Polyphaser sent me a document on request. Salient points:
The only way to test the IS-50 or IS-B50 without expensive test equipment is by seeing if the VSWR
(return loss) has increased or remove it from the line and see if the radio performance is better. Because
the product is a DC block device, there is no way to test adequately with a multi-meter. A properly running
product will test the same as a bad one. Normal operation, with an ohm meter you will get a short
between pins and short between any pin and the housing (ground). The same results if the gas tube has
failed. 95% of the time a gas tube will fail open which is also its normal state.
[...]
Since lightning is a natural event with varying current waveforms and return strokes there is no way to
predict the gas tube lifetime. The current IS-50 series has been in production since 1993. It is a very
reliable product. The first indication of impending failure will be increased insertion loss and SWR. The
protector should be replaced immediately if insertion loss and SWR are greater than a straight through
coax barrel connector (female to female) as full failure may or may not pass RF.
[...]
The MTBF (Mean Time Between Failures) is also effected by where and how the protector is installed and
the natural variation in the intensity and number of return strokes per lightning event. The tower size,
placement, ground conductivity, and number of storms per year can also affect MTBF. The gas tube
manufacturer can provide raw data on gas tube failures, but in the real world, it’s relatively meaningless.
The good news is the gas tubes will probably withstand 5 hits of 20KA each, but the N connector center
pin will be damaged with the first 20KA hit. Good news? Yes, since it is obvious that the protector has
done its job and should be replaced. How often could this happen? Probably never. That much current
could only ride on the shield, and indicates a failure to properly ground the coax cable shield or a poor
tower ground connection.