Honda Car Forum

No manufacturer rates a product's quantity or life
expectancy with a statistical average. Its not complicated.
When they say a battery is expected to last 48 months, they
don't mean half the batteries will last less than 48 months
and the other half will last more than 48 months. That would
be an 'arithmetic mean' of 48 months.

Instead, something like 99.99% of those batteries must last
48 months ... or longer. They provide a minimally acceptable
life expectancy number. It is called a Confidence level.
What's so difficult about this? It's a standard concept in
any first level statistics course. Why are you now getting so
emotional over a mathematical concept that is so trivially
simple? Why do you insult because you did not know this?

Expected average life is a Confidence level. They don't
publish a number that 50% of the batteries will not meet.
They publish a number that most every battery is expected to
meet. That is called a Confidence level.


You do remember the 1970s. If that 24 month battery failed
in 14 months, and if you still had a receipt, then the
pro-rated warranty was automatically honored - no questions
asked. All batteries either exceeded their rating, or would
be replaced by that warranty. So where does this nonsense
about batteries lasting too long have any relevance? Or is
that just another silly insult.

There is no lecture here. Just simple business math. Its
called a Confidence level.

A lead acid battery's performance is directly traceable to
temperature - which is why they are rated with a cold cranking
amperage. However performance and battery failure are two
completely different things. Temperature is completely
irrelevant when a battery will not even hold a charge.
Temperature makes no difference when one cell short circuits
from a pile of debris shorting electrode plates. Another
failure condition estimated by the Confidence level.

Please read carefully before jumping to conclusions. I
never said a battery not starting at -20 degree failed. I
noted that a battery that had frozen when new still managed to
well exceed its confidence level. Completely different meaning
from what you assumed. Frozen batteries (like over discharged
batteries) don't always last much longer. This one did. It
was still working five years later.

Your lecture about battery performance is noted as accurate,
but is simply preaching the the choir about something
completely irrelevant. A battery that should have failed
before its confidence level predicted (because it was frozen),
instead went on to long exceed its life expectancy. Now that
it is posted a second time, do you understand - and notice
how my tone has changed because you insult with arrogance
rather than learn what a confidence level is.

I guess you never took any introductory statistic lessons in
school. Shame. In Japan, simple statistics is required of
every Jr High School student - even future battery salesman.
Just another insult you earned.

Now, maybe you would like to stop with the insults?

Caprice Classic wrote:
> Apparently "w_tom" <w_tom1@hotmail.com> likes to make something
> very simple very complicated.
>
> In my previous post, my quotation of "expected average life" is in
> reference to aftermarket replacement of OE batteries and not
> confidence index as "w_tom" <w_tom1@hotmail.com> suggests.
>
> My purpose was not to imply any reference to warranty. Warranty
> in those days (70's)was secondary to brand popularity.
>
> He is the one that dreamed up the "confidence index"? and his
> accompanying doctoral dissertation. The battery industry has
> used for statistical purposes a 90 percent confidence level for
> published ratings.
>
> Based upon the thousands of scrap batteries I have had to personally
> prepared for proper disposal and shipment, I simply say rubbish to his
> lecture. By simple observation, one learns to read manufacturer date
> codes and deduce a scrap/spent battery's age. After palletizing
> about 20 per layer and 3 layers banded or shrink wrapped on a 40"
> wide by 42" long pallet loaded upon a semi trailer with about 19
> others at least 2 to 3 times per week was a lot of work 52 weeks per
> year for about 25 years.
>
> Maybe "w_tom" <w_tom1@hotmail.com> sat in a office playing with his
> slide rule and a chalk board dreaming of one day he would get a
> chance to lecture about the battery industry.
>
> No one cares about Coke bottles.
>
> "Failed" can be interpreted from within a limited warranty statement
> "failing to hold a charge based upon defects in materials or
> workmanship".
>
> Accept the fact if widget lasts longer than its warranted life, it
> has lasted longer what is average without a doctoral dissertation.
> You have simply beat the averages. So what about warranty! Maybe
> you would be glad to pay the manufacturer for that extra time that
> you enjoyed the benefit of the longer life ACDelco branded product.
>
> As for his "1976 GM Freedom battery (that 'no add water' battery,
> an early calcium lead type?, became the 1970s standard) was still
> working in 1981 when I sold the car." story.
> A lead acid battery's performance based upon the temperature of the
> battery is a fact of life.
> Just because a vehicle fails to start due to a -20F degree
> temperature does not mean the battery is defective. A lead acid
> storage battery performs close to 100% efficiency at 80 degrees F
> and about 40% at 0 degrees F. So depending upon Reserve Capacity,
> one may or may not get their car started in sub 32 degree F weather!
>
> snip

