Overdriving UV lamps

 I have built a dozen or so UV lights over the past ten years.  I always copied regular commercial lamps when I selected ballasts and other components, matching lamp wattage and ballast ratting.  Over the years I have heard a lot of discussion on "overdriving" UV lamps to increase their power output.  At least two manufacturers have comments in their literature that the lamps (tubes or bulbs) on their lights are overdriven.  This has lead me to a bit o' confusion, and a number of questions about overdriving lamps and ballast selection in general.  Possibly someone who actually understands the subject could enlighten me.

As I understand it, overdriving a lamp involves simply using a ballast designed for a higher wattage lamp.  In exchange for this greater power output there is some sacrifice in the life of the lamp. So, my questions are;

By how much does going to a more powerful ballast increase lamp output?  I have put 15 watt MW tubes into a lamp with 25 watt ballasts.  I see only a minor change in output over the same lamp run with a 15 watt ballast.  I have never actually measured the difference.

Does overdriving a lamp really cause a significant loss in lamp life.  Does it reduce the number of on-off cycles that the lamp can handle? 

 Is there a practical limit on how much a lamp can be overdriven by? What happens if I drive an 18 inch tube with the ballast for say an 8 foot lamp ?

Is there a difference in output driving a lamp with a newer solid state ballast over an old style magnetic ballast? 

I recently overheard a comment that leads me to think that different size and type lamps, U or H shape lamps or straight tubes, run at different voltages.  Not the 110v or 120v that the ballast draws, but the ballast output voltage.  Is this true, and does it factor into selecting a ballast to overdrive a lamp?  Assuming that one owns a VOM, how do you measure this?

Is there a significant difference between ballast manufacturers, or is price the main deciding factor in selecting a ballast?

Since I am about to reballast a couple lamps to eliminate an annoying hum any assistance would be appreciated.

 Thanks,  Jan

ed.a's picture

Re: Overdriving UV lamps

Hi All - There's also an excellent Wikipedia page on Fluorescent lamp technology. Some gems of information are hidden there.

Note that you do get about +15% output by going from magnetic to high frequency electronic ballasts, but as Stefan wrote, the ballast factor is now lowered to reduce consumption and increase efficiency. Overdriving the lamps is not as efficient as multiple fixtures, if energy load is an issue.

Maybe I can shed "a little more light" here. A ballast is basically a current limiter. Magnetic ballasts did this inductively. Moderm switching ballasts do it at high frequency using a smaller inductor and capacitive coupling. The tube wattage generally goes up with the length because the voltage across the ioni ed tube increases, but longer tubes also run at higher current (perhaps due to quirks of simple magnetic design designs). For example, some ballasts might have enough ignition voltage to run two tubes in series, and are designed that way, but most are optimized for one tube type.

I wouldn't bother with a magnetic ballast for overdriving a lamp. If it has taps on a transformer core for multiple tubes, wiring them together is likely to short the winding sections.

Because electronic ballasts are capacitively coupled, wiring two taps together effectively doubles the capacitive coupling to the tube. It mystefied me at first why doubling the capacitance didn't double the power, with even more diminishishing returns ganging four-lamp ballasts. I think it's not ballast inefficiency, since the ballast doesn't draw double the power, but that the current is partly limited by a single inductor internally.

Overdriving simulates using a high output (HO or VHO) ballast with double the Power Factor (tube current), but is far less expensive. Those ballasts are designed to match more expensive HO and VHO bulbs. A tidbit I gathered a while back is that HO bulbs are very similar to regular bulbs; mainly they use less mercury fill to achieve the same optimal vapor pressure at higher operating temperatures. Another design tidbit I realized myself is that since the thermal operating point is the coldest part of the bulb envelope, because that's where the vapor condenses,  an aluminum reflector mounted in contact with the bulb will seriously lower the operating temperature, if you're overdriving a standard bulb. With the right heat sink, you might even do away with those noisy cooling fans.

Lamp life is usually related to on-off cycles, and long term depreciaton (measured in years of on-time). SW filters would probably solarize faster, but I'd only be concerned about the decreased life span for a daily museum display. (And if that's the application, better to keep electrical code conformance in mind.) The electrodes are bound to run hotter when overdriven, but my recollection from aquarium forums was that overdriven regular bulbs are good for years of daily use.

Which ballast to use? My guess is an instant-start design to avoid filament power wiring problems and get maximum strike voltage. A ballast designed to handle higher current T12 bulbs (narrower T8 and T5 tubes run lower current at higher voltage) might have better design margin. Choose the one with the highest input power consumption for a standard lamp (not a low power factor energy-saving ballast). For overdriving, those feaatures may be more important than the type of tube the ballast is designed for (You might get slightly more output from a 20W tube using a 25W ballast, but you could also get less - that power spec is more about the matching tube than the ballast itself, though it's good to know the ballast is rated for higher power.

