Strobe theory -- xenon vs. LED

[TCJ]

I've never used an "LED strobe" worthy of entertainment/theatrical production before, but I'm curious how they compare with xenon, and the general opinions people have of each.


I'll be straight upfront with my initial assumption: I'm skeptical. Reason is, xenon technology seems pretty darn ideal when you consider the physics behind it. And there are a lot of "high-intensity LED" lamps out there (not necessarily strobes) that have serious longevity shortcomings.

The problems I see with both technologies:
(Don't take this the wrong way -- I'm not a pessimist; I just look for what can be improved. If inventors/engineers/innovators/scientists don't look for the shortcomings in various technologies and how to tackle them, technology would never improve!)

Xenon:
-Although lamp life could, theoretically, be infinite, lamps "wear out" from 3 factors:
-Gas diffusion (either xenon leaking out, or nitrogen & oxygen leaking in, or both?) through imperfect seals at the electrodes (or through the glass itself over thousands of years; faster if the glass develops small cracks).
-Electrodes vapourizing and depositing metal on the inside of the glass tube, darkening the tube a little.
-Electrodes vapourizing the coatings that improve electron emissivity (barium, lanthanum, etc.), resulting in lower flash energy for relatively low voltage drive circuits. (Though the strobe light I built charges up to 800V across the tube -- much larger than the typical 350 ~ 375, so this isn't really an issue on that light!)

LED:
-Although I suspect traditional LEDs in the mcd intensity range will last longer than humans do, "high intensity" LEDs do not.
-High intensity LEDs / overdriven LEDs develop small cracks in the semiconductor crystal lattice, which increases its effective resistance function.
-On a simple linear resistor-ballasted power supply, this results in decreased brightness
-On a constant-current regulated power supply, the forward-bias voltage slowly increases to compensate, causing increased power transfer into the LED for maintaining the same brightness... therefore increasing the temperature, and causing cracks to develop faster and faster, accelerating the LED's approaching end-of-life.


The peak intensity of a xenon flash tube is very high -- much higher than what any LED cluster of comparable size can continuously emit. Therefore an LED strobe must "overdrive" the LEDs, relying on the fact that the pulse is short to ensure the LEDs don't immediately self-destruct. But that raises the question of what is the specific heat capacity of the semiconductor die, and how effectively can the heat be conducted away, so even during the short pulse it does not cause significant cracking/fracturing in the lattice.

I also question: Why? What is the purpose of attempting to use LEDs for emitting a short pulse of white light? The only issue with xenon is the electrode decay... and possibly reduced heat, if that's even feasible.

Xenon's plasma emission spectrum is very close to a "daylight" colour temperature already. And while I don't have any data on its visible light efficacy (ie. "efficiency", if radiated heat and non-visible emissions are to be considered a "waste"), I don't believe it's all that bad. (It's better than mercury-vapour UV emissions striking fluorescent phosphors, ie. traditional "fluorescent lamp", and certainly better than incandescent!)

Considering that to get brilliant enough flashes out of poor little overdriven LEDs would need A LOT of thermal conduction, I'm not sure LED ends up being significantly better than xenon plasma arc, in the "efficiency" department. And when I see moderate intensity LED floor lights "burn out", I really can't see those overdriven blasted "strobe LEDs" faring much longer.

Some day I'll try to dig up the numbers and compare the quantitative facts. But for now, those are my thoughts.

What are yours?

 

[TCJ]

Wow, that was long. I should have split that into two posts

 

[icewolf08]

Just consider that an LED unit has far more versatility than xenon strobes. An LED unit does not require capacitors which means you could likely achieve faster strobe rates and there is less danger to the technician servicing the unit. Also, caps are usually the most common point of failure on a xenon strobe.

Now most manufacturers of theatrical strobes are not kind enough to give useful output data on their units for comparison. A standard set of strobe measurements like watt/seconds, t.1, and t.5 data would be nice. That said, I would bet that an LED array of equivalent output to say an Atomic 3000 would consume considerably less power.

Consider also that with an LED strobe, managing strobe duration is far simpler than with xenon. With an LED unit you could have an infinite duration whereas xenon strobe tubes are not designed for prolonged discharge. It means you could have a strobe that doubles as a work light or other continuous source for the production. Not a bad bonus feature.

 

[DuckJordan]

TCJ When is the last time you had experience with the LED strobes. I've seen several shows now that are incorporating them (rock concerts) and for the use they are perfect. Less power, Lighter, Sturdier. And just as bright and the ones I've seen that I really like and am looking to add to my rig is the 9 cell LED fixtures. The effect you can do when you have individual control of a 9 cell Strobe but takes the same space as a single atomic 3k? Sign me up.

