Re: Potential shutter speed problems

[P3D Michael Kersenbrock <michaelk@xxxxxxxxxxxxxxxxxx>]

> The device would basically be built around two light sensors (one for each 
> shutter), a clock chip, a counter and a few logic gates. The sensors would be 
> placed behind the shutter and the camera fired at a light source. The logic 
> gates would enable the counting of the clock signal as soon as the light sensor 
> switch tripped, and stop the count when the sensor turned off. The device 
> would then have three counters one for each shutter and one for the 
> differential.
> 
> If we ever get around to actually building something that works I'll post the
> schematic.
> 
> ...or better still if someone with more electronics knowledge (i.e.
> some :)) already has one, maybe they could post it..

This probably is the easiest approach, but if you're really concerned
about 'correctness', which your simultaneously-double-shutter tester
proposal would indicate, you might want to go a bit further and use an integrator
design.  If sensor linearities are manageable, this would be a much more
accurate method of measurement.  Implementing the timer-method yields
errors that include those talked about and nicely ascii-diagramed in the
one of the postings in the large-format camera usenet newsgroup thread
that someone mentioned a day or two ago. 

Circuit would be fairly simple -- The problem is that "calibration"
at run-time would be required for simple implementations (or so it
seems offhand).  Use of  one of those reasonably priced PIC-processor 
based boards with an A/D input and a digital timing input could make 
that task automatic.  Assuming one had a lot of time to burn.  :-)

Just a thought.  

IOW, measurements at the fastest speed or two with a timer-based
circuit might be taken with a grain of salt.  See that newsgroup 
thread for more details (summary: testers have their problems
and that many a shutter were ruined in attempt to fix them when
it was the shutter tester that had a "problem").

What I'm saying is that one might not want to take measurements overly 
seriously unless one can be sure the measurement method is valid.  For
slower speeds on most cameras, the timer method probably is valid.  To 
get an easily done and accurate measurement of the fastest shutter
speeds, one should probably use film-based tests using f/stops to 
compare the faster shutter speeds with measured slower shutter speeds 
(where tester accuracy would be better in most cameras).  This assumes
the f/stop settings have proper relationship to one another (absolute
accuracy not required).  This seems a reasonable assumption.  Does
anyone know what the usual accuracy of f/stop settings?  What I mean
is whether f/4 and f/6.3 have the proper relationship -- not if it's
actually f/3.8765 or f/4.12875 at the f/4 setting.

Of course, the effective shutter speed can be a function of f/stop
as well.... sigh.


Mike K.

P.S. - The "integrator" method is similar to what the film would
       do, so it'd take into account the changing brightness throughout
       the exposure-time -- then calculate an "equivalent" time as if
       the light "waveform" was a rectangular pulse.  If I understand
       correctly, the FED's shutter is the iris, and the iris starts
       closing the instant it gets to fully open -- so for any shutter
       speed (in auto-mode) the time the "shutter" spends completely 
       open is ~zero.  :-)  What a timer-based tester would read would
       be almost arbitrary.  :-(






------------------------------