🚀 go-pugleaf

amichiel@rodan.acs.syr.edu (Allen J Michielsen) writes:

>In article <4225@kitty.UUCP> larry@kitty.UUCP (Larry Lippman) writes:
>   ... in a truly esoteric arguement of lumens and baked beans... 8-)

>>In article <5232@rsiatl.Dixie.Com>, jgd@Dixie.Com (John G. DeArmond) writes:
>>> Here is how to extend your bulb life indefinitely with practically (<1%)
>>> decrease in output.  Place a negative tempco thermistor of the proper
>>> value in the lamp lead.  The thermistor limits inrush and self-heats to
>>> a low resistance value in about a second.

>>	While the above installation technique may function as intended,

>  If you can't clearly see that the above 'application' isn't a functional
>solution then you shouldn't make ANY coment on it.  I haven't reviewed the
>power ratings discussed, but believe (agree) that this type of application
>is outside the normal design parameters & specs, but probably within physical
>normally distributed (fire/explosioin) limits of the device(s).  

Allen, you just gotta overlook Larry the Lid (as he's referred to in other
groups.)  I induced him to make an ass out of himself in comp.dcom.telecom
a few months ago and he's been messing his britches to get even ever since.
He's quite fond of quoting the hypothetical and or calculated answers as
fact.   Just a spoiled rich kid not getting his way.

>    The funny part here is that while ul listed, these button type jobs are
>a common cause of fires, and very popular (winning) lawsuit item.  I have
>a friend that testifies and consults, and has made a ton of money off these
>little jobs blowing up.  The UL rating was either a result of poor government
>inoperation or just a simple bribe.  


I'll second that motion.  I tried some of the little buttons in a 4 bulb
chandelier in my den.  By the time they had failed and shorted under 
100 watt bulbs, they had emitted a large quantity of smoke and had 
fried to a blakened crisp.  *I* consider them a fire hazard after my
experience.

Let's get back to the thermistors.  I rarely post the entire history of
events that leads up to my recommendations.  After all, I'm not writing
magazine articles here on the net.  I do thoroughly test them whenever
possible.  This is such a case.  The thermistors specified are meant for
surge or inrush limiting applications.  I thoroughly tested the application
before I made the recommendation.

Because this is a new application for the parts, different than the
normal use in switchmode powersupplies, I set up a test environment in my
lab that consisted of a  light fixture representative of what is in my
house, a variac to vary the applied voltage and a thermocouple datalogger
to record what thermal events occured. One thermocouple was fastened via
thermal epoxee to the thermistor and the other to the base of the lamp
socket.  A 100 watt bulb  and a PAR reflector lamp were tested.   

The worst conditions  observed were with the 16 ohm thermistor and the 100
watt lamp.  The 60 degree C rise was the maximum measured.  What was 
interesting was that the base of the socket got hotter than the thermistor.
Considering that the socket was the type insulated with paper inside a 
brass-like shell, I'll let you draw your own conclusions.

Let's look at some of Larry the Lid's hallucinations and compare them to
experimental data.  From here on down, the single ">"s are Larry the Lid's
comments.


>	While the above installation technique may function as intended,
>readers should be aware that such an "aftermarket" installation: (1) most
>likely voids the UL approval on the lighting fixture in question; 

Since doing anything to a fixture that deviates from the UL test conditions
voids the approval, (such as putting all those 100 watt bulbs in "60
watt max" sockets), at least one can say that "voiding" is common and
does not cause many fires.  As we'll see, the *practical* safety is
not degraded and if anything, is enhanced.

>(2) is
>a National Electrical Code (NEC) violation; 

This one is a bit easier to refute.  A careful purusual of the 
1990 NEC will turn up only vague admonitions about fixture
designs such as 

410-24(a) "Insulation - Fixtures shall be wired with conductors having
insulation suitable for the environmental conditions, current, voltage
and temperature to which the conductors will be subjected."

41-68 "Fixtures shall be so constructed that the adjacent combustible
material will not be subject to temperatures in excess of 90 degrees
C."  (referring to the special case of recessed fixtures.)

In other words, use good engineering judgement and keep the surface
temperature of adjacent materials below certain values.

>(3) most likely violates local
>building codes because of (1) and (2); 

Hardly.  Most areas adopt the NEC in total.  I've had other occasions to
check here in Cobb county and know that this is the case here.  Any
hands wringers should probably check with their local electrical inspectors
before improving the life of their lightbulbs.  Don't be surprised if you
get a "Duh, I don't know" if you do ask.  While you're asking, be sure 
and ask about those extension cords that just sorta became a part of 
the lampand those non-waterproof outlets in your open garage and all the 
other little things that aren't strictly according to code.

>	Permissible wiring within lighting fixtures is tightly controlled
>and specified by the NEC due to elevated temperatures found within such
>fixtures.  I know of no lamp life extender which is UL-approved for the
>series wiring installation (as described above) within a lighting fixture,
>and based upon my knowledge of UL "philosophy", I doubt that such a device
>could ever meet UL approval.

Actually the NEC does not address the subject.  The closest references in
the index are to "thermal devices" and "thermal overcurrent protectors"
and these refer to motor protection.  And since we know that UL approves
the diode devices which WILL flame when they fail, we could postulate that
passing UL testing for a non-flammable, non-flaming and non-smoking
device would be trivial.


