Literature

Joint Industry Foam Standards and Guidelines

SECTION 9.0

Published: 7/94

FLEX FATIGUE OR IN-USE SOFTENING STANDARDS AND GUIDELINES

9.1 Introduction:
The fatigue properties of flexible polyurethane foams have long been a major source of controversy. Questions relating to the causes of fatigue, the relationship of fatigue to other physical properties, the ramifications of measuring errors, and the correct test method have been debated for years with little quantitative results.

For example, ASTM D3574 lists the following acceptable test methods for fatigue softening:

a---static force loss at constant deflection
b---dynamic fatigue test by the roller shear at constant force
c---dynamic fatigue test by constant force pounding

A plethora of other, non-ASTM tests abound in the industry for fatigue measurements on foam:

a---Squirming Irma (several versions)
b---Bouncing Betty (several versions)
c---Rotary Shear
d---cushion on springs (simulating actual furniture construction)
e---The Fire-House test
f---The Rest-Home test
g---the college dorm test
h---The lobby test
i---The tractor test
j---Fatigue of CFD size specimens

9.2
The constant load, roller shear test--at this point in time--seems to correlate well with in-use testing and performance. In 1982 and 1986, two papers were presented to the Annual Technical/Marketing Conferences of the Society of the Plastics Industry:

a. Fatigue testing of flexible foams by Dr. Herman Stone, proceedings of the SPI 27th Annual Technical/Marketing Conference--1982

b. SPI study--Flexible Foam In-Use Fatigue Testing for Chairs by J.E. Knight, proceeding of the SPI 30th Annual Technical/Marketing Conference--1986

Note: Excerpts from the above papers are used by permission from both authors. The research work and conclusions published in these two papers is and will be the beginning core for quantitatively relating laboratory fatigue testing to actual in-use softening or fatigue.

9.3
Following are the observations and conclusions from the 1982 paper:

a. Density measurement agreements between laboratories were good though some laboratories reported significant deviations from the average value. This was important because of the correlation determined between foam density and dynamic fatigue properties.

b. IFD determination--there was poor agreement between laboratories in determining initial IFD values. The 4" IFD measurements show better agreement than the 2" samples.

c. IFD losses in fatigue testing occur very rapidly regardless of the fatigue protocol used.

d. The percent loss after 1000 cycles may be used as a rapid indicator of quality. The losses after 1000 cycles were lower on 4" samples than on 2" samples.

e. Fatigue tests on 2" samples were more severe than on 4" samples.

f. There is a general effect of density which becomes more pronounced at the low end of the density range (1.2 PCF or less).

g. HR foam performs well but not better than an unfilled conventional foam of equal density (one data set reported in the study).

h. Filled foam did not perform as well as HR foam or conventional foam. Its performance was close to its unfilled density.

9.4
Following are the observations and conclusions from the 1986 paper:

9.4.1 In the 1982 work, some suggestions were made for future research, and many of these suggestions were incorporated in the 1986 research work. The 1986 work contained some very interesting ground rules and commentary.

9.4.2
The correlation between laboratory dynamic fatigue and the field evaluations of foams was of primary interest. For this reason the number of foams chosen as the subject of the study was restricted. Four commercially available foams were selected for study.

a. 1.0 PCF--30 lb IFD conventional

b. 1.5 PCF--30 lb IFD conventional

c. 2.0 PCF--30 lb IFD conventional

d. 2.5 PCF--30 lb IFD HR

9.4.3
None of the foams selected for study contained any inorganic fillers or fire retardants.

9.4.4
In the 1986 paper, the following observations and conclusions were offered:

a. Density--results from participating laboratories show excellent agreement, much better than was reported in earlier studies. The uniformity of test samples may have accounted for the improvement.

b. The 90% compression sets generally showed good agreement between laboratories with only an occasional outlying result. As would be expected of the conventional foams tested, the 1.0 PCF was much higher in compression set.

c. Tensile--the standard deviation shown between laboratories was generally good for the tensile property. The variation for tensile between the highest and lowest reported results from a foam grade was about 20%.

d. Elongation--the variation shown between laboratories for the elongation was not as good as for tensile. The high and low variation was greater than 30% with greater dispersion.

