General Xenon Flash and Strobe Design Guidelines

That will usually actually work!

updated slightly 12/29/99



These are approximate guidelines to determine voltages, energy levels, and capacitances to get almost any xenon flashtube working if you can't get the ratings for your flashtube. These are not guaranteed to work or even be safe for every flashtube.

Xenon flashtubes require high voltages, often at hazardous to highly deadly energy levels. You can be badly shocked and maybe killed. You must observe all appropriate precautions. Go here for some safety guidelines, not necessarily including everything you need to know.

No warranty, please read my disclaimer.


You should measure the discharge path length of the flashtube, from the tip of one electrode to the tip of the other, in millimeters. You should measure along the centerline of the tubing. At full power, the discharge has only a negligible tendency to concentrate to the inside of turns.

You also need to know the inside diameter of the tubing in millimeters. This is usually around 75-80 percent of the outside diameter. All references to diameter will be inside diameter, even if this is only estimated.

You don't need absolutely perfect measurements; these guidelines are broad enough to leave you room for error. If you measured something in centimeters, multiply by 10 to get millimeters. If you measured something in inches, multiply by 25.4 to get millimeters. All measurements in my guideline formulas below will be in millimeters.

These guidelines work best for flashtubes with inside diameters ranging from 2 to 10 millimeters.

You will probably also want to determine whether the flashtube is made of glass or quartz. If the flashtube is known to handle very high energy or power levels that greatly violate the guidelines below for glass flashtubes, then the flashtube is probably made of quartz or "Vicor" or a similar super-tough glass. You can hook the flashtube to a neon sign transformer or a spectrum tube power supply and light it up and see if you smell any ozone. If you smell ozone, the flashtube is made of quartz. (Some quartz tubes are made of specially doped quartz that blocks ozone-forming UV.)


The xenon discharge is close to being a "graybody" radiator. It emits a very broad, continuous spectrum. There are also emission lines in the spectrum, but they account for only a small fraction of the total radiation. The discharge is not opaque to its own radiation; in fact, it is usually closer to being transparent.

If you use less voltage than recommended, the discharge may not work well as a graybody radiator, and radiation may be heavily from infrared emission lines. If you use too much voltage, you increase ultraviolet output from the discharge, possibly causing overheating or thermal shock to the glass. Excessively high voltages will also make the discharge more like a blackbody, causing it to absorb much of the radiation emitted by the interior region of the discharge. This will cause uneven temperature, possible excessive pressure, and possible excessive mechanical shock effects.

If you use less energy than recommended, a large fraction of the energy may be used just to heat the xenon to the point it becomes a good graybody radiator.

My maximum recommended voltage, in volts: Raise the inside diameter to the minus .7 power, then multiply by the discharge path length. Multiply the result by 20, then add 20 volts to this. If this is less than 340 volts, then I recommend 340 volts as an upper limit.

My minimum recommended voltage, in volts: Raise the inside diameter to the minus .7 power, then multiply by the discharge path length. Multiply the result by 13, then add 20 volts to this. If this is less than 200 volts, then I recommend 200 volts as a lower limit.

To the limited extent in my experience that I actually got meaningful measurements, flashtubes seem to be more efficient at the higher voltages in the range I recommend above. I generally recommend at least 240 volts even if the tube should work at 200.

If the flashtube is small and short and (according to the formulas above) wants the flat rate voltages, it is probably filled with xenon to a higher pressure than usual. In my experience, most xenon flashtubes with discharge path lengths at least 1.25 inches (32 mm.) long and near or over 10 times the tubing inside diameter have similar xenon pressures not too far from 80 torr.

The rock bottom minimum voltage in volts that I recommend for reliable flashing, with some loss of efficiency: Raise the inside diameter to the minus .6 power, then multiply by the discharge path length. Multiply the result by 8, then add 20 volts to this. If this is less than 160 volts, then I recommend 160 volts as a rock bottom lower limit. Depending on the purity and pressure of the xenon in the flashtube, you may reliably achieve flashing at a much lower voltage.

If the flashtube is tightly coiled or against or extremely close to a reflector, I suggest reducing the upper limit by 10 percent (but not below 300 volts) and reducing the lower limit by 5 percent (but not below 200 volts). You don't need to do this for a straight flashtube that is close to a reflector as long as the distance from the reflector to the outer tube surface is near or greater than the discharge diameter. Example of what does not need to be derated: A straight flashtube in a typical parabolic-cylinder reflector.

My minimum recommended energy level: 4.5 millijoules per cubic millimeter of discharge or 300 millijoules per millimeter of *external tubing diameter*, whichever is greater.

