Stable DC Power Supply Design for Osram Hg 100 (45V 1A) Spectral Lamp for Spectroscopy

Hi to all!

I'm looking for a power supply to feed a high-pressure mercury vapor spectral lamp
(Osram Hg 100, 45V DC, 1A). I found many different circuit schemes over the web, however
the stability of them is unknown. I'd like to build a DC power supply which can ignite
the specified lamp and able to provide stable 1 A current, because the lamp will be applied
for spectroscopy, where the fluctuations in the light intensity is not acceptable.

Any help will be highly appreciated!
Thank you!

Osram Hg 100: http://www.osram.com/osram_com/products/lamps...troscopic-lamps/index.jsp?productId=ZMP_56381

OK, I have to ask, how stable, and over what sort of period. How does the lamp intensity relate to power supply and temperature, what sort of figures are we talking about. Power supplies can easily be made stable to parts per thousand. I would have thought that if the light stays relatively constant over the period of a measurement, that you would get the relative intensities of the various spectral components if you were simultaneously measuring the lamp intensity as well to get a baseline reading. Some maths would then correct for intensity changes. Absolute light output is not a trivial thing to measure.
In theory, if it is critical, you could use something like an integrating sphere to measure the light output and use it in a feedback loop to stabilize the lamp.
Don't know much about the subject, so further explanation might be helpful.
Cheers,
Richard

Hello Richard!

Basically I need +-5% emission stability over 8 hours. The lamp intensity is decreasing
slowly with time (50% over 8000 h), however it is not problem to us, because the
instrument where we like to apply the lamp is a polarimeter, were the 546 nm emission line have been used to carry out the measurement and the device is calibrating itself at the beginning of each measurement. What I need is a ballast for the lamp (the lamp is working as well with AC power supply, at 100 Hz). A few examples were attached in a pdf file. The question is: can we build a ballast that stand to the requirements, and if we can, then how? Power supplies for the specified lamp can be purchased from 500 €, so we like to stay under 150 € if it's possible, because the lamp is cost 500 € as well and if we destroy the lamp with a bad circuit, then it's not worth it.

Thanks for your help!

Interesting pdf. If you have little or no electronics background, then designing the type of circuits shown in the pdf is not trivial. If a commercial power supply is available, and Osram can confirm that it is suitable, then I would seriously consider going with it. Trying to do a custom supply will end up costing you far more than the commercial unit.
Cheers,
Richard

I guess you are right, so we going to purchase a commercial power supply.
Thanks for your help again!
Cheers,

Hi ---
Would you be able to suggest/advise on the commercial power supply you chose? Oddly enough, I am looking for the same thing and your message popped up during my search.
Thanks,
Eileen