how do bad contacts affect (falling) edges

When I set up a simple low-speed combinatorial pulse detecting circuit
using HC-logic on a breadboard I noticed that especially the falling
edges were horrible and caused glitches. The edges had some sort of
oscillation on them (unlike ringing) and I supposed they were due to
bad contacts of the breadboard but whatever I tried I couldn't get rid
of them until I finally soldered it all together.

Now my questions: _how_ do bad contacts do this? And why only the
falling edges?

Thanks,
Jenalee K.

 

Hello Jenalee,

As Jack hinted typical breadboards such as the white "plug-and-go"
boards or whatever they are called aren't good enough anymore for modern
logic. Take a look at VCC, maybe you'll see ringing there as well.

Anyway, a photo and a schematic would help. I really don't think it's
bad contacts. Then you should be seeing changes when you tug on the wires.

 

Breadboards -- the conventional variety -- are death over a few hundred kHz.

 

Joerg a écrit :

Thanks Joerg for your reply. I will try to post pictures and schematics
tomorrow.

Thanks,
Jenalee K.

 

Have you placed a decoupling capacitor across each IC (100nF)? It should be
right on top of the IC's, not at the end of the vertical tracks.

Meindert

 

Meindert Sprang a écrit :

C1
4n7
|| ___ |\ pulse with
pulse ------||---o---|___|---o--| >O- weird edges
15Vtt || | R1 | |/
| 4k7 | 74AC14
.-. |
| |R2 --- C2
| |10k --- 4n7
'-' |
| |
| |
| |
=== ===
GND GND
(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)

Original breadboard circuit without C2. The output pulses had weird
oscillating rising edges. As Joerg proposed, closer inspection of the
power lines reveiled some 100MHz or so ringing on the 5V and GND. After
the addition of C2 the problem went away, the ringing became one period
of "slow" 20MHz-ish bouncing. I had placed decoupling caps (100n & 10u)
as close to the 74AC14 as possible but they didn't seem to have any
influence.

There were also clamping diodes just after C1 but they didn't influence
the behaviour so I left them out for clarity. The purpose of R1 is
protection for the schmitt-trigger and its value doesn't seem to have
any influence.

BTW, I made an error, it is on the rising edges only, not on the
falling edges.

Thanks,
Jenalee K.