Unfortunately this is normal.
Same as any similar design low pressure gas discharge tube, evaporated
electrode material traps gas on the tube walls and output gradually
declines, but 100 hours is a ridiculously short lifetime. The lamp is
essentially a neon sign folded into a regular grid instead of a sign and
filled with an alternate gas mix, powered by an ordinary neon sign
transformer, The far superior mercury lamp monochromatic illuminators
are made the same way, and I know of some of them that ran almost 40
hours per week for nearly a decade without any problematic loss of
output; about what you would expect from a good neon sign.
Its a lapmaster CP-1 intended for use with optical flats up to twelve
inches .
LAP5-0010-004-0161
I have one of those that I got at an auction, only to recall after
getting it home that a former employer once described it as a nearly
unusable POS, speculating that a competitor sold it to their customers to
prevent them from inspecting their parts, while lamenting the loss of the
company which manufactured his large yellow-green mercury sources, which
were large enough for his 14" flats and filtered to a single mercury
line. I forget the exact wavelength, but when viewed from a practical
~20 degrees from perpendicular you saw 10 microinches per band to as
close as you could read it (~1/20 band). It produced far more distinct
bands than you will ever see with the lapmaster due to the the very low
contrast you get from it's unfiltered output. Also, if you use the base
of the lapmaster to support your work (as it is designed to be used) the
heat from the transformer can cause significant workpiece and flat
distortion. You can remove the lamp from the base to avoid this problem.
Last time I looked Edmund was selling smaller filtered mercury line
inspection illuminators, in a proper stand leaning ~20 degrees forward of
vertical, for less than $1000 IIRC.
For those not familiar with this method of inspection, you are looking at
the reflection of a diffuse source from the entire surface of a polished
work surface under the flat interfering with the reflection from the
flat, so the source needs to be significantly larger than the flat and
coherent over the length of the gap between flat and work (<.001") but
preferably not coherent between the work and the top of the flat (~ 1" or
more). After gently setting one side of the flat on the work and then
lowering the other side onto the air film you can let go of the flat and
wait for the air film to squeeze down to around a 10 band wedge, where
band curvature (flatness) is most easily read.
If I knew where to get the large mercury line filters I would get one and
have a neon sign maker refill the tube in my lapmaster with mercury. But
I wonder if you could do as well today by illuminating a sheet of ground
glass or an LCD backlight diffuser with laser diodes at lower cost.