Erratic PCB Behaviour

Thread Starter

Georacer

Joined Nov 25, 2009
5,182
Some months ago I started a project with a friend, but unfortunately we bit more than we could chew and now we 're over our heads.

The project consists of 30 or so 74XX(X) IC's (don't ask me why I chose 74 series, it wasn't my choice and was a prerequisite) and a 555.

We designed the whole circuit in Multisim, ran simulations, everything worked fine. So I ordered the dual layer PCB and soldered the components on it.

From that point, all hell broke loose. Erratic behaviour, different results from the same starting point, and so on. I even tried to remove the 555 and use and external manual clock, but the 555 circuitry (caps and resistors) kept on alternating the clock level even if I din't push the manual clock button.

One thing that I thought about (and hurts me to realise it now) is that I haven't used ANY (oh god, why?) debouncing capacitor. How serious a crime is this? Does it explain this kind of behaviour?

I could post a schematic in multisim file form or in an image (a large one though), but I doubt it will be of any help. If someone has a loooot of free time to try to understand what I have built, I would be happy to post a copy.

Thanks in advance!
 

bertus

Joined Apr 5, 2008
22,270
Hello,

Do you have decoupling capacitors at EACH of the chips accross the power lines?
These should be about 0.1 μF each.
This is to prevent pulses traveling on the powersupply lines.

Bertus
 

Thread Starter

Georacer

Joined Nov 25, 2009
5,182
Unfortunately I don't have ANY decoupling capacitor. However I used a Mastech workbench power supply. But I guess even the best power source can't absorb all the pulses from traveling through the lines.
 

tom66

Joined May 9, 2009
2,595
As well as the decoupling capacitors of 0.1µF each, you should also have a large ish electrolytic capacitor (10µF or so) every 10 chips, to compensate for the inductance of the PCB traces.
 

marshallf3

Joined Jul 26, 2010
2,358
Decoupling caps can be 0.01 uF and work just fine but, as mentioned, you need one at almost every logic IC or the switching transients will cause all sorts of problems. Not to worry though, I've done the same thing before and forgot a few. I keep 0.01 and 0.1 uF AXIAL 50V caps around, they're perfect to tack onto the back of an IC circuit between power and ground.
 

Thread Starter

Georacer

Joined Nov 25, 2009
5,182
Problem solved! Apart from some idiotic mistakes such as pins that punctured wire insulations causing shorts, hanging IC inputs and such, the board worked! I used 0.01μF ceramic capacitors on the most fast-switching IC's, wich were the counters. Even a meagly 20Hz needed decoupling, who knew! However, when I put caps in all of the counters, things went a bit crazy, but I kept the ones that gave me a result, and now it's fine. When I finish the hardware part of the project I might make a post of it.

Thanks guys!
 

windoze killa

Joined Feb 23, 2006
605
Even a meagly 20Hz needed decoupling, who knew!
Well, virtually everybody. They have been doing it for years. :D

It is not the speed that matters. Decoupling is used in some cases to stop ICs (even slow digital ones) from oscillating. It also keeps any stray signals generated by an IC from getting to other ICs.
 
Top