w_tom wrote:
>
> No manufacturer rates a product's quantity or life
> expectancy with a statistical average. Its not complicated.
> When they say a battery is expected to last 48 months, they
> don't mean half the batteries will last less than 48 months
> and the other half will last more than 48 months. That would
> be an 'arithmetic mean' of 48 months.

No, that would be the median, not the mean.

> Instead, something like 99.99% of those batteries must last
> 48 months ... or longer. They provide a minimally acceptable
> life expectancy number. It is called a Confidence level.
> What's so difficult about this? It's a standard concept in
> any first level statistics course. Why are you now getting so
> emotional over a mathematical concept that is so trivially
> simple? Why do you insult because you did not know this?

The "difficulty" with this is that you made up the 99.99% number.

> Expected average life is a Confidence level. They don't
> publish a number that 50% of the batteries will not meet.
> They publish a number that most every battery is expected to
> meet. That is called a Confidence level.

> You do remember the 1970s. If that 24 month battery failed
> in 14 months, and if you still had a receipt, then the
> pro-rated warranty was automatically honored - no questions
> asked. All batteries either exceeded their rating, or would
> be replaced by that warranty. So where does this nonsense
> about batteries lasting too long have any relevance? Or is
> that just another silly insult.

The duration of the warranty is not the expected life. The life
expectancy of a car in the US is around 17 years, the warranty is
significantly shorter. On the other end of the scale, my tire pressure
gauge has a lifetime warranty even on the battery. I do not expect the
battery to last another 50 years, but I sure hope I will.

"w_tom" <w_tom1@hotmail.com> wrote in message
news:3FF35AF6.DB6D6156@hotmail.com...
> No manufacturer rates a product's quantity or life
> expectancy with a statistical average. Its not complicated.
> When they say a battery is expected to last 48 months, they
> don't mean half the batteries will last less than 48 months
> and the other half will last more than 48 months. That would
> be an 'arithmetic mean' of 48 months.
>
> Instead, something like 99.99% of those batteries must last
> 48 months ... or longer. They provide a minimally acceptable
> life expectancy number. It is called a Confidence level.
> What's so difficult about this?
<snip>
> Now, maybe you would like to stop with the insults?

And maybe you should stop trying to blow smoke up everyone's ass. You bring
to mind the old saying that if you can't dazzle 'em with brilliance baffle
'em with bullshit.

In my previous posts, I shared my practical experience in direct observation
of scrap/spent secondary lead acid wet cell battery failure condition
starting in the 1970's. This was in reference to another poster making
allegation to present day batteries can be expected to last up to about 10
years.

It is amazing how warranty can be misconstrued to imply product quality by
many aftermarket marketers. Those aftermarket vendors go to market to
wholesale purchase commodities to satisfy a pent up demand. Many times they
choose how products will be warranty branded based upon competition and
perceived value to the consumer. Why do many consumers fall into the 7 year
battery trap when a 50 month or 5 year competitively rated product will
suffice? To quote P. T. Barnum, "There is a sucker born every minute."
So the "simple" statistical explanation by ""w_tom" <w_tom1@hotmail.com>"
doesn't hold "water" and is refutable.