 

[PeteEngel]

What I've found to be the case with Xenon strobes is their unfortunate tendency to thermal after a relatively short time. Your control over strobe duration and timing over a large rig of fixtures can depend upon various factors not always in your control. I agree that some LED strobes are not worthy of the name, but there are a couple out there that more than fit the bill. I have had some time on the Clay Paky Stormy (I work at ACT Lighting). The specs are impressive and being able to use it as either a strobe or an extended used audience blinder (and work light) makes it a multiple use item. I used the Stormy CC on a demo as a truss warmer and it looked great. There are others out there and you need to really examine each on their own merits since some LED Strobes are only flashing LED's as opposed to being designed to actually LOOK like a Xenon strobe. Research and demo's are your friends so don't just take my word for it. Here is the product page. If you haven't already read it over, check it out. I can still see some after images burned in my eyes from LDI.

http://www.claypaky.it/en/products/stormy

 

[GreyWyvern]

My thoughts are that you have never seen an LED strobe unit. If you had, you likely wouldn't be asking the question. You would just go with LED and forget about Xenon strobes, which are quickly becoming a thing of the past. I saw the Martin Atomic 3000 LED at LDI this past weekend, and it was impressive. While I have not seen the Clay Paky Stormy (or Stormy CC) other than at LDI, I have had hands on time with TMB's Solaris Flare. While it takes a different approach, it is still just as impressive as the others. It lit a 30' x 50' area of our warehouse better than the 6 mercury-vapor lights in the area. The things that LEDs can do, such as color changing, put them ahead of xenon.

 

[JD]

An LED unit does not require capacitors which means you could likely achieve faster strobe rates and there is less danger to the technician servicing the unit. Also, caps are usually the most common point of failure on a xenon strobe.

Actually, most of the high power non-photographic strobes don't use capacitors either, and instead fire the lamp directly across the AC mains. On one of these threads, I posted the schematic of the atomic strobe. Same was true with the old Super-strobes. Output is regulated by how late the lamp is fired in the AC cycle.
Professional photographic do all use big caps as they have to be able to vary the flash in a more critical way. This is done by charge voltage and the use of a shunt-snubber (basically a SCR fired across the capacitor to cut the length of the flash short.)

 

[TCJ]

Regarding high voltage and safety:
--------------------------------
The strobe light I built operates with a peak electrode voltage of 800V. It has not killed anyone.
Microwave ovens are ubiquitous in pretty much every modern kitchen... which is a wet environment too! These have magnetrons operating at around 2 kV, and it has not been a problem.
And for decades, people had CRT TVs in their home, with anode voltages at 25 kV or more. No one ever batted an eye.
Most commercially-available xenon strobe lights I've seen and for which I've had the privilege to get some technical data or see the schematics, typically have electrode voltages around 400V or less.
So I really don't think the voltage in these things is a problem. Just don't be an idiot.


Regarding capacitor reliability and longevity:
--------------------------------
Electrolytics don't fare well in this department, but they're also not exactly great for strobe applications. (Unless compact space is an issue. Then yeah, you're kind-of forced to go electrolytic.)

When I built my strobe light, I chose to use oil capacitors. These are great for high voltages, with very little chance of arcing / flashover failure. They last forever, and they have much better discharge rates than electrolytic. (The internal resistance and "equivalent resistance" or "current limit" is basically insignificant.) This is EXACTLY what you want designing a xenon flash system -- high voltage, and rapid discharge.

(The only drawback to oil capacitors is they are enormous. But I addressed that in my strobe light by building a separate "capacitor box", linked to the lamp with a special braided HV cable rated to 4 kV. This allows the light to be mounted to a truss or stand or whatever else, and easily aimed, without the bulk of the capacitors getting in the way.)

So with some creative design, a xenon strobe can be built with capacitors that will last indefinitely.


Regarding strobe rates:
--------------------------------
In a 60 Hz country, I can get up to at least 40 flashes per second on my strobe light. (Three half-cycles to get the capacitors up to a charge level ready to strike an arc.) But I suspect it could achieve a 60 Hz strobe rate (1 cycle to charge) on another controller (or if I retune mine)... and maybe, just maybe, 120 Hz (1 half-cycle to charge)... though I never tried tuning my strobe's control oscillator to go that fast.
Regardless, 40 Hz is waaaaay faster than I've ever found to be practical anyways.

So all that being said, xenon technology is definitely NOT too limited for practical strobe rates!

(I also recall reading something, somewhere, way back when I working on this strobe light design, that some epilepsy foundation was pressuring manufacturers to limit the strobe rates of their products to no higher than 15 Hz. So if one wanted to heed that suggestion, then it becomes the limiting factor far before the technology's inherent limits.)