>	How could one possibly have a fire, you may ask?  Simple.  Consider
>a light bulb whose filament opens, with a short filament section drooping
>against a support wire, thereby creating a *much* lower resistance.  While
>such lamp faults usually burn open within the lamp envelope, a virtual
>short circuit can also result.  Most readers have probably seen the effects
>of a burned out lamp which is then jostled while under power.  Consider
>now that we have such a low resistance fault drawing *much* more current
>than a normal lamp, and consider the effects of such overcurrent on the
>thermistor.  It is quite possible that the thermistor could burn under
>such circumstances.  
>
>>The button-type devices are deemed safe because
>they are not only enclosed within the metal screw area of the lamp socket,
>but are further insulated by the dielectric surrounding the socket (often
>a ceramic material).

But since we know that the button devices, when subjected to the high lamp
base temperatures, DO catch fire and burn, we know that this statement
is false.

But let's address the issue Larry the Lid raises above.  Even though we
know that a filament short is always blown away instantly upon contact
and because we could speculate that a lamp construction that would allow
a continuous short between electrodes could never be approved, it is
reasonable to wonder what a hard short would do to the termistor. So  in
the interest of completeness, I decided to test his hypothesis. 

Since power dissipation would be worst case with a cold, high resistance
termistor, my testing was done at ambient.  My test rig is simplicity
itself - a ptn-126 thermistor clipped in a  Jesus Cord (a 120 volt cord 
with aligator clips on the end - makes you yell "JEEEESUS" when they get
together and can aid you in meeting Him if you grab 'em :-) and laying on
a paper towel for fuel.   Plug the cord in and ... POP!  The thermistor
blows apart instantly. No flame and no smoke.   This is, of course, many
times more severe than conditions that exist in the light fixture.   

We must remember that the ambient temperature in a fixture is MUCH higher
than the 60 degree rise, (80 degree C absolute) measured on the surface
of the thermistor. 

I plan on setting up an accelerated life test with a cycle timer that
will cycle the lamp on and off at about 10 0r 15 second intervals. I'll
probably conduct at least part of the test with fiberglass insulation
wrapped around the thermistor to see if that makes any differnece.  I'll
post any interesting results. (Well hey! It's another excuse to throw
another 8048 at something :-) 

> Typical on resistance for the 16 ohm device is 1 ohm which means that it
> dissipates about a watt with a 100 watt bulb. ...
> The part will rise
> about 60 degrees C over ambient when in operation so you don't want
> to insulate it or pack it in a closed box.

>	The 16-ohm device is actually a Panasonic P/N PNT-126, and has
>a maximum rating of 0.6 watts.  Its R25/R50 deg C resistance ratio is 2.3,
>which means that its resistance is still about 7 ohms at 50 deg C.  I
>would find it difficult to believe that at 85 deg C ("60 degees C over
>ambient") its resistance would be as little as 1 ohm.  2.5 to 3 ohms
>seems more likely, which means we could be dissipating at least 2.5
>watts.  That's *FOUR* times the *maximum* rating of the thermistor!

The actual *MEASURED* (as opposed to hypothesized or made up) voltage
drop across a PTN-126 while wired in series with a 100 watt bulb ranges
from 0.8 to 1.0 volts as measured with a Fluke 8800A 5-1/2 digit
voltmeter with the true RMS option installed and an NBS (ne NITS)
traceable  calibration.  Since the measured current is just a bit under
an amp, the computed resistance is about an ohm and the  power
dissipation is about a watt.  We know from experience that Larry finds
experimental data "difficult to believe".  Fabricated numbers are *SO*
much more tidy. 

>
>	Also, the above thermistor series is primarily intended as a
>temperature measuring element, and not for any significant power dissipation
>in any protective circuit.  That's why its only rated at 0.6 watts.

Low ohm negative tempco thermistors with significant watt ratings are 
designed for inrush applications and make pretty poor measuring devices,
as a glance at the thermistor curves would show.  Since the 600 milliwatt
rating is for worst-case conditions, I have no problem at all operating
the device at higher dissipations under less severe thermal conditions.
After all, that is all that really matters.  One could run almost any
dissipation through the device if the cooling were adequate.  I've always
considered operating temperatures of less than 100 degrees C in 
resistance devices to be quite acceptable.
>
>	In my opinion, a much better, simpler and lower-cost alternative
>to any type of thermistor device (even the button-type) is to find a
>wholesale electrical supply firm which will sell you lamps rated for
>130 volts (instead of the usual 120 volts).  Such 130 volt lamps are
>also available rated for low-temperature and "rough duty" service, and
>those suckers will usually last for years in a home environment.  I have
>one above my kitchen sink (a bitch to replace because of a curtain rod)
>that has been running for almost *4* years with a *lot* of on-off cycles.
>
>	Another very robust lamp available from a wholesale electrical
>distributor is rated for traffic signal use.  The only disadvantage of
>this type of lamp which limits its application for home use is that it
>is available only in a clear envelope (i.e., not frosted). 

I can't argue with this advice.  If you can stand the reddish light and
the glare, go ahead.  I simply find that the the process of soft-starting
a regular bulb gives me the best of both worlds - a good color temperature
and long life.

John

-- 
John De Armond, WD4OQC        | "Purveyors of speed to the Trade"  (tm)
Rapid Deployment System, Inc. |  Home of the Nidgets (tm)
Marietta, Ga                  | 
{emory,uunet}!rsiatl!jgd      | "Vote early, Vote often"