Usually, the only time statistics mattered in my former job was when a
manufacturers
brand quotation submitted to a volume prospect requested engineering
ratings. Included in the Bid or Request For Quotation (RFQ), it was
required for my quote to be considered that I personally and corporately
certify the specifications by both my signatures and sometimes accompanied
by a performance bond. Warranty as somewhat erroneously explained by
""w_tom" <w_tom1@hotmail.com>" was a secondary consideration due to my
presenting a manufacturers brand. I cheerfully complied due to my
eagerness to help the cash flow of my business by paying creditors and
employees and finally myself.

The US battery industry has shared with many interested parties failure mode
analysis results of OE batteries and their average age upon replacement with
aftermarket batteries. I do not know any other simple manner of
explanation.

What planet did ""w_tom" <w_tom1@hotmail.com>" arrive from? Can one explain
how certain newsgroups spawn rocket scientists?

Since when does products that reach the end of the production line know
whether how each will be marketed based upon warranty printed on different
labels to different aftermarket customers?

In the future will you ""w_tom" <w_tom1@hotmail.com>" provide a syllabus?
"It's a standard concept in any first level statistics course." So what!

Who asked you to expand on trivial matters? "Why are you now getting so
emotional over a mathematical concept that is so trivially simple?" Who as
a consumer cares about your trivial simple mathematical concepts?

Apparently in your manner of pea brained deduction there is only the
""w_tom" <w_tom1@hotmail.com>" manner of thinking. I am comforted to be
informed by you of your vast knowledge! So when will you answer the help
wanted ad by NASA?

"Why do you insult because you did not know this?" Please get a life and
cease and desist from your babbling.

Last but not least please take a reading comprehension course. I have the
experience that you probably wish you had in your wildest dreams.

Sincerely,

A former battery specialist distributor manager!
snip


---

Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.557 / Virus Database: 349 - Release Date: 12/30/2003

If you have an automotive alternator driving a resistive load (no
battery in the circuit, negligible capacitance and inductance) the
ripple voltage would be about 1.4V (peak to peak) or about 0.5V (RMS).
This is assuming a three-phase alternator with the output DC voltage
regulated to 14.5 V RMS.

The ripple voltage changes very little with load current.

When it comes to judging the health of the alternator, the magnitude of
the ripple is not as important as the shape. Normally it will look
something like:
_ _ _ _ _ _ _ _
/ \_/ \_/ \_/ \_/ \_/ \_/ \_/ \_


If you have a failed diode (open) in the rectifier you will have
something like:
_ _ _ _ _
/ \_____/ \_/ \_____/ \_/ \_____


If you have a failed diode (short) in the rectifier you will have
something like:
_ _ _ _ _
/ \__ __/ \_/ \__ __/ \_/ \__ __
V V V

Some ripple is inherent in the system and does not cause damage. It can
sometimes cause noise in poorly designed audio equipment.

Voltage fluctuations CAN cause damage. If the alternator is full-fielded
under moderate load the output voltage can get very high. The service
manual for my '94 Civic warns against letting the voltage rise above 18V
during fullfield testing of the alternator. A common failure mode for
voltage regulators is that the alternator is fullfielded.


