2. Question asked by CPSC
• Should CPSC change the regulatory approach
from a smolder test on bench-scale models to:
– An approach that limits fire growth by
performance requirements similar to 16 CFR
1633?, or
– An approach that requires the use of fire barriers?
(for all fabrics, not just Class II fabrics).
3. Why is CPSC considering changing approach
from 16 CFR 1634 (2008 smolder standard)?
• Problems with scaling??
• Difficult to understand this point:
– Flaming tests are known not to scale well from
small-scale to large-scale, but cigarette tests do.
– California has consistently found that mockups do
predict real-chair behavior.
“Bench scale performance did not predict full scale
flammability…especially for smoldering ignition”
4. Smolder testing…
• CPSC considers that any regulatory test that
they adopt needs to accurately control the
real-article hazard and classify the products in
the same way as they behave in full scale.
• For tested cigarette mockups, CPSC concluded
that the mockup tests did not always predict
that full-scale chairs will fail by smoldering
when cigarettes are placed at certain locations
on the chair.
5. …Smolder testing…
• They identified these problems:
– Two different “standard” foams behaved very
differently.
– Mockups did not have doubled-over fabric layers
(which promote smolder), while some real
furniture did.
– Full-scale furniture itself showed significant
differences between supposedly identical
samples.
. . . .
6. …Smolder testing
– Fabric tension may not be well simulated.
– Side arms are generally constructed very
differently from back cushions.
– Fire barriers, when tested in mockups using
“standard” fabric and “standard” foam do not
reproduce the behavior of some actual chairs.
7. Three possibilities exist
• Require full-scale testing of chairs.
– Full-scale testing does not seem practicable, since
there is an enormous number of fabrics used by
most manufacturers.
• Stop all standard development
• Fix mockup problems…
8. Improving mockup testing…
• Specifications for “standard” foam have not been
adequate.
– It is unreasonable to suppose that
density, ILD, permeability, etc. would suffice to
describe the smolder propensity.
– It should be possible to define specifications for a
“standard” foam, but the specifications must include a
cigarette test.
– Smoldering is a chemical phenomenon, and physical
tests do not measure chemical properties.
9. …Improving mockup testing…
• Lack of stringency concerning doubled-up
fabrics is simple to fix.
– Make mockups use doubled-up fabric.
• Specimen-to-specimen variations in full-scale
chair construction.
– Have to accept this. Upholstered furniture is hand-
built and not made on an assembly line.
10. …Improving mockup testing
• Wide differences in construction of real
chairs, depending on the exact locale.
– Need to select a near-worst-case mockup
configuration. Then most actual variations will
tend to give conservative differences, while the
few opposite results will be rare.
• Some smolder barriers behave too well in
mockup tests, compared to real furniture.
– May need to revisit specifications for “standard”
fabric and foam.
11. Recommendations
• Needed improvements in mockup testing for
smolder resistance should not be difficult.
– With such modifications, a smolder standard
should prove to be reliable.
• I do NOT believe that flame-resistance testing
for barriers is economically justifiable.
• I do NOT believe that HRR testing of full-scale
furniture is economically justifiable.
12. Flame tests or HRR tests are not justified…
• CPSC reported that there are about 30
“addressable” fire deaths per year from small-
flame ignitions.
• Added costs would have to be very small to
create a positive benefit/cost situation.
• Economic and engineering analyses on large-
flame ignitions are non-existent.
• Added costs of producing TB133-type furniture
would be large.
• Comfort and choice would be reduced if TB133-
type furniture were mandated.
13. …Flame tests or HRR tests are not justified
• Testing to a peak-HRR type requirement
would involve high costs.
• A manufacturer may offer thousands of fabric
variations.
• Peak HRR is dependent on fabric type
– Babrauskas (1983) showed a factor-of-4 effect!
• Could not adopt the 16 CFR 1633 strategy of
ignoring fabric variations.