Regarding forced "cool down":
--------------------------------
I have fractured tubes in my light from prolonged continuous running at high strobe rates, with the unit set to max brightness. But that "thermal issue" exists with LED technology too. Only difference is instead of cracking quartz glass, you're cracking some kind of doped silicon semiconductor. (I never bothered putting a thermistor and associated control circuit in my strobe light to force the operator to take a break if the tube gets too hot.)

So when commercial strobes go into "cool down" mode for a while, that tells me they just don't have the necessary cooling required to run continuously and still prevent the glass tube from cracking.
I've heard of xenon tubes running at ultra-fast strobe rates, continuously, for pumping large industrial/scientific lasers. These enormous tubes are liquid-cooled and run continuously. So that tells me it is possible to build a xenon strobe that can run continuously without damage. (There's no universal law of physics against it.)

So if LED suffers similar thermal issues and needs an exotic cooling system to prevent destruction, then either commercial LED strobes could end up going into a "cool down" mode too, or if a commercial manufacturer designs a light with sufficient cooling to never need it... then it should be within their capability to design a xenon strobe light with sufficient cooling too.


Regarding power consumption:
--------------------------------
I had a talk with a physicist today regarding the light efficiencies. He explained the electron band-gap phenomenon that occurs in LEDs, and how a gas plasma emission from an arc is due to heat from the electrons colliding with the gas molecules. He said the efficiency of LEDs is better... but he did agree that their semiconductor crystal lattice develops fractures... especially in high-intensity LEDs... causing them to fade over time and then eventually die outright.

Which brings me to my next point...


Regarding longevity:
--------------------------------
I still feel any attempt to build an LED strobe, with appreciable peak intensity, will have a much lower lifespan than a well-built xenon tube.

And after looking at the Clay Paky Stormy, I couldn't help but facepalm and think "Why on god's green earth are they using a limited number of LEDs, all scrunched together in a line where they'll cook themselves to death, when they could just do away with the reflector and use far more LEDs spread out evenly across the entire front, each running at a safer / more-reliable / lower peak intensity and with the potential for better cooling!??". They're just asking for the semiconductor lattice to develop fractures.

And I'll bet my balls... not just the left one, but both of them... that replacement LED modules for the Stormy will cost more than xenon tubes too.

 

[TCJ]

Actually, most of the high power non-photographic strobes don't use capacitors either, and instead fire the lamp directly across the AC mains. On one of these threads, I posted the schematic of the atomic strobe. Same was true with the old Super-strobes. Output is regulated by how late the lamp is fired in the AC cycle.
Professional photographic do all use big caps as they have to be able to vary the flash in a more critical way. This is done by charge voltage and the use of a shunt-snubber (basically a SCR fired across the capacitor to cut the length of the flash short.)

Maybe in 240V (RMS) countries, where the peak voltage can just barely reach 340V, the xenon tube can be placed directly on the AC mains. But I've yet to see a flash tube that'll strike an arc at just 170V across the electrodes (peak voltage for 120V RMS countries.)

Also, that must be terrible for the power lines and electromagnetic interference! Imagine the transients! I don't know what the peak flash current is, but it's enormous... and very short. That must create RFI hell in the building! And although I've never played around with AFCBs yet (arc-fault circuit breakers), I'd imagine such a strobe design would wreak havoc on them and cause nuisance tripping... pretty much always. After all, the xenon plasma flash is literally, an arc.

 

[TCJ]

My thoughts are that you have never seen an LED strobe unit. If you had, you likely wouldn't be asking the question. You would just go with LED and forget about Xenon strobes, which are quickly becoming a thing of the past. I saw the Martin Atomic 3000 LED at LDI this past weekend, and it was impressive. While I have not seen the Clay Paky Stormy (or Stormy CC) other than at LDI, I have had hands on time with TMB's Solaris Flare. While it takes a different approach, it is still just as impressive as the others. It lit a 30' x 50' area of our warehouse better than the 6 mercury-vapor lights in the area. The things that LEDs can do, such as color changing, put them ahead of xenon.

I've seen LED flood lights / work lights, and LED retrofits for incandescent light bulbs. I've seen them "burn out". Which is crap, because an LED should last longer than a human does. Now compare the peak intensity of a continuously-illuminated work light (which is not much), to that of a decent strobe light (which is enormous). Tell me, how long do you expect those strobe LEDs to live up to being abused 15 times per second? (Especially when they're clustered in a little bar, instead of spread out and affixed to a big heat sink substrate of some sort?)

