11/23 clock issue
jnc at mercury.lcs.mit.edu
Fri Feb 6 07:42:18 CST 2015
> From: Brent Hilpert
>> Well, try removing W1 to isolate the output of the xtal oscillator E15
>> from inverter input pin E4.4.
Sigh, the crystal output is dead as a doornail. Total flat-line. Guess I'm
going to have to find a new one...
I actually think there's an open in there, because the input to the buffer
inverter stays at +2V, whether the crystal is connected or not. If there was
some sort of internal short in the crystal taking its output to ground, I'd
have expected it to pull the inverter's input down.
> TP4 looks to be an 'input' test point, rather than an output. The
> pull-down resistor value (150 or 180 ohms) has been selected such that
> it is low enough to allow enough current to flow through the tri-state
> control input E4.1 to pull it low normally, but high enough to allow
> one to connect TP4 high, to enter tri-state
Ah, got it. (I tended to assume test points were outputs, but I need to
remember that they may be inputs.)
> What the point of going to tri-state is, is not clear, considering that
> W1 is there for an external clock. A conjecture is there may have been
> some external test fixture that tri-stated it for some sort of
> synchronous single-step clocking.
Yeah, but couldn't they have lifted the W1 jumper, and fed their test clock in
that way? Eh, not important.
> Could also check the V reading on the open E4.4 input.
See above - +2V is a floating TTL input, ISTR?
> From: Holm Tiffe
> Hmm, may be since his hints are standard debugging technique and you
> aren't really familiar with debugging??
I cheerfully admit to being primarily a software person. But I have been
debugging broken hardware off and on for 30+ years - although not as a
principal occupation, of course. I think it's more just that my mind does not
do hardware intuitively (the more-so, the further one gets from the ideal -
aka digital at the design level - to real hardware) - I have to think about
> You don't ned no pullup for +5. All open TTL inputs are reading High
> w/o any pullup.
That's why all those boards use pullups on unused inputs that need to be 1,
right? :-) But you're probably correct for a quick test.
More information about the cctech