My maximum recommended energy level for cheaper glass flashtubes having external diameter 3 to 6 mm: .5 joule per millimeter of discharge length. I have seen .67 joule per mm of discharge length usually succeed, and .75 joule per mm of discharge length to crack a few flashtubes in just a few flashes. I am aware of some flashtubes claimed to survive as much as 1.1 joule per millimeter of discharge path length, and I have personally experienced such or similar flashtubes cracking with .75 or more joule of discharge path length over a range of tubing external diameters from 3.15 to 6 mm. .67 joule per millimeter of discharge length is an aggressive level that I have seen in "disposable" cameras.

My maximum recommended energy level for flashtubes made of quartz or "vicor" or similarly tough materials such that thermal shock problems are minimized: .1 joule per square millimeter of discharge surface in most cases, even higher if the cathode is a nice big piece of solid tungsten. I know of flashtubes that can take .22 joule per square millimeter, even one for which I have been told a figure amounting to .4 joule per square millimeter.
A guide based on cathode dimensions:
A cylindrical solid cathode around 3.5 mm. long by 4.5 mm. in diameter permits a flash energy in joules roughly equal to 60 percent of my maximum recommended voltage in volts, maybe more.
A cylindrical solid cathode around 5 mm. long by 5 mm. in diameter permits a flash energy in joules roughly equal to my maximum recommended voltage in volts, maybe more.
If the voltage is lower than the average of my recommended minimum and maximum values, I recommend reducing the energy proportionately. This typically results in reducing the maximum recommended energy by about 20-22 percent at my minimum recommended voltage.

If the flash energy will be near the minimum that I recommend above, I recommend higher voltages near the maximum as determined above. Going 10 percent higher in voltage should not hurt if the flash energy is less than 1.6 times the minimum I recommend above. You can probably go 20 percent higher in voltage if the flash energy is below the minimum I recommend above. Please note that the spectrum may vary slightly for higher voltages and lower energy; this may cause photographs to be a bit bluish or greenish.

Maximum average power input: I recommend multiplying the discharge length by the external tubing circumference, and multiplying by 6 milliwatts per square millimeter of this. You can increase this to 9 mW/mm^2 if you don't need really long life expectancy. You can double this if the flashtube is cooled by lots of forced air. Quartz tubes can easily take 50 mW per square millimeter, 100 with forced air cooling.

If you can determine/predict peak flashtube glass temperatures as a function of average input power (and everything else such as flashtube dimensions/cooling), you may want to consider this:
Cheaper glass flashtubes largely don't like any part of the glass to get hotter than about 200 Celsius. Quartz doesn't seem to mind temperatures up to 600 Celsius (400 if things are touchy), although the limit is often lower in the ends of the tube at and near where the quartz is sealed around metal. The maximum temperature here is uncertain, but usually around 350 Celsius but it is preferred to not exceed 200 Celsius. A few tubes have a solder seal with more extreme temperature limits, but this extreme case does not include tubes where it looks like a wire is simply sorrounded with fused quartz.

Please consider that the electrodes may have limited average current capacity. If you are pushing a glass flashtube and relying on some major cooling, please also consider that the tube may fail if the inner and outer surfaces of the glass have too great a temperature difference between each other. This largely limits high-power use of glass flashtubes.


If you only have a loop or two of thin wire around the middle of the flashtube, then you will need lots of voltage. Perhaps something like 1,500 volts plus 6 times my maximum recommended voltage (As determined for "normal use", not using higher limits determined for lower flash energy).

If there is good trigger coverage over the flashtube, then 1,500 volts plus 2.5 times my maximum recommended voltage usually works. I generally find flashtubes can be reliably triggered by 1,000 volts plus 1.5 times my maximum recommended voltage.

UPDATE 2/7/99: I have seen some unusually untriggerable tubes that need more, amounting to requiring at least 4 KV and preferable 5 KV.

To get good triggering with minimal voltages, I recommend that the trigger circuit provide a positive pulse with a risetime no more than a couple of microseconds and an available peak current at least in the tens of milliamps. I also recommend that the trigger electrode exposes most of the length of the discharge path to the high voltage pulse. If you don't have this, you may need a higher trigger voltage or higher main voltage or both to get reliable triggering. Your mileage may vary; reliable triggering often occurs if you fall short of the above trigger guidelines.

Go here for some trigger coil info.

Go here for a strobe trigger circuit.


Generally, to use a flashtube in any manner described above, you don't need one.