Also, since LEDs begin to fail with a slowly-increasing equivalent resistance, then depending on the drive circuit, this could likely manifest itself as a slowly fading LED. Suppose an LED strobe that fired 100 000 flashes over its life is paired with one fresh out of the box. Do you think they'll look the same?
I don't. (Of course, I'm guessing here... so no testicular wagers... but unless I see some data to prove otherwise, I think my skepticism is justified.)

Now take a well-built xenon flash tube after 100 000 flashes, and compare it with one fresh out of the box. Do you think you'll notice anything different in their flash intensity or colour?
Unless the electrode seals are crap, or the glass cracked, and some xenon diffused out or nitrogen or oxygen diffused in, I doubt you'll be able to tell the difference in the resulting output.

 

[PeteEngel]

So, LED's don't last longer than Humans. 50,000 hours is 5.7 years of 24/7 operation. I've tried to find the lamp life of the MAX-7 Xenon lamp for the Atomic 3000 but I have yet to read a document that will list that.
LED's do fail due to heat. Not sure where you are trying to walk with your line of questioning here. Since LED strobes and LED fixtures in general are built to different standards and you seem to be lumping everything into one cauldron, I guess you are not Actually looking for an answer. Do some research. For me, the unit I've played with has 144 7w LED's. The math behind this says that means 1008w of LED's. max ouput of the fixture is 980w, and continuous operation is 720w. True, you can overdrive LED's and you can get failure but how do you rate a LED life if you are only using instantaneous and non-overdriven life measured in milliseconds? I would be willing to guess the other LED strobe manufacturers will have similar data available.
Also, Xenon lamps (using my followspot experience) tend to be sketchy in nature, prone to failure due to mishandling, high pressure explosions if there is an envelope failure, and unstable arc due to deterioration of the anode and cathode.
Now, does that apply to the xenon strobe lamp? Not sure but a quick search does not find any basic lamp info that can be found on any other lamp or LED type. What are the hiding? Having been near xenon followspot lamps when they fail isn't pleasant.
as for pairing different aged units, we are really just at the beginning of finding out what LED strobes can do and how they can be effectively used. Thermal management is important for any fixture.
Your use of negative LED verbage ("Also, since LEDs begin to fail with a slowly-increasing equivalent resistance, then depending on the drive circuit, this could likely manifest itself as a slowly fading LED. Suppose an LED strobe that fired 100 000 flashes over its life is paired with one fresh out of the box. Do you think they'll look the same?") vs positive xenon verbage (Now take a well-built xenon flash tube after 100 000 flashes, and compare it with one fresh out of the box. Do you think you'll notice anything different in their flash intensity or colour?") tends to make me think you have an ulterior motive in your line of questioning. What up with that? You mention you have seen a number of LED units, but don't mention actually seeing Theatrical/Entertainment LED units (professional, not crappy). Get your hands on some gear. I'm not saying you will come away with an opinion other than what you already have, but at least you can speak from experience.

 

[JD]

Maybe in 240V (RMS) countries, where the peak voltage can just barely reach 340V, the xenon tube can be placed directly on the AC mains. But I've yet to see a flash tube that'll strike an arc at just 170V across the electrodes (peak voltage for 120V RMS countries.)

Actually, they will fire with as little as 40 volts across them as long as they get the 4kv trigger pulse. It's all in the design of the lamp. 120 volts for most of the old Diversitronics and a lot of the Atomics (although they do have a 240 volt model that uses a different lamp. Sample link for Atomic replacement lamp:
http://www.fullcompass.com/prod/103097-Martin-Professional-97010308

Here is the schematic for the strobe:
http://www.controlbooth.com/threads/martin-atomic-3000.6978/page-3#post-303432

 

[MikeJ]

Elation, yes...elation makes a LED strobe, the Protron 3k or something, that outperforms the atomic 3000(208v lamp) in all aspects, especially "stab"(blinder) function. It also patches as the same fixture profile as an Atomic. It looks like a frosted glass rectangle though, not the really bright line down the center that the atomic produces. The Clay Paky stormy emulates the xenon lamp well, bit I have only used the CC version, and while in deep colors they may do better than an atomic with a scroller, they are not very bright as a white strobe.

Also, the idea that manufactures are "over-driving" LEDs seems to be common among people who have a predisposition to not liking LEDs, and rarely founded in facts or research.

 

[MikeJ]

and unstable arc due to deterioration of the anode and cathode.
Now, does that apply to the xenon strobe lamp?

While flash tubes for Atomics seem to last a LOOOONG time, this is how I have had them fail. They start flashing erratically an both ends of the tube will be glowing for much longer than normal. They at first look like someone has set them to the wrong mode or address. It seems to get bad pretty quickly once it stats.