that the R-value of down is a function of areal density, and not necessarily “Loft”. (Within common sense limits, of course.)]

 

 

[Greg Mihalik]

“I guess what I’m really looking for is the reason why baffles never inflate to their maximum volume,”

Part 2:

I look at “fill power” the same way I look at US EPA mileage estimates, And assume there are a few “Volkswagon” engineers in the mix.

I do make relative comparisons between 2  numbers provided they come from the same protocol.

[R]

Re. James’ initial comment:

I think my ‘stupid’ question may not have been completely clear.  I understand the definition of fill power and its intended use in designing a quilt/bag.  What I don’t follow is whether one ounce of 800 power down occupies 800 cu. in. before or after the compression piston is applied in the test. Presumably, it’s after, but this means it will expand significantly beyond rating in use.

Here’s a test: https://www.youtube.com/watch?v=n49aNBncpqU

At 0:50, you can read the top of the scale printed on the cylinder to be 750.  Starting at 5:40 you can see the tester dumping a small bag of down in to the cylinder which expands significantly.  It appears to occupy not quite half the cylinder – say 40%; that would be 300 on the scale if we could read it.  By 7:50 air lofting has ended, and it might have increased to 50% of the cylinder; 375 on the scale.  At 9:50 compression has completed and you can read the scale at 155 – half its initial volume.

Is it then suggested that the magic treatment down test samples receive prior to testing doubles their volume?  I have a hard time believing that.

Jerry – thanks for the link.  I saw that several years ago, probably worth another look.

Greg:

To be clear, by ‘maximum volume’ I’m referring to the maximum volume the total assembly of constrained tubes could contain.  Sure the lower bulges compress, just as a tire does against the road, but that just means there’s more down available to inflate the upper side.  And yes, I’m talking ideal condition, laying on a flat surface, not in use around a sleeper.  But again, no catalog photo, or yours, resembles an inflated air mattress (even just on top) – maximum volume for the shell.

These bulges are steeper at the sides, and relatively taller in the middle, than any sleeping bag I’ve ever seen.

Obviously, all the math and physics of business as usual works out.  I just wish I understood why…  But not bad enough to do all the careful work it’d take to convince me :)

[John Vance]

The fill is measured with the 1oz top in place in the cylindrical tube.  Fill is stated in the semi compressed condition.

[Greg Mihalik]

“These bulges are steeper at the sides, and relatively taller in the middle, than any sleeping bag I’ve ever seen.”

The lofting power of down will never equal the pressure from lungs (~2 psi) or a pump.

[Jerry Adams]

“I think my ‘stupid’ question may not have been completely clear.  I understand the definition of fill power and its intended use in designing a quilt/bag.  What I don’t follow is whether one ounce of 800 power down occupies 800 cu. in. before or after the compression piston is applied in the test…”

If you calculate the volume of the baffles with no bulging, multiply by 1.25, divide by fill power (European), that’s how many ounces of down you need.

Other questions are just geekiness, which is a good thing : )

The 25% overfill accounts for all that other stuff

[James Marco]

Rene,
” What I don’t follow is whether one ounce of 800 power down occupies 800 cu. in. before or after the compression piston is applied in the test. ”
Well, it is obvious you use the 800 or semi-compressed number, as was said. This is the “standard” number. Note that there are rather loose standards for this. In one video they used an air injector to stir the down, randomizing any clumps. The second video, they used stirring rods to randomize the clumps. One will add static charges also, not mentioned. Also, they don’t mention the level of cleaning/washing. Fill power numbers can vary by 50 or more points easily. For MYOG stuff, deduct 50 in all calculations for volume. You really don’t care about the numbers, you want WARMTH. Don’t forget, the weight of the fabric tubes/baffles will act as roughly the same as the weight applied during testing.

Humidity can collapse down as much as one third. But warmth will only be about 15% effected. This is useful to know to calculate the high temp number a bag/quilt can be used comfortably at. Higher fill downs are more susceptible to humidity than lower fill downs.

Maximum fill numbers are pretty useless. With enough weight/vacuum pressure, you can squeeze down to a very small size. Close to 100:1 over a fully fluffed volume. I normally carry my bag/quilt compressed to a small volume (~20:1) in a dry bag with no problems.

[R]

Has any one ever measured down purchased from a reputable source?  What were your findings?

[Jerry Adams]

Maybe the relevant measurement is that when I overfill by 25%, 850 down from Wildernesslogics, the baffles stay pretty much filled.  On a quilt, after several days of use, the down can shift down to the sides leaving spots on top with little down, so I have to fluff it up a little, maybe each night when I set up camp, or even just once at the beginning of a trip.  I wouldn’t want to do less than 25%.

 

[James Marco]

Sorry for the delay, been busy…
Short answer, no. It is certainly possible to make a tube with cap and weighted disks. Then calculate the gradient lines to measure the volume. Air injection could be fun. I never bothered to do it though. The precision required for the weights, tolerances for the disk, and the ability to recover the down afterwards plus the space needed for storage for a more or less two or three times in a lifetime type tool always led me away. Like I say, it is low tech, but not easy.

[Ivo Vanmontfort]

Did a test,..long ago

https://backpackinglight.com/forums/topic/59920/

[Mark Fowler]

Having run a small factory producing sleeping bags many years ago I have done many fill power tests – I wouldn’t bother to do it provided I purchased the down from a reliable supplier. I would suggest Jerry is on the money with 25% overfill – my experience is that you can go a little lower – say 20% but 25% is fine – it depends on just how accurate you are in measuring your tube volume.  After filling, loosely tack closed the bag, shake it out to loft it up and you can then decide whether more down is needed and add it quite easily. Once happy finish it properly.

