6000 hours.

Need help.

Two test laboratories have interpreted IEC requirements/ definitions
differently.

The first lab declares a sample as "not complying" because 5th out of
ten lamps failed at 5800 hours. (Test was discontinued after 5th
failure)

The second lab declares another sample as "complying" because 50%
lamps failed before 6000 hours against the requirement of "Maximum 50%
failures in 6000 hours."

The definition of Average Life is: "The length of time during which
50% of the lamps reach the end of their individual lives."

This has been interpreted by the first lab as the highest life of the
first 50% failures.

The second lab argues that, as per the definition, if 50% lamps reach
the end of their lives after 6000 hours, then the average life is

6000 hours.

This means that one of the reports (I feel, the first one) is
incorrect. Hope I am not biased.

I have both the reports. But the first lab report is available with a
regulatory body who had initiated the life test.

I have given the second lab report to the regulatory body to convince
them that the sample COMPLIES. While the regulatory body is yet to
conclude, I request expert opinion on the issue from the lighting
community.

Regards.

Pankaj

Need help.

Two test laboratories have interpreted IEC requirements/ definitions
differently.

The first lab declares a sample as "not complying" because 5th out of
ten lamps failed at 5800 hours. (Test was discontinued after 5th
failure)

The second lab declares another sample as "complying" because 50%
lamps failed before 6000 hours against the requirement of "Maximum 50%
failures in 6000 hours."

The definition of Average Life is: "The length of time during which
50% of the lamps reach the end of their individual lives."

This has been interpreted by the first lab as the highest life of the
first 50% failures.

The second lab argues that, as per the definition, if 50% lamps reach
the end of their lives after 6000 hours, then the average life is
6000 hours.

This means that one of the reports (I feel, the first one) is
incorrect. Hope I am not biased.

I have both the reports. But the first lab report is available with a
regulatory body who had initiated the life test.

I have given the second lab report to the regulatory body to convince
them that the sample COMPLIES. While the regulatory body is yet to
conclude, I request expert opinion on the issue from the lighting
community.

Regards.

Pankaj

Did the lamp sets run at both labs have the 5th lap fail at
5800 hours? That would be rather surprising.

I would have to read the full IEC spec to determine which
interpretation is correct

We can clearly see that your disappointment in the actual results has
clouded your thinking.

Does the phrase "reach the end of their individual lives" mean "reach
6000 hours (rated average life)"? Well, even if the definition were this
imprecise and vague, your sample lot did not achieve this.

Did the lamp sets run at both labs have the 5th lap fail at
5800 hours? That would be rather surprising.

Thanks for your interest.
The test report from the second lab does not mention actual hours of
individual lamps which failed. It only says "50% (5 nos.) lamps failed
before 6000 hours..."
The debate among the labs is whether it is the first five lamps or the
last five lamps that will decide the average life.
In the definition ("The length of time during which 50% of the lamps
reach the end of their individual lives."), does the phrase "reach the
end of their individual lives" mean "reach 6000 hours (rated average
life)"? Or should we read the definition as "The length of time during
which FIRST 50% of the lamps fail."

I would have to read the full IEC spec to determine which
interpretation is correct

Awaiting your comments.

Regards.

Pankaj

Did the lamp sets run at both labs have the 5th lap fail at
5800 hours? That would be rather surprising.

Thanks for your interest.
The test report from the second lab does not mention actual hours of
individual lamps which failed. It only says "50% (5 nos.) lamps failed
before 6000 hours..."
The debate among the labs is whether it is the first five lamps or the
last five lamps that will decide the average life.

In the definition ("The length of time during which 50% of the lamps
reach the end of their individual lives."), does the phrase "reach the
end of their individual lives" mean "reach 6000 hours (rated average
life)"?

Or should we read the definition as "The length of time during
which FIRST 50% of the lamps fail."

We can clearly see that your disappointment in the actual results has
clouded your thinking.

Before "We" encompasses every member of this group, I would like to
clarify that the fight is between the two labs for opposite
interpretation. The regulatory body has already given me benefit of
doubt considering the ambiguity in the definition. So I don't have to
be disappointed.

Does the phrase "reach the end of their individual lives" mean "reach
6000 hours (rated average life)"? Well, even if the definition were this
imprecise and vague, your sample lot did not achieve this.

The sample did not achieve this because 50% lamps (surviving 5 lamps)
were not tested for 6000 hours. This makes the test inconclusive.

Need help.

Two test laboratories have interpreted IEC requirements/ definitions
differently.

The first lab declares a sample as "not complying" because 5th out of
ten lamps failed at 5800 hours. (Test was discontinued after 5th
failure)

The second lab declares another sample as "complying" because 50%
lamps failed before 6000 hours against the requirement of "Maximum 50%
failures in 6000 hours."

The definition of Average Life is: "The length of time during which
50% of the lamps reach the end of their individual lives."

This has been interpreted by the first lab as the highest life of the
first 50% failures.

The second lab argues that, as per the definition, if 50% lamps reach
the end of their lives after 6000 hours, then the average life is

6000 hours.

This means that one of the reports (I feel, the first one) is
incorrect. Hope I am not biased.

I have both the reports. But the first lab report is available with a
regulatory body who had initiated the life test.

I have given the second lab report to the regulatory body to convince
them that the sample COMPLIES. While the regulatory body is yet to
conclude, I request expert opinion on the issue from the lighting
community.

Regards.

Pankaj

We can clearly see that your disappointment in the actual results has
clouded your thinking.

Before "We" encompasses every member of this group, I would like to
clarify that the fight is between the two labs for opposite
interpretation. The regulatory body has already given me benefit of
doubt considering the ambiguity in the definition. So I don't have to
be disappointed.

