HomeDemonstrating a Neon Glow Lamp's Dark Effect

This is a Sidebar of Main Article, A Closer Look at Fender Optical "Vibrato" (Really Tremolo)

Introduction. Appropriately, I first noticed the dark effect during a blackout caused by the remnant of Hurricane Ike blowing through this part of Ohio in September, 2008. For eight days, I kept two freezers in the basement and the kitchen's refrigerator happy on generator power. When these and other loads caused the gasoline-powered 4-KW generator's output to drop from about 125 VRMS to about 115 VRMS, the neon indicator lamp at the bottom of the old Kenmore upright freezer became photosensitive. In the dark, the lamp lit up only when I pointed my flashlight at it.

All neon glow discharge lamps have a dark effect (see General Electric's Glow Lamp Manual, either the First Edition [1963] or the Second Edition [1966], the 1979 Signalite Glow Lamp Manual, and the neon lamp application notes PDF at Visual Communication Company, LLC). Due to a special coating on the lamp's nickel electrodes, ambient light causes the surface of the (electrode serving as) cathode to emit photoelectrons when there is a potential difference across the lamp. These electrons are accelerated in the electric field, and some collide with atoms of the filler gas (neon), ionizing them. In turn, secondary electrons released by such collisions are accelerated, causing more collisions and ionization in a cascade or "avalanche" culminating in the lamp starting. Starting is also called "striking," or "breaking down" (into a relatively low-resistance state), or "discharging." The gas near the cathode begins to glow (called the "corona"). In ambient darkness, the lack of photoelectrons makes the lamp's static breakdown voltage significantly greater (see Table 1 in main article and associated discussion), and also delays striking once this threshold is exceeded. This is called the "dark effect."

As explained near the end of Section 7 in the main article, many neon lamps (including all high brightness indicator types) contain a radioactive additive, "seeding" ionization and reducing, but not eliminating, the dark effect. For every neon lamp, there is a critical, narrow voltage range near the breakdown threshold, in which illuminating the previously dark lamp causes it to strike. In other words, breakdown voltage depends on the ambient light intensity. As neon lamps age during use, cathode sputtering usually causes their breakdown voltage to increase. Most of this effect happens near the end-of-life in high brightness types (see G.E. manuals linked above). Operating on AC line power (where a dynamic breakdown threshold is approached or exceeded 60 times per second for each electrode), striking may become erratic--the lamp flashes randomly after aging has increased its breakdown voltage to approach the applied voltage. The next time you encounter an erratic neon lamp (perhaps on an old AC power strip), try changing the ambient light to see how the lamp reacts. As for the one in the old Kenmore freezer mentioned above, I happened to catch it at just the right age to observe the dark effect at 115 VRMS in 2008. Within a couple of years after that, the lamp no longer struck even at nominal 120-VRMS line voltage in bright ambient light.

A Demonstration. To demonstrate the dark effect, I used a neon lamp (the high brightness new-old-stock NE-2H described in main article) in a relaxation oscillator circuit. In this simple oscillator, the lamp is shunted by a 33-nF capacitor, and is in series with a 100-K resistor (see right-hand side of schematic diagram, below). When unlit, the lamp has very high resistance, and the capacitor charges through the resistor until it reaches the lamp's dynamic breakdown voltage. At this point the lamp strikes, and for a brief time while emitting a flash of light, it has a low resistance. The capacitor thus discharges until the lamp's extinguishing voltage is reached, when the lamp regains a high resistance and stops emitting light. The capacitor immediately begins to recharge and the cycle repeats. Since ambient light lowers the breakdown voltage, oscillator frequency increases as light intensity increases, because the time required to charge the capacitor to the breakdown voltage decreases. The G.E. manuals (see links above) say that neon lamps are most sensitive to near-ultaviolet (near-UV) light, so I used a Radio Shack 276-0014 "ultrabright" near-UV LED (typical wavelength = 405 nm) as the light source.

