In Reply to: umm, Schmitt what? posted by Mike on 08/09/02 at 9:39 AM:
There's a lot to answer here, I'll try and tackle the issues one by one:
1) Schmitt trigger. This is a circuit that is able to accept a slowly changing input voltage, and still deliver a clean, fast output transition when it's input threshold is crossed. If you've ever played with a light switch (like I did when I was a child), maybe you found the no matter how slowly you moved the lever, a spring inside would make it snap on or off quickly - to prevent excessive arcing across the contacts. A schmitt trigger is like this - the rapid output transition reduces the possibly of spurious oscillation arising in your circuit as the R-C network charges. More detailed explanations of the Schmitt trigger can be found in text books.
A 74HC14 can be used in place of a 74HC04 if the two 74HC04 gates are just used as logic inverters - which is what you're doing. BUT, if you were using the 74HC04 gates as part of a crystal oscillator, or some other analog function, swapping the '04 for a '14 would not work.
2) The 100R resistor. Okay, I understand what it is doing. But, if you use the schmitt trigger gate, and connect the counter reset to it's output (instead of to the 2n3904 emitter), you will not need the resistor - the 2n3904 will then only be turning the LEDs on and off.
3) Short pulse length. The 2n3904 base will be drawing a relatively large current via the 10uF capacitor, which makes the pulse length short. Using the schmitt trigger gate will fix this - in fact, you'll need to put the resistor (in parallel with the diode) back in! But you'll also find you can reduce the capacitor value and still get a longer reset period.
4) Unreliable counter reset with larger capacitors. This is the most puzzling problem of all - but what I guess is happening is this: during the reset pulse, as the 10uF capacitor is charging, the 2n3904 is bursting into high frequency oscillation, maybe @100MHz or so. An emitter follower running at several tens of mA with capacitance at the base (which is what you've got) is prone to do this, and possibly upset your counter. If you have a fast 'scope, you might see it - then again, you might not because applying a 'scope probe might be enough to stop it happening! However, using the schmitt trigger avoids this problem in 2 ways: a) the counter will be driven by the clean output from the 74HC14 gate (see (1) above), and b) you will no longer have capacitance at the 2n3904 base since it will also be driven by the 74HC14 gate via a series resistor @ 1K.
Just a reminder: if you do use the 74HC14, don't forget to turn the R-C network upside down, see my previous reply. If you want to send me the circuit, I'll be willing to look over it.