Okay folks, I ran the procedure once again, but this time I used tape and fixed the americium buttons along with the SBT11A. Had the tube mounted in a metal box with Aware Electronics HV/discriminator PCB. Running AW-RADW software. First, I did both buttons for over 8 minutes, then lifted the tape and flipped off one button for several more minutes (thanks to a phone call). Finally, I resecured the button and flipped off the other for 10 minutes.

I took an 8 minute average for each reading. Button A, button B and buttons A+B. Just copied the text numbers into Excell and used a formula in A2, B2 and C2, =average(d1:d2) etc.

This time, I put the =abs() function into Bryan's formula.

Averages were as follows:

Buttons A+B 241Am buttons = 68,660 CPM

Button A = 41,983 CPM

Button B = 28,275 CPM

Deadtime in Microseconds uses the formula:

=abs((((A2*B2)-SQRT((A2*B2)*(C2-A2)*(C2-B2)))/(A2*B2*C2))*60000000)

The result was a surprising 41 microseconds, far better than my previous result, and a pleasant surprise. 440 microseconds is the dead time for a 900 volt LND7313-1 pancake. The LND7317 is given as 80 microseconds, as is my Aware Electronics RM-80 using the 500 volt LND7313-3 GM tube.

So, not only is the SBT11A inexpensive and tiny, it is very efficient.

My window sill background readings with an SBT11A are averaging 15 to 16 CPM, while the larger LND7313-3 is averaging 30 CPM.