Counter not starting without touching

Thread Starter

Anjum9694

Joined Aug 28, 2021
7
241715802_875042009817930_7442579401193547569_n.jpg

Hi all. My decade counter made with 74HC73 flipflops doesn't start counting (or showing up on the display) until I touch my hand to the positive terminal on the breadboard. I assume I am giving it ground for the circuit to work. How can I make this circuit work without touching or giving external ground?
 

Ian0

Joined Aug 7, 2020
9,667
At a guess (and without any circuit diagrams, it can only be a guess) you have an input that is not connected.
ALL CMOS inputs must be connected to either V+ or ground to ensure reliable operation.
 

Thread Starter

Anjum9694

Joined Aug 28, 2021
7
At a guess (and without any circuit diagrams, it can only be a guess) you have an input that is not connected.
ALL CMOS inputs must be connected to either V+ or ground to ensure reliable operation.
Well here it is. That sounds like a good possibility.

Screenshot_1.jpg
 

zophas

Joined Jul 16, 2021
165
It would also be a good idea to put some decoupling capacitors on your power lines. Most people will recommend different values but I tend to go with 0.1µf and 100µf on each power strip.
 

DickCappels

Joined Aug 21, 2008
10,153
When I read your description of the problem bypass (decoupling) capacitors came to mind immediately. In addition to zophas's brilliant recommendation, I sugges a 0.01 or .001 uf ceramic capacitor between Vdd and Vss for each chip.

A long time ago a much more senior engineer than me looked at my circuit and said "Why are there so few decoupling capacitors?" I answered "Because I am trying to keep costs down." To which he replied "Our job is not to save the odd few cents on capacitors but to keep ourselves out of the manufacturing department." If you think you won't need this decoupling capacitor or that one, wait until you have everything working ok, then you can remove the capacitors and see what happens.
 

Thread Starter

Anjum9694

Joined Aug 28, 2021
7
It would also be a good idea to put some decoupling capacitors on your power lines. Most people will recommend different values but I tend to go with 0.1µf and 100µf on each power strip.
You guys are amazing. I just shorted blanking input of the 7 segment driver IC to the ground with a 3.3uF capacitor and the numbers started rolling. Now I don't have to touch it to work. Thanks! Learned something new.
 

olphart

Joined Sep 22, 2012
114
Yes to all above, can't have too many, only too few. Clocked devices (F/Fs, counters, etc) are notoriously flaky without sufficient decoupling.
 

Thread Starter

Anjum9694

Joined Aug 28, 2021
7
When I read your description of the problem bypass (decoupling) capacitors came to mind immediately. In addition to zophas's brilliant recommendation, I sugges a 0.01 or .001 uf ceramic capacitor between Vdd and Vss for each chip.

A long time ago a much more senior engineer than me looked at my circuit and said "Why are there so few decoupling capacitors?" I answered "Because I am trying to keep costs down." To which he replied "Our job is not to save the odd few cents on capacitors but to keep ourselves out of the manufacturing department." If you think you won't need this decoupling capacitor or that one, wait until you have everything working ok, then you can remove the capacitors and see what happens.
Yep. I searched google regarding this and found the perfect image to describe the situation. decoupling-capacitor.jpg
 

dl324

Joined Mar 30, 2015
16,844
My decade counter made with 74HC73 flipflops
Your second picture shows you mixing HC CMOS with LS/TTL. Guaranteed TTL HIGH outputs aren't guaranteed to be high enough to drive CMOS inputs. To guarantee sufficiently high output voltages, you need to use pull-up resistors on the TTL outputs driving CMOS inputs. I've found that what you're doing usually works, but it isn't guaranteed if you get an atypical part. CMOS outputs can drive TTL inputs, but fanout is low.

Driving HCT inputs with TTL is guaranteed to work because the threshold voltages of the HCT CMOS inputs are designed to be compatible with TTL outputs. HCT outputs are CMOS levels.

More judicious component placement would let you breadboard your circuit in half the space (one board). Put the LS08 between the flip flops and your power supply and put the decoder between the flip flops and the timer.

What is the purpose for the tape/plastic wrap? Is the wire you're using 22AWG? It looks like 24 and I've found that to be just a little too skinny for reliable connections. I've used 24, but am moving away from it because it pulls out too easily.
 

Thread Starter

Anjum9694

Joined Aug 28, 2021
7
Your second picture shows you mixing HC CMOS with LS/TTL. Guaranteed TTL HIGH outputs aren't guaranteed to be high enough to drive CMOS inputs. To guarantee sufficiently high output voltages, you need to use pull-up resistors on the TTL outputs driving CMOS inputs. I've found that what you're doing usually works, but it isn't guaranteed if you get an atypical part. CMOS outputs can drive TTL inputs, but fanout is low.

Driving HCT inputs with TTL is guaranteed to work because the threshold voltages of the HCT CMOS inputs are designed to be compatible with TTL outputs. HCT outputs are CMOS levels.
Thanks for the tip about not mixing the ICs. I didn't think it would be a big deal because well they both do the same thing. I should be reading datasheets beyond the pinouts.

More judicious component placement would let you breadboard your circuit in half the space (one board). Put the LS08 between the flip flops and your power supply and put the decoder between the flip flops and the timer.
This is my first ever breadboard circuit so didn't focus on optimizing IC placement but then wouldn't you still need the second board for the display and the 220ohm resistors.

What is the purpose for the tape/plastic wrap? Is the wire you're using 22AWG? It looks like 24 and I've found that to be just a little too skinny for reliable connections. I've used 24, but am moving away from it because it pulls out too easily.
It's 21 gauge solid wire stripped from cat5 cable. I initially wired the board with stranded 23 gauge and it kept moving as you mentioned hence the plastic wrap. The circuit works but thanks I will remember the tips from this thread next time and do better.
 

dl324

Joined Mar 30, 2015
16,844
This is my first ever breadboard circuit so didn't focus on optimizing IC placement but then wouldn't you still need the second board for the display and the 220ohm resistors
If you remove as much space as possible between the IC's and don't take so much space for resistors, it should fit on one board.
It's 21 gauge solid wire stripped from cat5 cable. I initially wired the board with stranded 23 gauge and it kept moving as you mentioned hence the plastic wrap.

You could use 1/4W resistors on the display.
AFAIK, CAT5 cable uses #24 wire.
1631470239686.png

Stranded wire isn't good for solderless breadboards. If the wire moves so easily, you could get unreliable connections. Solderless breadboards were designed for solid #22 wire. The blue jumper you're holding is around the same gauge as #22 wire. If it's square, it's probably 0.025" square to make it compatible with female header pins and solderless breadboards.
 

DickCappels

Joined Aug 21, 2008
10,153
Just to inject my own 2¢, solderless breadboards are not very useful for fast or high frequency things, this is particularly true for circuit with edge sensitive like latches, shift registers and counters to name a few.
 

Ian0

Joined Aug 7, 2020
9,667
I just shorted blanking input of the 7 segment driver IC to the ground with a 3.3uF capacitor and the numbers started rolling. Now I don't have to touch it to work.
Not sure that's going to work long term. There might well be enough leakage current to charge it up. Try a resistor.
 
Top