e. Tear--the standard deviation for the tear property was good. The dispersion was better than for tensile and elongation.

f. IFD--the variation between IFD values reported was no better than what would be expected if ASTM procedures were followed. The 25% IFD showed about the same variation as the 65% IFD. The sample density did not appear to influence variation. The accuracy in IFD was better in this study than others and may be attributed to sample uniformity. The accuracy in the compression modulus data reflects the limitations of the IFD test method.

g. Roller shear dynamic fatigue-- after 1000 cycles the variation in IFD between laboratories was about the same as the initial IFD. The percent loss was significant. After 20,000 cycles the variation in all IFD's showed significant increase. The lower the foam density, the greater the percent loss (more fatigue). Allowing the foam to rest 24 hours after flexing showed the foam recovered some of the IFD loss. All test foam showed very little height loss. The percent loss at 40% and 65% IFD was not as great as the percent loss at 25% on all test foams. The variation between laboratories showed that the percent loss in all foam grades calculated at 25%, 40%, and 65% IFD was inordinately large. The reasons for the large variations are not clear. It may be a) humidity since most roller shear apparatus are not in humidity controlled atmospheres, b) due to inaccuracies of the IFD measurement or c) difference between the roller shear equipment between laboratories.

9.4.5
Observations on the in-use testing were: Cushion height used in the test chairs was selected to be 4". This was done to provide a more direct correlation between laboratory dynamic fatigue data and field evaluation data. While the use frequency or severity of use of the test chairs may vary from one test site to another, averaging the data still provides good correlation information.

9.5
The following observations are offered for the reported data. (From the research paper--1986)

a. One very surprising finding was the height loss observed in the 1.0 PCF density foam. In the other test densities, height loss was very low and close to that observed in the roller shear fatigue test. With the 1.0 PCF foam, height loss approached 10.0%.

b. The percent loss in IFD was very rapid at all indentations. Approximately 75-90% of the IFD loss was seen in the first 30 days.

c. As expected, the lower the foam density, the greater the loss in IFD. The percent loss in IFD is much greater for foam densities below 1.5 PCF.

d. The average percent loss after 180 days for a particular foam grade was very close to the average percent loss in IFD for that grade as determined by the roller shear procedure.

e. The variation in the percent losses were not as great as the variation observed between laboratories with the roller shear procedure.

9.6
Summary, conclusions, and commentary on the test results and observations from both research papers is as follows:

9.6.1
It is becoming more and more statistically evident that low density foams cannot be expected to perform as well as higher density foams from the standpoint of fatigue or loss in IFD with flexure from any source.

9.6.2
Thinner foams simply take a greater beating than thicker foams, and seats should be designed accordingly, or foams should be chosen accordingly.

9.6.3
More research should be performed on the reproducibility and accuracy of IFD measurement.

9.6.4
The roller shear test may adequately correlate with in-use fatigue losses.

9.6.5
More work yet is required on the roller shear fatigue test; as the results from lab-to-lab are not yet statistically dependable.

9.6.5.1
Because of their ultimate importance, excerpts of this research were reproduced here in their entirety (with only numbering changes to accommodate this publication.) This research is and will be the core information source for the ultimate solution to fatigue problems, and, thus, may soon allow the furniture industry to design seats and constructions without the total subjectivity which presently exists.

9.6.5.2
Many fatigue or softening complaints may be caused by the furniture manufacturers and their sales and marketing philosophy. Generally, the sales, marketing, and merchandising personnel perceive that upholstered furniture must have a soft, showroom feel in order to interest the buyers. In order to obtain that soft, showroom feel, IFD's are sometimes lowered. This lowering of the IFD may cause as much as 30-75% of the field softening complaints. First, a lower IFD deflects more under the same load than a higher IFD foam, and it is the amount of deflection which causes fatigue softening.