There are a few cases where an inductor is used. The effect is a slightly more constant light output during the flash, and the flash light output dropping off more rapidly. With a more constant current through the flashtube, one might hope for higher efficiency, although I think the efficiency gain is slight and unlikely to make up for losses in the inductor.

With an inductor, the current rises less suddenly (sometimes this results in less mechanical shock), and the peak power dissipated into the flashtube is less (less shortwave UV impact on the tube material). An inductor also permits a higher voltage for the energy storage capacitors, ideally 72 percent more for the same peak current. This would permit reducing the capacitance nearly by two-thirds and the flash duration even more.

Another effect of using an inductor, occaisionally advantageous, is a more complete discharge of the energy storage capacitor before the discharge extinguishes. In a few cases, this can prevent having the discharge fail to extinguish and keep on arcing as the storage capacitor is recharged.

If you are going to try this, I would use a voltage around 1.25 times the maximum recommended value determined above. Choose a quantity of energy, then determine the capacitance you will need for this.

You will probably want to use an oscilloscope to display the storage capacitor and/or flashtube voltages during a flash. If the voltage dips below zero (reverses polarity), you need less inductance. If the discharge dies with the voltage staying well above zero, you want more inductance.

It may be tricky getting the scope to work properly and not be interfered with by the trigger pulse or stray induced voltages.

It may also be quite a challenge to get or make an inductor that works satisfactorily and has low losses and does not make much noise or vibration nor saturates with the high peak current involved.

For a technical and theoretical determination of the inductor value, look just below in the laser pump flashtube section. The inductor stuff works (to whatever extent necessary) for all flashtubes.

EG&G ELECTRO-OPTICS DIVISION LASER PUMP FLASHTUBES and use of inductors with flashtubes of known impedance characteristic

These flashtubes usually have a high xenon pressure of 450 Torr. These flashtubes are intended to be operated with series inductors. Flashtubes have a significantly nonlinear resistance, described with a resistance characteristic they refer to as "K", having units of ohms-amps^.5. The resistance in ohms is K divided by the square root of the current in amps. This is only valid at higher currents and at higher energy levels. The resistance will be higher than K/sqr(I) during the beginning of the flash, since it takes a few millijoules per cubic millimeter to warm up the xenon fill to approaching a typical operating temperature of about 5500 Celsius or so.

The inside diameter ("bore") is 2 mm. less than the tubing's outside diameter for any of these tubes with outside diameter of 4 to 15 mm.

Xenon flashtubes in general, and especially these, have a "K" value of:

1.28 times the arc length in millimeters, divided by the tubing inside diameter in millimeters. In the event the xenon pressure is known but of a pressure other than 450 Torr, then multiply K by
(pressure/450 Torr)^.2  to correct it for pressure. That is no misprint - K is proportional to pressure raised to the 1/5 power! So if you have only a rough idea what the pressure is you can get a fairly good idea of what K is.

The usual design procedure is to pick an energy level and a flash duration. For a flash duration from .01 to 10 milliseconds, the maximum energy the flashtube can withstand is 3.5 joules, multiplied by the arc length in millimeters, multiplied by the inside diameter in millimeters, multiplied by the square root of the flash duration in milliseconds. The maximum safe energy is less than the above indicates for flash duration longer than 10 milliseconds and may also be less for flash duration under .01 millisecond. It is generally recommended by EG&G to not exceed 30 percent of this in order to get really long flashtube life and a good chance that the tube will not eventually fail catastrophically. These tubes are known to explode.

With the energy and desired flash duration known, the capacitance *in farads* needed is:

(.09102*E*t^2*K^-4)^(1/3) (from the EG&G flashtube catalog but simplified.)
(I have a computer model which calls for 15 percent more capacitance than this.)

Where E is the energy in joules, K is the impedance characteristic, and t being the approx. time *in seconds* at which the flashtube current is in excess of 1/3 of its peak value (the nominal flash duration).

The voltage (in volts) required is 1.414 * sqr(E/C).

The recommended series inductor is:

2.44 * C * K^4 / V^2

(From EG&G's catalog, reworked to be in terms of C, V, and K.)
(My computer model recommends 16.5 percent less.)

where V is the initial storage capacitor voltage if there is no resistance besides that of the flashtube. You will need more inductance to account for extra resistance and K being higher during the first few millijoules per cubic millimeter of the discharge.

If C is in farads, inductance is in henries. You can use microfarads with this formula, and get the inductance in microhenries.

If you know the resistance of the inductor, wiring, and the capacitor's internal resistance, then the appropriate inductor according to my computer model is:

2.04 * C * ( A*R + (K^2 /V))^2.

