Hey there!

I'm diving into a fun project and trying to design a digital alarm clock. I've been scouting the web for inspiration, hoping to find a simple design that I can tweak a bit. The thing is, a lot of the projects out there use transformers, Arduinos, or dive deep into programming with microcontrollers like PICs.

I'm aiming for something a bit different. Ideally, I'd love a purely DC-powered design WITHOUT all those mentioned "transformers, arduinos, programming, microcontrollers. Quick to build, but still a solid DIY project, you know? Oh, and a reset button would be super handy!

Do you think it's possible to create something like this? If you have any ideas on the components or the schematic design, I'd really appreciate your insights!

Thanks in advance!

 

A very common project a few decades ago in a first digital course was to build a clock (an alarm clock is quite a bit more complex). Since programmable logic is a later course, these were done using SSI and perhaps MSI chips (look up those terms to see what is meant).

This is a significant undertaking if you aren't going to use programmable logic. You will learn a lot, but you will spend a LOT of time wiring things and debugging your wiring of those things -- and that's after you've gotten all the kinks worked out in your design itself.

 

Of course, I apologize for the confusion earlier. What I intended to convey was my experience with code programming. I'm open to exploring projects involving logic gates, but I might steer clear of code programming since it's not my strongest suit. Thank you for understanding and guiding me in the right direction!

 

Welcome to AAC!

I'm open to exploring projects involving logic gates

It's going to involve a lot of logic gates and counters.

Implementing a 4 digit clock will require 4 counters for keeping time, 4 BCD to seven segment decoders, 4 7-segment LEDs. You could use a single 4 digit multiplexed LED display, but you add the complexity of multiplexing the digits (more logic gates).

You'll need a crystal oscillator for the time base, but using AC voltage would give you a more accurate (long term) timing signal.

Implementing set/reset logic will require yet more gates.

Implementing an alarm clock would require more multiplexers, storage, etc.

That adds up to a lot of area and wiring.

Here's an example of a 24-hour clock using 74LS162.


If you wanted to go more basic than counter chips, you'll need 4 flip flops (and a bunch of logic) for the 2 decade (0-9) counters. You'll need a mod6 counter, which will take 3 flip flops, and a flip flop for hours (12 hour clock) or 2 for a 24 hour (and more logic to decode 23).

On the other hand. Implementing my single digit scrolling LED (matrix) clock that also displays the date on a Raspberry PI Zero W took 1 CD4017 counter (for multiplexing the columns of LEDs), an OR gate for a column reset, 12 transistors, 35 LEDs, and a bunch of resistors; and some code. I don't have to worry about set/reset logic or code because the program gets time from the operating system and the operating system gets it from a time server when it boots (and I suspect periodically to keep time correct).

 

Thank you so much for shedding light on this! Your insights are incredibly valuable. The challenge of extensive wiring and space is daunting but I'm up for the challenge. It's a tad frustrating that there's limited material online about digital alarm clocks without relying on embedded systems or transformers. But I'm holding onto hope that there's a feasible way to achieve this without code programming or transformers, if possible. I'm also leaning towards using a 4-digit LCD, but my current stumbling block is kickstarting the schematic design. Any advice or pointers would be greatly appreciated. Thanks again for your guidance!

 

But I'm holding onto hope that there's a feasible way to achieve this without code programming or transformers, if possible.

How low in technology to you want to go? Counters, flip flops, gates?

There's someone who took several years to make a pretty neat clock using only transistors, diodes, resistors, and LEDs. Can't remember what he did for the time base. Might have used AC to simplify because he'd only need an extra mod 60 counter to convert 60 Hz to seconds.

https://www.techno-logic-art.com/clock.htm

 

I'd love a purely DC-powered design

I'm also leaning towards using a 4-digit LCD

If by DC you mean battery powered, then you certainly have to go with an LCD display, unless you want to use a push-button to momentarily turn on an LED display when you want to read it.

 

Datasheet for MM5370 digital clock chip.

 

But I'm holding onto hope that there's a feasible way to achieve this without code programming or transformers, if possible.

The use of AC and a transformer is to (1), get a DC supply that isn't limited by the battery capacity, and (2) almost always to get the time base using the very stable, at least in the long term, in a very simple way.

If you want to be DC powered, then you need to design a time base circuit. For a clock, you need pretty good accuracy, and so you are looking at a crystal oscillator. So you are just trading one complexity for another.

It's going to be a lot of parts and a lot of wiring to make just a clock. To add in the alarm functionality, you are basically doubling the complexity because you need a separate set of registers to store the alarm time and then a comparison circuit to compare the alarm time to the current time, and a small state machine to keep track of whether the alarm is currently sounding or not based on the comparison and the user's action to silence it.

 

Thank you so much for the recommendation! I'll definitely dive into that link and see if it sparks any inspiration. I'm aiming to keep things as straightforward (“easy”) as possible

I actually graduated in electronics a while back, but veered off into other interests. I've kind of lost touch with what I learned about circuit designing. Now, I'm circling back to the world of electronics, realizing it might be a boon for my career. Thanks again for pointing me in a direction. Really appreciate it!

