3 Sep
2009
3 Sep
'09
4:12 p.m.
On 3 Sep 2009 at 20:41, Tony Duell wrote:
I have (both the HP one and clones), but I don't
own one. To be
honest, it _looks_ cool, but it's fairly useless. The ones I used had
no pulse-stretching circuitry, the LEDs just displayed the state of
each pin (on for '1', off for '0' and on-but-dim for toggling -- the
brightness depended on the mark/space ratio). It was OK on slow
circuity, but that's easy to debug anyhow. On anything with a
reasonable clock frequency, it didn't tell you much.
I can see that without pulse stretching or even a "pulse" as opposed
to "steady state" indication that the utility would be marginal.
HP made a couple of other devices that I don't own
and could never see
the point of. One was the 'signature analyser' which will tell you if
a logic signal differes from the correct version (well, provided you
have a 'correct version' to also test) but won't tell you _how_ it
differs. So actually finding the fault doesn't seem to be any easier.
I remember reading about signature analysis in the HP Journal. It
looked like a promising technology--not for the lone self-employed
technician, but for the field engineer who saw the same equipment day
in and day out.
It's a method of automating the diagnostic procedure; in a sense, a
predecessor to modern boundary-scan techniques.
--Chuck