A is about .9 to .95 if R is small compared to K^2/V, and about .8 to .85 if R is about the same as K^2/V. Of course, you should verify proper actual operation before finalizing the inductor.

Since these flashtubes have a higher xenon pressure than usual, they require higher trigger voltages. EG&G recommends 15 KV for flashtubes with 450 Torr pressure and arc length no more than 100 mm and inside diameter no more than 4 mm, 20 KV for longer arc lengths up to 230 mm and larger tubing inside diameter up to 7 mm.

I believe somewhat lower trigger voltages will usually work for external triggering, something like 2 KV plus about a kilovolt per centimeter of length plus external diameter if this is lower than EG&G's recommendation.

Maximum average power: .125 watt times the arc length in mm, times the inside diameter in mm. for convection cooling. Multiply whatever you get by 3.75 if the flashtube is the version optimized for convection cooling. For really substantial forced-air cooling, multiply by 7.5 instead of 3.75 regardless of the flashtube model.

EG&G recommends a minimum voltage of 300 volts plus 100 volts per inch of arc length (3.93 volts per millimeter).

These flashtubes will work without inductors, but with a substantial reduction in maximum safe flash energy. They will work with my original guidelines above, except that the voltages have to be increased by approx. 30-50 percent and the minimum energy for efficient flashing must be at least doubled due to the higher-than-usual xenon pressure.


Use a quartz flashtube and Good Luck! Homebrew flashlamp-pumped dye lasers are quite difficult to get to work. Ruby and neodymium (or YAG) lasers are much easier to get to lase. The laser threshold is tremendous. In addition, you get "triplet absorption" which refers to dye molecules being excited to their "triplet" state, in which they absorb laser radiation slightly, usually enough to impair lasing. The laser cavity typically gets enough dye in the triplet state to impair lasing in something like a microsecond of typical flashlamp pumping.

Now, how does one try this? I recommend a straight quartz flashtube. Refer to the section above for laser pump flashtubes in general, but keep these guidelines in mind:

You want just enough energy for efficient flashing, maybe 15-40 millijoules per cubic millimeter, even less if you know the xenon fill pressure is well below 450 Torr. Maybe do some experimenting and use just enough energy to have the discharge fill up the flashtube in a nearly uniform manner. Pick a flash duration as short as practical but for which the amount of energy you are using is safe. Use the laser pump flash guidelines to pick a capacitance, voltage, and inductance. You may not need a coil - the wires going from the capacitor to the flashtube might have enough inductance.

Another idea: Pick a "nominal" flash duration of a few microseconds to 10 microseconds or so, and a flash energy no more than 20 percent of the maximum energy from the formula for determining this from flash duration. Determine a capacitor from this nominal flash duration and this energy. The flashtube can probably survive this with no inductor, so try to minimize the inductance of any wiring, etc. Minimizing inductance will increase output during that critical first microsecond.

If the capacitance is low enough for xenon ion spectral lines to be significant, this helps with some more popular laser dyes. Rhodamines work well from mid-green and fluorescein (in alkaline solution or in the sodium salt form known as "uranine") works well from blue-green. The xenon ion spectrum has strong line clusters in these regions of the spectrum. I show this spectrum as the xenon one in my Spectrum Page.

If you use more energy than a flashtube can withstand for a given very short flash duration, you will vaporize traces of the inner surface of the flashtube. This phenomenon is known as "wall ablation". The vaporized quartz will decompose in the arc and release oxygen, which will impair flashing. If this happens, you may be able to "rejuvenate" the flashtube by forcing a steady current through it (high voltage at 100's of mA is required) to make the electrodes get glowing hot. The hot electrodes will oxidize and remove most of the oxygen. The flashtube won't be quite as good as new, but this "repair" usually works.


If the flash energy is low, the xenon may not get hot enough (roughly 5500 degrees C or so) to become a good graybody radiator. Or, a large fraction of the flash energy may be used just to get the xenon this hot. As a result, efficiency is less with lower flash energy.

To maximize the efficiency of a flashtube with energy levels near or below about 10 millijoules per cubic millimeter of arc volume, I recommend a different set of guidelines. These are largely not valid for most laser pumping flashtubes nor others that are both large and having higher xenon pressure.

These guidelines generally result in the spectrum being rich in xenon ion lines.

If the flash energy is around 3 to 7 millijoules per cubic millimeter, I recommend a storage capacitor voltage around 8 to 14 volts per millimeter of discharge path length, plus 20 volts. But this is only if the distance between electrodes is at least 32 mm or this distance is at least 20 mm AND the tubing inside diameter is at least 3 mm.