Dave wrote:
>
> Just thought I'd throw in another alternator question in this thread. I've been
> getting intermittent voltage fluctuations in my electrical system. The
> fluctuations can be seen the in lighting system and by the in-dash volt meter.
> An oscilloscope on the battery terminals with the car at idle shows +/- 0.2
> volts at approximately 6Hz (higher Hz at higher RPMs). I had my charging system
> tested at an auto parts store and it turned up ok, with the battery being a
> little weak. About 8 months ago I had a voltage regulator go out and was
> over-charging the battery for a little while. Since then the alternator was
> replaced.
>
> My question are:
> - What is the acceptable +/- voltage that a properly functioning alternator
> supposed to output, if any?
>
> - Will the voltage fluctuations cause damage to other systems in my car?
>
> - Is it possible that my slightly weak (and possibly damaged) battery isn't
> "filtering" the voltage fluctuations?
>
> I plan on pulling the alternator and having it bench tested, but I'm waiting
> until the weather clears.
>
> Thanks,
> - Dave
>
> Tibur Waltson wrote:
> > Today I load test the alternator by an instruction in an aftermarket repair
> > manual. The instructions says to turn on headlights, defoggers, fan, radios,
> > cigarrete lighter, etc and measure the voltages. It reads 13.6V at 2000 rpm
> > or 11.2V at 750 rpm. It's within normal range according to the book. But the
> > headlight seems a bit weak unless rpm is above 1300 rpm or sometimes it just
> > too weak for the starter. We hardly drive in the highway. It has an alarm
> > system. Is the alternator fine?
> > TIA
> > Tibur

The 99.99% number was provided as example. The 99.99%
ballpark number was created, as example, to keep the concept
simple.

The term 'average life expectancy' as it was introduced by
another referred to how batteries were sold in those days.
Yes, the number of months stamped on a battery meant every
battery should last beyond that date. But in the early 1970s,
if a warranty was for two years, then the product was often
failing in less than three. Therefore many assumed that 24
months was the average life expectancy. Reliability in the
1970s was so bad that many cars needed a whole valve job or
new cam by 20,000 or 40,000 miles - typically just after the
warranty had ended.

There should be little relationship between warranty and
life expectancy. Warranty is some number less than the life
expectancy.
But in the 1970s, life expectancy for some products was often
little beyond the warranty period for too many products.
Therefore for a 24 month battery, many simply assumed that 24
months was, also, its life expectancy.

This was a period of Firestone 500 paraplegics and
quadriplegics when Sears promoted a different concept. Tires
that did not fail prematurely (Michelins) and a battery that
really did last longer (Sears Die-hard). Today we don't worry
so much about warranty periods because batteries typically
last well beyond those numbers. Concept of 'failure is
normal' that was so prevalent in the 70s is really unknown to
so many today. Today, car batteries last on the order of
seven and nine years. In the 1970s, if I tried to sell an
eight year battery, one would probably assume it was a scam.
It was a different time when premature failure (and even cars
rusting out starting in two years) was quite normal.

Randolph wrote:
> w_tom wrote:
> ...
>> Instead, something like 99.99% of those batteries must last
>> 48 months ... or longer. They provide a minimally acceptable
>> life expectancy number. It is called a Confidence level.
>> What's so difficult about this? It's a standard concept in
>> any first level statistics course. Why are you now getting so
>> emotional over a mathematical concept that is so trivially
>> simple? Why do you insult because you did not know this?
>
> The "difficulty" with this is that you made up the 99.99% number.
>
> ...
> The duration of the warranty is not the expected life. The life
> expectancy of a car in the US is around 17 years, the warranty is
> significantly shorter. On the other end of the scale, my tire pressure
> gauge has a lifetime warranty even on the battery. I do not expect the
> battery to last another 50 years, but I sure hope I will.

AC Delco seems to believe that with an average life of 63 months, they
have the industry average beat by 33%. See
http://www.acdelco.com/html/pi_batt_auto_study.htm

