Online RadAlert geiger counter

You should read this at least once (updated 31 Mar 95)

Here at World Power Systems' galactic headquarters is a small RadAlert, Inc, geiger counter. It generates a pulse upon each count of the geiger tube. A small interface I built with a Parallax Inc BASIC STAMP computer tallies the pulses, and is connected to via serial interface. (Unlike the toilet graphical interface, this is not a hoax.)

The geiger counter is located approx. 20 feet north of the NW corner of 16th and Mission Streets, in San Francisco. Sometime I'll figure out exact lat/long. The geiger counter window is pointing up, and at the moment is indoors on the 2nd of three floors. So there is zero alpha activity, and very little beta at the moment. I will at some point figure out how to mount this up on the roof and at least let beta particles in (thin window, etc).

Using unix's cron scheduler, Fido accumulates counts from the geiger counter every 10 minutes; at one minute past the hour a report is generated; these are available below.

The daily summary log, contains only the daily summaries of the 24-hour average of the counts in a given day; the data log that contains the data reduced to hourly/daily counts in human clock time; and the raw counter data log contains the data read from the geiger counter interface with a unix binary time stamp. (The time stamp should be more than adequate for any purposes; it is synced to Lawrence Livermore Labs' atom clock via NTP protocol twice daily.)

The interface though homemade was carefully designed to be reliable. Both the interface and the geiger counter are battery powered with line-power backup, and will keep counting for days even if Fido is powered off or rebooted. The interface carefully counts pulses, has a large count range (0 - 99,999,998) and properly checks for overflow. It may start to lose count pulses due to overrun at very high pulse rates -- what rate I do not know, but I know it will count OK at 60 pulses/second, or 3600 counts/minute. With typical background radiation at approx. 10 - 20 pulses/minute (0.01 - 0.02 milliroentgens/minute) it will at least keep accurate background radiation counts, which is what this was designed for.


I am still testing and double-checking the accuracy of the counts and tallies, cross-checking the accumulated counts against hand-calculated values from the RadAlert's built-in display. This is slow and tedious, and will probably take a few weeks (starting 24 Mar 95). This paragraph will disappear and if it looks like the data is suspicious, I'll delete it. Consider this data preliminary...