3

I have wired up a simple circuit here to test the cmos one-shot multivibrator. The top half of the chip seems to work but the bottom half does not. I even replaced the chip with a new one and I get the same results.

R1 is 27K, R2 is 10K, and the rest of the resistors are 1K each. All caps are 22uF, and I'm powering the device with 5VDC.

When I press the key (raising pin 12 to +5V), LED1 functions by lighting up for a brief moment correctly.

When I release the key, I expect LED2 to do the same but for a different amount of time, but instead, it stays on and never shuts off.

How do I fix this?

enter image description here

2 Answers2

1

Your schematic does not show pin 16 of the 74HC4538 chip. This must be connected to the VCC rail for the chip to work.

Similarly the schematic does not show pin 8 of the 74HC4538. This must be connected to GND for the chip to work.

Also note that as the circuit is configured the top section is meant to trigger on the release of the switch. The bottom section is setup to trigger on the press of the switch.

It is possible that the top section does not operate correctly due to a long fall time on the signal to the B input. You could consider replacing the C3 capacitor with something more reasonable like a 0.1uF device. The large 22uF capacitor at this site will eventually cause pitting and destruction of the contact surfaces in the S1 push button switch.

Michael Karas
  • 56,889
  • 3
  • 70
  • 138
  • 1
    That's eagle that decides to not show those pins, but I connected pin 8 to ground and pin 16 to +5VDC when I tested my circuit. –  Jul 26 '16 at 06:11
  • Also, the reason why I used the capacitor in series with the key is because I wanted to prevent random key-spikes in the circuitry. Would some kind of input buffer help (even something as simple as an cmos inverter). Maybe that capacitor is messing things up, but I want the circuit as simple as possible. –  Jul 26 '16 at 06:18
  • An inverter or buffer with a schmitt trigger input between the button and the rest of the circuit should do it. That will give you some noise immunity on the input and a nice sharp edge for the rest of the circuit. – Andrew Jul 26 '16 at 08:44
  • @Mike: there **must** be a way in Eagle to connect the power lines of an IC (and show it). How else could you draw a schematic that can be turned into a working PCB circuit with Egale? – Curd Jul 26 '16 at 09:51
  • 1
    @Mike - Well whatever CAD package you choose to use it behooves you to use one that you can be 100% certain about your PWR and GND connections. At over 40 years in this electronics engineering thing I cannot tell you how many times I've seen circuit boards with missing/wrong power and ground connections because the schematic was not explicit. I can tell you I never use the default symbols in a CAD package that do not show PWR/GND connections on the symbols. As a matter of fact when I start a new design I generally create new symbols and footprints for every part I have not used before. – Michael Karas Jul 26 '16 at 11:59
  • @Mike - I know you want the circuit as simple as possible but if you cannot live with switch bounce then you really need to consider eliminating it will cost you some complexity. Note that these days it is common to consider a low cost MCU to be the simplest design because the switch can be debounced in software and you can custom tailor the LED drive according to what is needed for the application at hand. In days gone by when things like CD4538's were much more prevalent in designs it was common to always use switches with a Form-C contact (continued) – Michael Karas Jul 26 '16 at 12:10
  • (continued from above) arrangement and debounce said switch using a cross coupled NAND or NOR gate pair. Gate choice made by the use of pullups with grounded switch or pull downs with switch common to Vcc. – Michael Karas Jul 26 '16 at 12:12
  • @Curd In Eagle, with ICs that have multiple gates like this one, the power and ground connections are shown on the schematic separately. Clicking on one of the gates and selecting Invoke from the pop-up dialog will place power and and ground pins on the schematic which you can hook up to the positive rail and ground. – tcrosley Jul 26 '16 at 16:23
  • @tcrosley: I was sure that there has to be some way. So the OP is wrong when he says "That's eagle that decides to not show those pins". – Curd Jul 26 '16 at 20:49
  • @Curd Eagle won't show the pins unless you explicitly use the Invoke command for ICs with multiple gates. Failure to do so (and I'm sure there are legions of designers that either forget or don't initially know about the feature) *will* result in the power pins not getting connected. However this should show up in an ERC (Electrical Rule Check) run. If they don't show in the schematic, they won't be on the PCB. – tcrosley Jul 26 '16 at 20:54
  • Well I ended up showing those pins when "invoking" the IC. –  Jul 27 '16 at 02:47
0

The bottom half is wired incorrectly for falling-edge trigger. When your switch is closed, the input is +5V and when you release the switch, the input will fall "slowly" from +5 down to ground. Thus, the input is a negative-going signal. Thus you want the bottom to trigger from a "falling edge". Looking at the datasheet, you'd want the "falling edge" to go into the "B" input with the "A" input grounded. That is the opposite of your bottom part. ALSO the "A" input is Schmitt-triggered but the "B" input is NOT. Hence you should use the "B" input for the "rising triggered" (UPPER) part and the A input for the falling. But you can't because your switch generates fast-rise (when pressed) but slow-fall (when opened). You need the RISE to be slow (so you can use the A input) and the FALL to be FAST (so you can use the B input). Change the input cap to head towards ground instead of V+. The resistor will now go to V+. And rewire the A/B inputs accordingly for both sections.

Atomique
  • 396
  • 1
  • 4
  • At least for the TI version, both the A and B inputs are Schmitt triggered. https://www.ti.com/lit/ds/symlink/cd74hc4538.pdf?HQS=TI-null-null-mousermode-df-pf-null-wwe&ts=1606431472341&ref_url=https%253A%252F%252Fwww.mouser.com%252F – WhatRoughBeast Nov 26 '20 at 22:59