Does the phrase "reach the end of their individual lives" mean "reach
6000 hours (rated average life)"? Well, even if the definition were this
imprecise and vague, your sample lot did not achieve this.

The sample did not achieve this because 50% lamps (surviving 5 lamps)
were not tested for 6000 hours. This makes the test inconclusive.
Since you have an even number of values, the life should be rated as the

On Feb 16, 6:40 am, Pankaj <pankaj...@yahoo.com> wrote:
Need help.

Two test laboratories have interpreted IEC requirements/ definitions
differently.

The first lab declares a sample as "not complying" because 5th out of
ten lamps failed at 5800 hours. (Test was discontinued after 5th
failure)

The second lab declares another sample as "complying" because 50%
lamps failed before 6000 hours against the requirement of "Maximum 50%
failures in 6000 hours."

The definition of Average Life is: "The length of time during which
50% of the lamps reach the end of their individual lives."

This has been interpreted by the first lab as the highest life of the
first 50% failures.

The second lab argues that, as per the definition, if 50% lamps reach
the end of their lives after 6000 hours, then the average life is

6000 hours.

This means that one of the reports (I feel, the first one) is
incorrect. Hope I am not biased.

I have both the reports. But the first lab report is available with a
regulatory body who had initiated the life test.

I have given the second lab report to the regulatory body to convince
them that the sample COMPLIES. While the regulatory body is yet to
conclude, I request expert opinion on the issue from the lighting
community.

Regards.

Pankaj

This is not an average, it is a median. Average has a strict
mathematical definition, and would require measure ALL lamp failure
times to determine what it is.

Clearly the intent is to use median life here. Back when I was taught
about this in math classes, the median for 10 lamps would be the
average of the 5th and 6th lamp. This is essentially the best
estimate of what a larger population would have as a median.

So, you would need to measure up to the 6th lamp failure. Without
that, you cannot calculate the median.

Regards,

Mark

On Feb 16, 6:40 am, Pankaj <pankaj...@yahoo.com> wrote:
Need help.

Two test laboratories have interpreted IEC requirements/ definitions
differently.

The first lab declares a sample as "not complying" because 5th out of
ten lamps failed at 5800 hours. (Test was discontinued after 5th
failure)

The second lab declares another sample as "complying" because 50%
lamps failed before 6000 hours against the requirement of "Maximum 50%
failures in 6000 hours."

The definition of Average Life is: "The length of time during which
50% of the lamps reach the end of their individual lives."

This has been interpreted by the first lab as the highest life of the
first 50% failures.

The second lab argues that, as per the definition, if 50% lamps reach
the end of their lives after 6000 hours, then the average life is

6000 hours.

This means that one of the reports (I feel, the first one) is
incorrect. Hope I am not biased.

I have both the reports. But the first lab report is available with a
regulatory body who had initiated the life test.

I have given the second lab report to the regulatory body to convince
them that the sample COMPLIES. While the regulatory body is yet to
conclude, I request expert opinion on the issue from the lighting
community.

Regards.

Pankaj

This is not an average, it is a median. Average has a strict
mathematical definition, and would require measure ALL lamp failure
times to determine what it is.

Clearly the intent is to use median life here. Back when I was taught
about this in math classes, the median for 10 lamps would be the
average of the 5th and 6th lamp. This is essentially the best
estimate of what a larger population would have as a median.

So, you would need to measure up to the 6th lamp failure. Without
that, you cannot calculate the median.

Regards,

Mark

On Tue, 19 Feb 2008 16:41:01 -0800 (PST), redbelly



redbell...@yahoo.com> wrote:
On Feb 16, 6:40 am, Pankaj <pankaj...@yahoo.com> wrote:
Need help.

Two test laboratories have interpreted IEC requirements/ definitions
differently.

The first lab declares a sample as "not complying" because 5th out of
ten lamps failed at 5800 hours. (Test was discontinued after 5th
failure)

The second lab declares another sample as "complying" because 50%
lamps failed before 6000 hours against the requirement of "Maximum 50%
failures in 6000 hours."

The definition of Average Life is: "The length of time during which
50% of the lamps reach the end of their individual lives."

This has been interpreted by the first lab as the highest life of the
first 50% failures.

The second lab argues that, as per the definition, if 50% lamps reach
the end of their lives after 6000 hours, then the average life is

6000 hours.

This means that one of the reports (I feel, the first one) is
incorrect. Hope I am not biased.

I have both the reports. But the first lab report is available with a
regulatory body who had initiated the life test.

I have given the second lab report to the regulatory body to convince
them that the sample COMPLIES. While the regulatory body is yet to
conclude, I request expert opinion on the issue from the lighting
community.

Regards.

Pankaj

This is not an average, it is a median. Average has a strict
mathematical definition, and would require measure ALL lamp failure
times to determine what it is.

Clearly the intent is to use median life here. Back when I was taught
about this in math classes, the median for 10 lamps would be the
average of the 5th and 6th lamp. This is essentially the best
estimate of what a larger population would have as a median.

So, you would need to measure up to the 6th lamp failure. Without
that, you cannot calculate the median.

Regards,

Mark

In this case the "correct" answer depends upon the test
procedure specified by the appropriate regulatory agency,
not mathematical purity. Unfortunately I do not have a
copy of the IEC spec and have no current plans to acquire
one.

--
Vic Robertshttp://www.RobertsResearchInc.com
To reply via e-mail:
replace xxx with vdr in the Reply to: address
or use e-mail address listed at the Web site.

This information is provided for educational purposes only.
It may not be used in any publication or posted on any Web
site without written permission.