9.6.5.3
Another very important point to consider is that even well made, good quality flexible polyurethane seat-cushion grade foams will lose 4-6 pounds of the 25% IFD (4" THICK), with time in actual use. If the original IFD were 26 lbs @ 25% and the fatigued 25% IFD were 19-21 lbs, the 19 to 21 pound values would not be adequate to prevent an average weight person from feeling the springs and/or the under construction of the sofa. The customer complaint is likely to be "that cushion has become too soft." Complaints of this nature were most likely caused by incorrectly choosing the lower end of the acceptable IFD range. Conversely, in this case, had the lower end of the IFD range been selected to be 28 lbs at 25% deflection, the foam would fatigue soften to 22-23 lbs at 25% and would just barely (but acceptably) keep the average weight person from feeling the springs and under construction. Thus, manufacturers whose 4" 25% IFD specifications are 25-26 pounds minimum are walking on the edge of potential field complaints at all times--particularly in the summer when IFD's regress because of higher temperatures and humidity.

9.6.5.4
High density (3-3.5PCF) HR type foams which fatigue no more than 3 pounds in the same in-use period may be used to minimize field problems with fatigue.

9.6.5.5
With HR type foams and their high compression moduli, there is a temptation to use softer foams. Caution should be exercised in using softer or lower IFD foams because they do deflect more under the same load, and deflection is the cause of fatigue. Thus, while lower IFD foams are indeed softer to the feel, they may exhibit more fatigue because of their greater in-use deflection.

9.7
Edge set at the edges of cushions:

9.7.1
The mass of the foam in the very edges of perfectly rectangular (non-radiused or buffed) cushions and the degree of compression at the cushion edge leads to a very interesting type of fatigue failure called "cushion-edge-set" or just "edge set." When the foam core is stuffed in the cushion casing, the volume of the foam used is always substantially more than the volume of the cavity of the cushion casing or jacket. The foam is then, after stuffing into the cushion jacket, always under a constant, static load equal to the volume stuffing ratio differences.

9.7.2
Since the edges of the foam are in first contact with the cushion jacket, and since much of the static load or force is transmitted into the foam through the now-compressed edges, the edges see the brunt of the compressive, static load. As was mentioned earlier, the edges do not contain much mass and, thus, are not very resistant to compression; so the edges deflect a great deal under the compressive forces caused by simply stuffing the foam core in the cushion jacket. This compression on the edges is further compounded by the act of normal sitting on the cushion. Regardless of the beginning quality of foam or foam type used, some edge set will be found. Most of the edge set is usually not recoverable when the compressive forces are removed from the edges of the cushion.

9.7.3
Edge set can contribute to fabric movement on the foam core and thus, lead to cushion jacket rotation and/or wrinkles in the fabric which, in turn, can lead to excessive fabric wear.

9.7.4
To minimize edge-set, a radius can be cut or buffed on the edge of the cushions. Another way to minimize edge-set is to wrap the cushion with polyester fiber.

9.8
Other fatigue considerations: while not published, statistically planned laboratory results are available. It is becoming generally accepted that the hysteresis measurement can be a good indicator of fatigue for all types of flexible polyurethane foams. The theory, in this case, states that if the 25% return hysteresis loss is, for example 30%, the actual in-use fatigue will be very close to 30%. More testing and research is required, but obviously if hysteresis is in fact correlatable with fatigue, much expensive and time consuming testing will be obviated.

9.9
Criteria for fatigue losses:

9.9.1
Even though the research into the in-use fatigue losses of flexible polyurethane foams has produced some viable trend data, the state of the accuracy and reproducibility of the measurement of IFD is still inadequate enough that specific, quantitative criteria for IFD losses in fatigue and in-use testing cannot-- and should not--be proposed. Some guidelines, however, can be suggested for use until more reliable and quantitative data is available.

9.9.2
For 4" thick seating grade foams tested using any of the test methods in ASTM D3574, the fatigue loss in the 25% IFD should be no more than 6 lbs. The percentage loss values on typical 4" seating IFD's are found in chart 9.9.2.

Note: The use of two integers and a decimal fraction in some of the following data does not imply that the data is to be taken as accurate to three significant figures. There are presently no data available which factually indicates the number of significant figures in an IFD test or a percentage using IFD as the basis of the percentage.

25% IFD on 4" thick cushion %IFD loss if actual loss is no more than 6.0 Lbs
24
25
25
24
26
23.1
27
22.2
28
21.4
29
20.1
30
20
31
19
32
18.7
33
18.2
34
17.6
35
17.1
36
16.7