Re: turn signal

[John Denker <jsd@AV8N.COM>]

David Bowman wrote:
> I haven't carefully surveyed the turn signal behavior of all the
> vehicles on the market, but in my experience I have noticed that a
> single burned out bulb (i.e. a finite increase in circuit resistance)
> can cause a turn signal flasher to flash a) significantly faster,
> b) much slower, c) not at all, or d) with no noticeable change in
> frequency depending on the particular vehicle and flasher in
> question.  Overall, with the cars I've had an association with, it
> seems that case a) is a little more common than the other cases b),
> c), & d), but that may be due to my limited sampling statistics.

This is an admirable example of the scientific approach:
 -- more thorough than most other discussions in this thread,
 -- separating the _observations_ from the _interpretations_
    placed upon the observations,
 -- and yet simultaneously modest, pointing out the limitations
    of the reported evidence.


David T. Marx wrote:
> A write up of the turn signal circuitry and mechanisms is at
>
> http://auto.howstuffworks.com/turn-signal.htm/printable

That writeup has more glitz than substance.  It alleges there
is a delay between switching on the blinker and seeing the
light come on.  This behavior is uncommon and IMHO objectionable.
To suggest that that is "the" way turn-signal flashers work
is not right.

I have had a number of bad experiences with "Marshall Brain"
and his howstuffworks site, including
 -- stuff that was apparently plagiarized, and
 -- stuff that was wrong.

Beware.  Just because a site is large and well-known doesn't
make it reliable.  Just because it has good "production values"
doesn't make it respectable.

====================

Carl Mungan asked:

> Wouldn't it be great to have a [failure] indicator for
> headlights?

My car tests the headlights and brake lights on every
startup, and complains if there's an outage.  The car is
eight years old, so this isn't exactly a new invention.
It's a nice feature, and it can't be very hard to
implement, so I'm surprised it hasn't become very common.

======================

Not all flasher units work on the same principle.
 *) Some use a heater + bimetallic strip.
 *) Some use a motor + cam.
 *) Some use an RC circuit driving a relay.
 *) Nowadays more and more are all-electronic.

So don't imagine that the first thing you run across is the
general case.

I'm reminded of the old riddle:  Which is more likely to fall
on Friday the 13th:  Christmas or Easter?  (The point is that
an amazing number of people stop their analysis too early, as
soon as they come up with a pet theory that fits _some_ of the
data.  In contrast the scientific approach is to test each
proposed theory against _all_ of the data.)