Schematic of experimental setup to demonstrate neon lamp dark effect
As shown in the left side of the above drawing, the near-UV LED was directly driven by a function generator, which was set to linearly ramp from 0 V to +10 V and then back again about six times per second. With a 330-ohm series resistor in place, the LED drew negligible current and was dark at input voltages below about 2.4 V, and approached its maximum rated current of 20 mA at 10 V (thus the LED was dark about 1/4 of the time; otherwise its light ouput approximated a linear function of input voltage when the latter was greater than about 4 V, or about 6/10 of the time). The following photo shows the LED illuminating the neon lamp (normally the pair was covered by a light-tight blanket; also, the capacitor was disconnected for this photo and the lamp lit to show position of its cathode [with orange corona], which is the photosensitive electrode).

Near-UV LED illuminating NE-2H neon lamp

As indicated in the schematic, one oscilloscope channel monitored the output of the function generator (LED drive waveform), while another the output of the relaxation oscillator (at the neon lamp's anode). A photo of the oscilloscope display is shown below. The relaxation oscillator makes a "sawtooth" waveform, with each "tooth" (cycle) being part of the exponential curve describing the voltage across a capacitor as it charges through a resistor. The bottom of each "tooth" is the lamp's dynamic extinguishing voltage, which appears to be about 75 V in this case; it is not light-sensitive. The top of each "tooth" is the dynamic breakdown voltage, which is light-sensitive. In the dark phase of the LED's cycle, the lamp's dynamic breakdown voltage approaches the oscillator's supply voltage (109 V). At maximum LED output, the dynamic breakdown voltage drops to about 106 V. As you have probably deduced, choosing a 109-VDC supply was not arbitrary; it was fairly critical and I found it empirically as the voltage giving the most presentable demonstration. (A volt or two lower made the oscillator too slow in the dark, and a little higher made the dark-light frequency difference less dramatic.)

Photo of oscilloscope display demonstrating neon lamp dark effect

As reported in Table 1 of the main article, the static breakdown voltages of this new-old-stock NE-2H lamp were 95 VDC in darkness and 90 VDC in bright near-UV light. Thus, the dynamic breakdown voltages are significantly greater than the static (by about 14 V in darkness). As explained in Section 7 of the main article, this is due largely to the lamp's ionization time, which delays discharge until the voltage has risen past the static breakdown threshold. (One expects the dynamic breakdown voltage to approach the static level at sufficiently low frequency.)

The red asterisks (* or **) on the above 'scope display indicate where I measured periods to calculate relaxation oscillator frequencies. I averaged the three cycles marked by single asterisks (*) to obtain a frequency of 121 Hz for performance in maximum light. For darkness, the three cycles marked by double asterisks (**) averaged 52 Hz. At such relatively low frequencies (where capacitor discharge time is very small compared to charging time), the 1963 G.E. Glow Lamp Manual gives the frequency for this type of relaxation oscillator as [the following image was copied from p. 16 of the manual, where it is called Equation 2.5]:
Equation 2.5 in G.E. Glow Lamp Manual, 1st edition (1963)
where f is frequency, R is resistor value in ohms, C is capcitor value in farads, V is the DC supply voltage, Ve is the lamp's dynamic extinguishing voltage, and Vf is the lamp's dynamic breakdown voltage. Solving this equation using the R and C values of the demonstration, and setting V = 109 and Ve = 74.9, plugging in Vf = 105.9 for maximum light intensity gives f = 126 Hz; using Vf = 108.9 for darkness gives f = 51.9 Hz. These frequencies are in close agreement with the ones I measured on the oscilloscope display.

A Note About the Oscilloscope Photo. Perhaps you are wondering why the vertical grid lines on the 'scope display are sloping. Unfortunately, the analog oscilloscope was in a strong magnetic field when the photo was taken, and I compensated the resulting distortion by skewing the image by four degrees. This also applies to the 'scope photos in the main article. To justify this graphical transformation, please see my sidebar on correcting magnetically skewed 'scope. As a result, I must ask you to consider the left and right edges of the photo as parallel to the vertical axis (voltage); don't follow the sloping vertical grid lines on the display. I apologize for the inconvenience.

Back to Main Article on Fender Optical Tremolo

Forward to Follow-Up Article on Fender Optical Tremolo

Sidebar on Magnetic Distortion of Oscilloscope

Back to Clark Huckaby Main Page (shows my email address)

Copyright 2013 Clark Huckaby. All rights reserved. Commercial use of this content without first obtaining the expressed permission of Clark Huckaby is prohibited by law. Distribution of copies beyond personal use and/or re-posting of this page is prohibited by law.