Gas molecules inside a GM tube become ionized at the passage of a subatomic particle. The ionized gas lets electrons, pushed by the high voltage on the tube, pass from the wall of the GM tube to the center conductor. This makes the GM tube act like a switch at the passage of a cosmic ray. While the tube remains ionized, it's unable to detect other radiation events. How quickly the GM tube clears out this ionization is called the tube's dead time. The shorter the dead time, the more frequently the tube can detect radiation. In near space, I have detected up to 800 counts per minute. On average then, there is 75 milliseconds between detections. As long as the GM tube's dead time is less than that, it should accurately detect radiation levels in near space.
I'd like to try placing a capacitor across the GM tube to smooth out the voltage spikes. If that works, then the flight computer doesn't have to divide the number of counts by six to get the real radiation levels. Perhaps it will also let the tube clear out faster (reducing its dead time).
It's a short video about my experiments to date. Look for an article in Nuts and Volts this year.