[R]

Rats – just lost a bunch of typing and number crunching…

Well, I took a look at the 900fp sample and converted to the US test spec.  The compression load was about 9% above spec., and the cylinder diameter was 30% below spec.  Correcting for these, converting to cubic inches and scaling to a full one ounce sample, I found the uncompressed volume to be 33% over spec, but the compressed volume 39% under spec – just 550. As with the video I linked previously, the difference between uncompressed and compressed volumes is still about 2:1.

I don’t think the compressive behavior is linear, but there can’t be too much error in treating it as such over just 9% delta.  I think the narrower tube favors a higher score as there’s more friction against the tube wall resisting compression – but I don’t think that accounts for much.  This implies the magic fluff process applied to a commercial test samples accounts for the bulk of that 39% deviation.  Quite a bit.

As a reference, I calculated that the compressive load of a bag’s upper shell plus vertical baffles amounts to 37% of the compression applied in the test.

One interpretation of this says that our usual 25% – 30% over fill makes up for the majority of the magic fluff, and the remaining loft volume (bulging shell) results from the expansion allowed by the reduced compressive load.

Kinda hand-wavy, but still informative.  It’d be interesting to see another sample.  I’d also like to see incremental compression data. And it might be interesting to see how much bulge you get with zero over fill.  We’ll see how interested I am by the time my order shows up :)

Thanks for the test, Ivo!

 

[Aaron W]

Maybe this will help answer your question.

Im assuming your baffle is a normal baffle?
You have aomething like a 1 1/2 inch baffle, want 2-2 1/2 inches of loft, right?

Well, unless your baffles are an inverse “U” shape, (meaning the top fabric width between each baffle is significantly wider than the bottom distance), then every tug or pull on any part of that bag is going to result in bringing the thickness of the bag down to the 1 1/2 inch baffle hieght.

The forcing up of down in the center of the baffle is fighting a lot more than just the weight of the fabric.

[Jerry Adams]

what do you mean ” The compression load was about 9% above spec., and the cylinder diameter was 30% below spec.”

Maybe results aren’t linear with load or diameter

Nice testing

[Jerry Adams]

another thing about the test is to “condition” the down – dry it out and fluff it in a particular way, possibly unrealistic for real world use.

how did you condition the down?

[James Marco]

Yeah, like I said, low tech, but not easy. Even a purchased standard amount will vary a lot with a home measurement. I would simply use the tables at EE or another reputable manufacturer and add some percentage of overfill to suit yourself.
For example:
EE, Nunatak, Feathered Friends, Western Mountaineering, etc all have between 13 and 15oz of 850fp down in a 20F bag. Not a huge difference. Only features, really make the difference.

Personally, I would purchase a pound and divide it evenly if you want a 20F quilt. It will likely take you lower, but it *will* work at 20F. Just plan on being surprised when you actually receive the 1 pound bag. It might be vacuum sealed and quite small.

[R]

My post was analysis of Ivo’s test.  Sorry if that wasn’t clear.  The spec I was comparing against is the US fill power test, found here: https://en.wikipedia.org/wiki/Fill_power#Measurement

My assumption is that compression is non-linear.  Having only one data point, I have no option but approximate to linear over the 9% difference.

From Ivo’s post, his conditioning of the down sample was just to hold the tube sideways and shake.  The whole reason I asked if any one had conducted a test at home was to see how a sample measured outside of a lab, where we can’t do all the crazy conditioning they do –  that’s what I meant by ‘magic fluff’.

[Jerry Adams]

ahhh…  thanks

[James Marco]

“I don’t think the compressive behavior is linear, but there can’t be too much error in treating it as such over just 9% delta. I think the narrower tube favors a higher score as there’s more friction against the tube wall resisting compression – but I don’t think that accounts for much. This implies the magic fluff process applied to a commercial test samples accounts for the bulk of that 39% deviation. Quite a bit.”

Compression would NOT be linear. From an ideal of say 1000fp, you start compressing it. It requires only the weight of itself, or a perforated disk or screen to start this. As you increase pressure, each fiber will bend more and more, exerting its resistance to the weight. To compress my quilt, I typically use about 120lb of pressure to achieve a 20:1 compression, assuming a 30lb pull on all four straps.

This is with a 20f EE Revalation, so some is also the fabric and not the raw down, but for general purposes, we can consider fabric to be a theoretical plane of infinite density. Only the folds and creases add to the compressive forces, but these are few, much less than 1%, considering how many down plumes and fibers are involved in resisting compression. But the fabric would also resist compression more.

Using a 4 strap compression bag, to achieve a 20:1 it is around 110-120 pounds total pressure or around 30lb on each strap. To achieve a 10:1 compression, only about 40-50 pounds of force is needed or around 10 pounds per strap. To achieve a 5:1 compression, only about a 5-6 pound force is needed. (I assume a 100:1 compression would be a maximum since this is roughly destruction of a down plume.) So by plotting just the three reference points, we see:

This appears to be exponential or Gaussian based. For each compression point, you need more force to achieve it than the one before. More and more down plumes become involved with pushing back. And, the ones already pushing back push back stronger. This is only based on three measurements using the Mark I eyeball, but it is also reasonable and conforms with standard physical laws. After many years using down bags, this behavior only changes in the nature of the exact curve, not the theory, sooo…take it with a grain of salt if you like, or ignore it. Especially at very small compression pressures it will respond quite linearly. With larger amounts and larger pressure, linearity breaks down. Myself, I believe it is actually exponential, but this is only a guess…

[Jerry Adams]

nice analysis on three data points : )

good enough for anything practical one would want to do

[Author]

Posts