w_tom wrote:
>
> The 99.99% number was provided as example. The 99.99%
> ballpark number was created, as example, to keep the concept
> simple.
>
> The term 'average life expectancy' as it was introduced by
> another referred to how batteries were sold in those days.
> Yes, the number of months stamped on a battery meant every
> battery should last beyond that date. But in the early 1970s,
> if a warranty was for two years, then the product was often
> failing in less than three. Therefore many assumed that 24
> months was the average life expectancy. Reliability in the
> 1970s was so bad that many cars needed a whole valve job or
> new cam by 20,000 or 40,000 miles - typically just after the
> warranty had ended.
>
> There should be little relationship between warranty and
> life expectancy. Warranty is some number less than the life
> expectancy.
> But in the 1970s, life expectancy for some products was often
> little beyond the warranty period for too many products.
> Therefore for a 24 month battery, many simply assumed that 24
> months was, also, its life expectancy.
>
> This was a period of Firestone 500 paraplegics and
> quadriplegics when Sears promoted a different concept. Tires
> that did not fail prematurely (Michelins) and a battery that
> really did last longer (Sears Die-hard). Today we don't worry
> so much about warranty periods because batteries typically
> last well beyond those numbers. Concept of 'failure is
> normal' that was so prevalent in the 70s is really unknown to
> so many today. Today, car batteries last on the order of
> seven and nine years. In the 1970s, if I tried to sell an
> eight year battery, one would probably assume it was a scam.
> It was a different time when premature failure (and even cars
> rusting out starting in two years) was quite normal.
>
> Randolph wrote:
> > w_tom wrote:
> > ...
> >> Instead, something like 99.99% of those batteries must last
> >> 48 months ... or longer. They provide a minimally acceptable
> >> life expectancy number. It is called a Confidence level.
> >> What's so difficult about this? It's a standard concept in
> >> any first level statistics course. Why are you now getting so
> >> emotional over a mathematical concept that is so trivially
> >> simple? Why do you insult because you did not know this?
> >
> > The "difficulty" with this is that you made up the 99.99% number.
> >
> > ...
> > The duration of the warranty is not the expected life. The life
> > expectancy of a car in the US is around 17 years, the warranty is
> > significantly shorter. On the other end of the scale, my tire pressure
> > gauge has a lifetime warranty even on the battery. I do not expect the
> > battery to last another 50 years, but I sure hope I will.

not to stray from the already strayed, but.. you can full field load
test those old honda alternators with a paper clip. there's a hole in
the back that you can push a contact and bypass the regulator.
Chip

On Thu, 01 Jan 2004 22:07:02 GMT, Randolph <trash@junkmail.com> wrote:

>AC Delco seems to believe that with an average life of 63 months, they
>have the industry average beat by 33%. See
>http://www.acdelco.com/html/pi_batt_auto_study.htm

As I've already pointed out, "average" includes a large number which are
abused. When I see people walk away from their cars leaving the headlamps
on because they know they have a handy-dandy auto-off switch I see umm,
ignorance. Of course the same people will leave the headlamps and likely
radio, vent fan etc in on position when they next start the engine - IOW
simple abuse.

Obviously the timeframe means that observations are difficult to accumulate
but personally I expect to get 7-8 years from a battery nowadays, just by
observing the obvious precautions. My '92 Integra original battery lasted
71/2 years and it had those nutty auto-retracting seatbelts. I used to
leave the door open while starting the engine... to avoid whacking the
battery with the retracting mechanism jolt. When it did eventually need
replaced, it failed quite gradually - the door open trick was enough to get
it going 1/2hour after my wife had found the engine wouldn't start with the
door closed.

Rgds, George Macdonald

"Just because they're paranoid doesn't mean you're not psychotic" - Who, me??

"Tibur Waltson" <Toush@hi.com> wrote:

> Today I load test the alternator by an instruction in an aftermarket
> repair manual. The instructions says to turn on headlights, defoggers,
> fan, radios, cigarrete lighter, etc and measure the voltages. It reads
> 13.6V at 2000 rpm or 11.2V at 750 rpm. It's within normal range
> according to the book. But the headlight seems a bit weak unless rpm
> is above 1300 rpm or sometimes it just too weak for the starter. We
> hardly drive in the highway. It has an alarm system. Is the alternator
> fine? TIA
> Tibur
>
>
>
>

Once you do this if you turn off all the loads does the voltage quickly
float up to 13.6 volts? If so then the alternator is prolly' okay. You
may just need a battery. How old is it and of what quality? Are you near
and Autozone or Sears? They will check the system out for free, or a
small fee.