The aquaria lighting nuts have probably figured out a good cheap ballast to use. I haven't looked around since I first heard about the fad. Really new ballasts might be too smart, shutting down because they sense there's a bulb problem.

I'd definitely like to see more powerful portable lamp designs avaliable, even if it's only in brief flashes. Low voltage multilamp DC ballasts aren't common. My portable design simply boosts the voltage to the ballast from 12V to 18V when I press the "Turbo" button. (I use a tough Iota ballast that can handle it. I burnt out a few cheap ones.)  A custom intermittent VHO ballast design is probably the way to go for a new portable lamp design. It seems like that would have broader (visible light) applications, too.


Re: Overdriving UV lamps

Hi Jan,

did you read this article here? reocities.com/Heartland/Pines/7557/overdrv1.html

Overdriving means that you simply run a higher current through the lamp, leading to more ionization and therefore more light output. On the drawback the electrodes wear out faster, the lamps gets hotter and so on.

If you want to fight hum, use an electronic ballast. Output is the same as with the old magnetic ones if both are rated the same. But input is lower with the electronic ones, you save energy.

I am overdriving my outdoor one with a higher rated ballast. In combination with an instant start you definetely shorten the lifetime.

Hope thats a start....




Re: Overdriving UV lamps


Thanks.  That is exactly what I was looking for.


Re: Overdriving UV lamps

Hi Jan - sorry for the late reply to your request for info; it's hard to have a lively forum conversation without notification of new messages in the forums. 

Ed answered the question basically but no one gave a recommendation on a brand of ballasts.  I use Fulham Workhorse ballasts in all my AC powered lamps.  (www.fulham.com).  They have a complete line of ballasts for driving most any wattage unit.  Most commonly I use the Workhorse 5 - it can drive up to 120W (4+ x 25W for example) and is a nice compact package.  I use WH7 for larger lamps (up to a total of 240W I think).

You can overdrive lamps simply using these ballasts.  They come with a yellow wire which is common to one side of all the lamps, and an individual red wire for the other side of each lamp.  Tying two of these red wires together to one lamp will overdrive the lamp (double the capacitive drive) and you will get about 40% increased drive.  When replacing an older style magnetic ballast you must connect both wires at each socket end to either the red or yellow wires.   You can even add a third and fourth red wire to drive the same lamp, but with significantly diminished return (really not worth it).  Lamp life is effected, but not dramatically in my experinece.  I've been doing this for years in both my aquarium hobby and mineral hobby.

Just replacing the old magnetic ballast will be a major imrovement - eliminating flicker (20khz drive instead of 60hz) and higher efficiency, and lots lighter weight.  There are several brands out there, but I know and trust Fulham (and they're priced right).

Overdriving DC powered lamps - I would not do this by driving a ballast at 18VDC.  Not a good thing to do.  Most are designed to operate up to 14VDC (the charging voltage of a car battery).  Beyond that and you run into heat sinking issues, or simply smoke issues - no matter how "good" the design.   A good way to overdrive lamps with a DC ballast is to add a second capacitor in parallel with the existing output cap.  These caps are usually around 1500pf, 500V - a 2nd cap will have the same effect as tying the red wires together in the Fulham ballast.  I do this in all my dual bulb 26W handheld lamps.

Fans - I highly recommend using a fan in all UV lights.  UVC bulbs prduce ozone (as we all know by the lovely smell from our lamps) even if they are spec'd as "non-ozone producing";  a little is always generated.  Ozone blocks UV (ask Al Gore).  It only makes sense to evacuate this ozone.  Convection cooling will not do this effectively.  Fan cooling will cycle the air inside your lamp, and keep your lamps operating at optimum temperatures (104f) .

Feel free to ask any questions off-line - markcole@minershop.com (or let me know you've asked something in this fourm)



Don's picture

Re: Overdriving UV lamps

Interesting discussion.  The 2007 Journal of the FMS Vol. 27 on page 31 has some information on "Over Driving a Lamp".

My guess is without electrical and radiometric measurements with actual lamps and ballasts it might be difficult to come to any conclusion.  Just observing a lamp and ballast combination with our eyes many not be adaquate.  Our eyes can easly be fooled.  I have access to all of the proper test equipment (clarke-hess model 2330 V-A-W meter, International Light ILT1700 radiometer, Photo Research PR-650 spectraradiometer, etc.), but I just do not have the time. 

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Potential Exploration Links