 

Absolutely! I'm actually leaning towards incorporating pushbuttons or even keypads. I think it would not only enhance the aesthetics but also make it much more userfriendly, especially when it comes to adjusting the time. Thanks for the suggestion!

 

If by DC you mean battery powered, then you certainly have to go with an LCD display, unless you want to use a push-button to momentarily turn on an LED display when you want to read it.

Absolutely! I'm actually leaning towards incorporating pushbuttons or even keypads. I think it would not only enhance the aesthetics but also make it much more userfriendly, especially when it comes to adjusting the time. Thanks for the suggestion!

 

Welcome to AAC!

It's going to involve a lot of logic gates and counters.

Implementing a 4 digit clock will require 4 counters for keeping time, 4 BCD to seven segment decoders, 4 7-segment LEDs. You could use a single 4 digit multiplexed LED display, but you add the complexity of multiplexing the digits (more logic gates).

You'll need a crystal oscillator for the time base, but using AC voltage would give you a more accurate (long term) timing signal.

Implementing set/reset logic will require yet more gates.

Implementing an alarm clock would require more multiplexers, storage, etc.

That adds up to a lot of area and wiring.

Here's an example of a 24-hour clock using 74LS162.
View attachment 305912

If you wanted to go more basic than counter chips, you'll need 4 flip flops (and a bunch of logic) for the 2 decade (0-9) counters. You'll need a mod6 counter, which will take 3 flip flops, and a flip flop for hours (12 hour clock) or 2 for a 24 hour (and more logic to decode 23).

On the other hand. Implementing my single digit scrolling LED (matrix) clock that also displays the date on a Raspberry PI Zero W took 1 CD4017 counter (for multiplexing the columns of LEDs), an OR gate for a column reset, 12 transistors, 35 LEDs, and a bunch of resistors; and some code. I don't have to worry about set/reset logic or code because the program gets time from the operating system and the operating system gets it from a time server when it boots (and I suspect periodically to keep time correct).

Absolutely right! I recently reached out to one of my former professors about this. They were aware that programming isn't my forte, so they challenged me to construct the project without delving into the realm of embedded systems. While I initially thought it would simplify things, I'm now seeing that avoiding programming might actually add layers of complexity. Nonetheless, I'm committed to the challenge and will avoid using any embedded systems or code. Thanks for the insights!

 

Certainly do what you wish to in keeping in line with your desires, motives, and goals. But one thing to consider is that the only way to become better at something that you aren't good at is to practice doing it. You might find that tackling this project via some kind of embedded logic (and Arduino seems pretty reasonable, but since I've never used one, I might be overlooking something) not only improves your coding ability, but opens up a whole world of projects that are simple to do via programmable logic but are monsters without it.

 

Another version of a transistor only clock: https://nutsvolts.texterity.com/nutsvolts/200907/MobilePagedReplica.action?pm=2&folio=42#pg42

 

Another version of a transistor only clock: https://nutsvolts.texterity.com/nutsvolts/200907/MobilePagedReplica.action?pm=2&folio=42#pg42

Thanks for sharing the link! I must admit, I don't distinctly remember delving into transistor clocks, though it might have been something I came across years ago. I truly appreciate the refresher. I'll definitely explore the information using the link you've provided. Thanks once again!

 

Digital clock using CMOS chips

 

they challenged me to construct the project without delving into the realm of embedded systems.

So how far back do you want to go? It sounds like you aren't interested in going back to the transistor level. Now your choices are gates only, flip flops with steering logic, counters with logic.

As a point of reference. I designed a hexadecimal to 7 segment decoder using logic gates. IIRC, it took 18 gates for each decoder, and you'll need one for each digit.

I started working on a version of that resistor, diode, transistor clock. Designed a discrete clocked R#S# flip flop, breadboarded a couple of the counters, and decided it was more work than I wanted to take on. It would have made for an impressive clock. It even fired up the artistic part of my brain and I worked on wire bending skills to sign my art.

You really should consider moving up to the last decade and learn how to program Arduino's and Raspberry Pi's. I didn't take the plunge until after I retired. But programming was never an obstacle for me. I took 2 programming classes while I was in high school and one required programming class (that was more rudimentary than the class I at the community college while I was in high school) for the EE program. I think it's much easier for an EE type to pick up programming than it is for a CS type to pick up electronics.

 

As a point of reference, here's a 6 digit clock I started using BCD counters over 40 years ago when I was a technician. I didn't need any 7 segment decoders because the display I chose to use contained decoder logic. If you could find those displays now, they'd probably cost more than $25 each.



I was doing it with wire wrapping. I couldn't imagine doing it with soldering point-to-point routing or laying out a PCB.

 

Avoiding complications, hundreds of flavors :
----> https://search.brave.com/images?q=automotive clock digital

Buy your preferred. Dismantle it to its components. Now you have all the parts to build it, adding features if preferred. Put it back together.