For tubes with distance between electrodes 13 to 32 mm and tubing inside diameter under 3 mm, use 350 to 550 volts for flash energy 5 to 12 millijoules per cubic millimeter.

For lower flash energy, use slightly higher voltage. Don't be afraid to use over 20 volts per millimeter or over 800 volts when the flash energy is only a millijoule or two per cubic millimeter.

Two tubes that I know of with only 10 mm between electrodes are special strobe tubes that are unusually intolerant of excessive voltage. Do not exceed 360 volts with 15 millijoules per cubic millimeter. Maximum voltage will be slightly higher at much lower flash energy. I do not recommend higher flash energy.

Please note that higher voltages are only recommended for straight and U-shaped flashtubes. Ring-shaped ones probably tolerate somewhat higher voltages. Coiled ones should not have high voltages since high peak currents result in much higher electromagnetic forces that may cause mechanical shock to the tube. Coiled tubes should have voltages near the minimum for these low energy guidelines.

Please note that the xenon ion spectrum is low on deep-red light. Human eyes don't notice this much, but photographic film is more sensitive to longer red wavelengths and less sensitive to shorter red wavelengths than the human eye is. (More correctly, film is nearly equally sensitive to all visible red wavelengths and the human eye is more sensitive to shorter red wavelengths, less sensitive to longer red wavelengths.) The result is that photographic film will notice that the xenon ion spectrum is low on red even if the human eye does not notice this much. The result is that color photos taken with xenon ion light will look blue-greenish (or more likely green-bluish since even as seen by the eye, flashtubes operated to produce xenon ion light generally look a bit bluish).

Lower value capacitors should be especially conductive types! Vishay/Sprague TVA series and similar axial lead electrolytics are beter than most other aluminum electrolytics for this. For smaller capacitance still, good are "glassmike" types, extended foil types and types rated for use in pulse forming networks. Good series include Vishay/Sprague 715P, 131P, 735P and Cornell Dubilier WMF and WWMA series.


To produce a short flash, you need a small capacitor. To get lots of light, you need to charge it to a high voltage, probably approaching the maximum voltage at which the flashtube reliably does not flash without triggering.

Smaller camera flash tubes with diameters near 3 mm. and discharge path lengths of 18 to 32 mm, as well as smaller U-shaped flashtubes, should work with voltages around 1000-1600 volts or so. Do not exceed 45 volts per millimeter. The energy level should be around a half joule for small, straight tubes and about a joule for U-shaped tubes with a 40-50 mm. arc length and 6 mm. or so tubing outside diameter.

If you need a higher voltage but the tube self-flashes, and it has any conductive paint on it for triggering, then you can carefully remove the trigger electrode and scrub off all conductive paint with steel wool and/or a single-edge razor blade. After this, simply wrap a loop of fine wire around the center of the flashtube (maybe slightly closer to the cathode than to the anode) for a less sensitive trigger electrode. With the high main voltage, the tube should still trigger easily.

If you successfully use a higher voltage, use a smaller capacitance to keep the amount of flash energy down to levels recommended above. This may get you even shorter flashes.

Be sure to use very conductive capacitors such as axial lead foil types or "glassmike" types or a parallel bank of ceramic discs. Good series of capacitors include Vishay/Sprague 715P, 131P, 735P and Cornell Dubilier WMF and MMWA. If you put capacitors in series for higher voltage, put resistors (1 to 4.7 megohms) in parallel with the capacitors to prevent any unequal leakage from overcharging one of the capacitors.

If you have an oscilloscope and can determine the flash current waveform, you can adjust the series inductance for best results. The flashtube, its leads, and wires connecting the flashtube to the energy storage capacitance may have enough inductance for this. To increase inductance, use longer leads or spread them further apart. To decrease inductance, use shorter leads, bunch them closer together, or use thicker wire.

Please note that the energy level is likely to be low, and you may only be able to illuminate your bullets at shorter distances. This may not be so bad since bullets are rather small. Please note that the flash duration is likely to be around a microsecond to maybe a few microseconds, which may make bullets look a bit smeared since they generally travel at speeds of a goodly fraction of a millimeter per microsecond. Please note that with the very high voltage and significant ion spectrum content, the flash will make color photos look bluish or green-bluish.

WARNING At higher voltages of around a kilovolt or two, you don't have to touch things to get shocked. This high voltage can spark through a fraction of a millimeter to sometimes as much as a millimeter or two of air.

Back up to my Xenon Flash Page.

Back up to my home page.

Written by Don Klipstein.

Please read my Copyright and authorship info.

Please read my Disclaimer.