DC on pilot lamp filaments = short life

<< < (2/3) > >>

Tom WA3KLR:
Don,

From my old product catalog files:

GE Sub-Miniature lamps Catalog 3-6016 Revised May, 1976
Page 4   Lamp Life

Rated average life is that obtained in closely controlled laboratory testing of lamps on 60 Hz. alternating current at their design voltage.  Very long life lamps are generally rated on the basis of extrapolated laboratory test data.  Service conditions such as shock, vibration, voltage fluctuations, temperature, etc., may contribute to a shorter average life.

Ordinarily, for still-rack operation, normal tungsten filament evaporation is the basic force or mechanism controlling incandescent lamp life.  Where normal filament evaporation is the dominant failure mechanism, lamps should reach their design-predicted lifetimes.

In recent years, another life-influencing lamp filament mechanism has become more prominent.  This is commonly referred to as “filament notching”. Its prominence is due to at least three factors, primarily associated with sub-miniature type lamps:
1.   Low temperature filament operation, less than that for significant normal evaporation.  (Long life lamp designs, such as 10,000, 25,000, 50,000, and 100,000 hour designs.  This does not apply to filament temperatures below 1600 degrees C.)
2.   Small filament wire sizes, less than one mil (0.001”) diameter in many cases.
3.   Increased use of D.C. voltage operation (generally resulting from advances in solid state technology).

Notching is the appearance of step-like or saw-tooth irregularities, appearing on all or part of the tungsten filament surface, after some burning.  These notches reduce the filament wire diameter at various points.  In some cases, especially in fine-wire diameter lamps, the notching is so severe as to almost penetrate the entire wire diameter.  Thus accelerated spot evaporation due to this notching (as well as reduced filament strength), now becomes the dominant mechanism for influencing lamp life.  Because of its abnormal evaporation and/or reduced strength efforts, lamp lifetimes due to notching may be one-half or less of so-called ordinary or normal, predicted lamp lifetimes.

Photo below – “An example of filament notching”.

N8LGU:
If one is going to use DC on filaments that have been rated in AC terms then you should only use .707 of the pk AC rating. Does this seem reasonable?
Happy Thanksgiving all! We Am'ers have much to be thanlful for- such as the new gift from the FCC in terms of 80M reassignment! Dec 15 is just around the corner!

k4kyv:
Thanks.  The key word is "filament notching."  I did a Google search, and found loads of information on the subject.  This about sums it up:


MECHANICAL PHENOMENON
      Question:
      What is filament notching?
      Answer:
      DC notching is a phenomenon that occurs to an incandescent lamp when operated on DC voltage. The Tungsten material flows in the direction of the current and stops when it has reached its elasticity point and then starts to flow again. This movement creates various weak points within the filament wire, shortening the average rated life by approximately 50%.

But:

Question:
      What are the benefits of using noble gasses (xenon, krypton, argon) for increased life, and how does their use affect cost?
      Answer:
      Halogen extends operating life by recycling tungsten, evaporated from the filament and deposited on the inside of the glass lamp, back on to filament coil. It keeps the lamp from turning gray/black and allows brightness to remain at or near initial brightness specs for up to five times that of standard incandescent lamps. Lamps using an inert gas such as argon physically suppress evaporation of tungsten from the filament. The heavier the atomic weight of the gas used, the greater its effectiveness. This also increases the cost of the lamp, but provides the added benefit of NO reduced life operating on DC voltage, and a whiter, more efficient light that is very desirable where color is a serious consideration.

http://jkllamps.com/techLibrary/index.cfm?action=techLibraryFAQs


The lamps in question are in my Gates BC1-T.  I run the relays on DC, and the same DC power supply lights up the filaments-on and plate-on indicators.  The original lamps are rated at 15 watts @ 220 v.a.c. but I run mine on 110v DC, and have not been able to find  replacement 220v  lamps, so I use the equivalent 110 volt ones.  In my case the solution will be to increase the series resistance, thus dropping the filament voltage below the rating, to compensate for life reduction due to the notching effect.  A series resistance also limits surge current, another factor that shortens filament life.

I run my 75A-4 with the equivalent 12-14 pilot lamps instead of the stock 6-8 volt ones.  Not only does this increase life from a few months to several years per lamp; it gives them a nice, warm, mellow amber glow instead of the harsh bright white glow of stock lamps.

W1RKW:
Not to steer this thread in a different direction as it is an interesting thread nonetheless but this now this begs the question of using DC on tube filaments.  The audiofools are sucking up the classic transmitting tubes for their audio projects and run them on DC from the few tube audio articles I've read to reduce noise and hum. Metal migration, just another factor to consider in the tube drain down by these folks.

Tom WA3KLR:
The mechanism is only a problem with filament wires around 0.001" diameter or less.  Tubes that the audiophools would use are not in this boat, unless they have fallen in love with 1R5's now - a 50 milliampere filament.

I don't have a table of tube and pilot bulb filament diameters, but I would guess that a filament greater than about 100 milliamperes is out of range of the problem.  Perhaps someone else has access to tube or bulb filament mechanical specifications.

12AU7, 12AX7   150 milliamperes
6SN7    600 milliamperes
6F5    300 milliamperes
6A3, 6B4  1 ampere
300B   1.2 amperes
45      1.5 amperes
805    3.25 amperes
810    4.5 amperes
811, 812  4.0 amperes

Shucks.

Navigation

[0] Message Index

[#] Next page

[*] Previous page