[Date Prev][Date Next] [Chronological] [Thread] [Top]

Experience in South Carolina

R. Pickett experienced a problem in a South Carolina precinct on a Tuesday morning election (May 4, 1999).  It consisted of ballot being dropped from an Accu-Vote loaded with 1.94w (the latest firmware level).  We have tried to reproduce it here at McKinney with our own unit and ballots, but have been successful.  R. Pickett will be sending that unit, some sample ballots, and a memory card to McKinney so we can try to reproduce the problem with that particular Accu-Vote unit.
 
The following is the sequence of events as stated by R. Pickett who witnessed the event (he was standing next to the machine):
  1. The Accu-Vote had already counted four (4) ballots.  The poll counter on the Accu-Vote stated that as well.
  2. A voter (big guy) attempted to insert his ballot skewed (right corner leading).
  3. The reader motor started but the reader could not take the ballot at that angle.
  4. The LCD read "BALLOT NOT READ/PLEASE RE-INSERT"
  5. The message remained after the ballot was cleared of the reader.  (It is unknown as to whether the reader motor stopped after the ballot was cleared).  R. Pickett should be able to fill in the info.
  6. With the message still  displaying, the ballot was fed into the reader again but this time without skew.
  7. The ballot went through the reader and into the ballot box.  (It is unknown as to whether the ballot stopped at the exit and ejected in a regular fashion).  R. Pickett should be able to fill in the info.
  8. The LCD poll counter still read four (4).
  9. R. Pickett checked the voter registration table for a count of voters.  It stated there were five (5) voters so far.
  10. R. Pickett opened the ballot box and found five ballots inside.
  11. R. Pickett was able to determine which ballot was the last one fed and he re-fed that ballot through the reader.
  12. The LCD poll counter now read five (5).
 
I don't know if the memory card/Accu-Vote was set in the feeder mode.  R. Pickett may have that info.
The memory cards were set up to reject on undervoted ballots.  Further testing with the sample ballots and memory card from that election may result in a bug report.
 
 
 
POINT:    According to Pickett, this did happen.  A ballot made it from the front of the Accu-Vote to the Ballot Box without counting.  We do not know the criteria for the failure as yet.
 
Perhaps others can watch for this or test for this and report your findings (whether positive or negative) to the Support List.
 
Ian