-----Original Message-----
 From: cctalk-bounces at 
classiccmp.org [mailto:cctalk-
 bounces at 
classiccmp.org] On Behalf Of Josh Dersch
 Sent: Wednesday, October 30, 2013 8:49 PM
 To: On-Topic and Off-Topic Posts
 Subject: Re: Random UNIVAC item...
 On 10/29/2013 1:07 PM, Brent Hilpert wrote:
  On 2013 Oct 28, at 7:01 PM, Josh Dersch wrote:
  Grabbed this as it triggered both my
"interesting numerical display
 device" and "old computer" reflexes:
 
http://yahozna.dyndns.org/scratch/misc/univac.jpg
 It's labeled on the back (roughly, as things don't line up
 particularly well):
 Magtrol, Inc.
 DYNAMOMETER         HDE-500
   MODEL             HD-705
   SERIAL NO         511B341
 TACHOMETER          811E342
   MODEL
   SERIAL NO
 And riveted to the back is a placard reading "SER 2U02-2"
 Inside it's completely tube-based, no solid state electronics at all.
 Not counting the Dekatrons, there are 9 tubes in this thing, one
 branded "UNIVAC" on the base, the rest appear to be RCA.  It looks
 like it contains an electronic tachometer of some sort.  I can't seem
 to find a date code on anything.  Any ideas on what this might have
 been used for, and how it would relate to UNIVAC?
 I can provide more pictures later if anyone's interested, once I get
 a chance to disassemble it a bit... 
 Very cute. Between 'dynamometer', 'torque adjust', and 'Magtrol'
I was
 going to guess something to do with calibrating the mechanics of the
 tape drives.
 The dekatrons are likely the display for the tachometer function.
 Photomultiplier tubes were sometimes used as the tachometer sensor for
 such counters.
 I can provide schematics of a couple of dekatron counters to give an
 idea of how dekatrons were used in-circuit, if of interest.
 
 Sure, that'd be very interesting to look at.  When I get some time I want to
 figure out how this thing works...
 Thanks,
 Josh
  
For those who aren't familiar with dekatrons: the folks at TNMOC were explaining to me
how they work in computer circuits.  The dekatron's state transitions with an input
pulse, taking one of ten states.  One cannot directly (electronically) read a
dekatron's current state (although it's obvious by eye!).  So in order to read the
state, one cycles the tube until it emits its 'index' pulse, meaning it is now
back to its base state.  Subtract the number of pulses needed to get it to that state from
ten and you know where it was!  Of course, if you want to retain that state, you must now
cycle it with that number of pulses, to return it to where you found it.  A similar idea
to the read-restore cycle of a core memory unit, just with more states....  -- Ian