Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Slow Clocks

Status
Not open for further replies.
(yes this is a spin off from another thread).

I'd like to capture everyone's options here for slow clock generators. We already have Hackers Bench 1Hz generator and the MM5369 with a 3.579545 MHz color burst crystal. Anyone have any other options or links?
 
Last edited:
You would get relatively good accuracy using a pair of 4017 Decade Counter/Divider chips running off the 60 Hz. AC line frequency. 60 Hz / 6 = 10 Hz. Then 10 Hz / 10 = 1 Hz. The popular way to do this is start with a 6.3 volt secondary line transformer. 120 VAC 60 Hz to 6.3 VAC 60 Hz. Then wave shape using a simple operational amp (even an old 741 would work). That gives you a 60 Hz. square wave to divide down to 1 Hz. Any cheap transformer will work or low voltage source of 60 Hz. The uncertainty is as good as mains frequency. So that is yet another option in the quest.

<EDIT> For those who may be curious, things began in this thread the object being to develop what I see as a run time meter. </EDIT> :)

Ron
 
Last edited:
The first thing you should specify is the accuracy needed for the 1 pps over what temperature range.

The best you can get for low cost is a GPS receiver module. Most provide a 1 pps output. Long term accuracy is a fraction of part per billion. Pulse to pulse accuracy is in the order of 20 nsec but depends on particular manf. unit as some have software interrupt delivery of pulse that can effect pulse to pulse accuracy. Long term average will be based on GPS timing.

A simple 32.768 kHz clock crystal with ripple divider will give about 100 ppm if calibrated at 25 deg C.
A color burst crystal will give you about 50 ppm over 0 to 50 degs C if calibrated at 25 degs C.
The best AT cut crystal oscillator will give about 10 ppm if calibrated at 25 deg C.

The best free running TCXO module is about 0.5 ppm over 0 to 50 degs C. (type of TCXO used in commerical GPS receivers).
 
Why does the project need to be done WITHOUT a microcontroller?

All yuor problems would be solved from a cheap little PIC with a xtal, rather than try to build a complex multichip solution without a micro?
 
I don't have access to particularily good tools to program or debug micros. My customers generally have even less and some of them don't want the hastle of software/firmware revision control. So I am working in a double niche, what I have happens to be want they want.
 
Last edited:
You could use a 32,768Hz watch crystal and a MC14060b 14-bit counter to generate a 2 Hz pulse. For a 1 Hz pulse, you just add a divide by 2 flip-flop at its output.
 
I don't have access to particularily good tools to program or debug micros. My customers generally have even less and some of them don't want the hastle of software/firmware revision control. So I am working in a double niche, what I have happens to be want they want.

At the customer end it is just another chip. Like the processor in a microwave oven.
 
Go that one. I'm looking to catalog all our ideas. Got any others?
Certainly there are many ways to generate a 1 Hz clock, but I don't think there are any that are more simple than the ones already suggested in this thread.
 
At the customer end it is just another chip. Like the processor in a microwave oven.

My customers are designers and manufacturers not end users. To them the program and programming are two more pieces that in most cases they don't want to deal with.
 
I can supply you with two options.

1. A 6 pin (sot-23) solution with no additional external components accurate to 1%.
2. An 8 pin solution that requires a 32,768 crystal and 2 caps and is as accurate as the crystal.

Both cost around $1.00:D

Edt, after reading the other thread, both solutions can output a pulse every 6 mins if needed.

Mike.
 
Last edited:
In that case I suggest to follow Crutschow's way to use a 32.768KHz crystal with a CD4060 which will put out a 2Hz square wave at Q14. Divide that by two either using 1/2 CD4027 or 1/2 CD4013 you'll obtain the desired 1Hz clock frequency.

If you pull the oscillator frequency to the desired value using one trimable cap you'll have a pretty accurate time piece.

Boncuk
 
I can supply you with two options.

1. A 6 pin (sot-23) solution with no additional external components accurate to 1%.
2. An 8 pin solution that requires a 32,768 crystal and 2 caps and is as accurate as the crystal.

Both cost around $1.00:D

Edt, after reading the other thread, both solutions can output a pulse every 6 mins if needed.

Mike.

By going with a PIC/micro?
 
By going with a PIC/micro?

As they come preprogrammed from the factory they are no different to any other chip. The actual cost is about 47c per chip in quantities over 100, 58c for single units. These prices include programming.

Edit, I'm sure there are a few people on this forum willing to supply the necessary hex file for upload to Microchip.

Mike.
 
Last edited:
I don't have access to particularily good tools to program or debug micros. My customers generally have even less and some of them don't want the hastle of software/firmware revision control. So I am working in a double niche, what I have happens to be want they want.

Thank you for explaining that. Lots of people post questions and try to do things "the hard way" and don't have a good reason. It's refreshing to hear someone actually explain their good reason.

I would second Reloadron's suggestion of mains sync (assuming it is mains powered). A divide by 60 counter is a pretty easy way to give you a 1 second pulse.

Then you can use the binary counters/bcd counters and 7-segment drivers as used in many people's discrete IC clock projects.

The trouble with microcontrollers is that they make things so darn easy in comparison... ;)
 
As they come preprogrammed from the factory they are no different to any other chip. The actual cost is about 47c per chip in quantities over 100, 58c for single units. These prices include programming.

Mike.

Can you provide us with any other details on the chips?
 
The 6 pin 1% chip is the 10F200 and the crystal version is the 12F508. Both available from Microchip Direct.

If you decide to go this route I'm happy to write the code and provide the source/hex files required. In fact I might just do it anyway as it comes up in the forum quite often.

Mike.
 
My customers are designers and manufacturers not end users. To them the program and programming are two more pieces that in most cases they don't want to deal with.

I see your point however you might like to look at an arduino or it's derivatives. Incredibly easy to use and program.
Excellent tutorials here: https://tronixstuff.wordpress.com/tutorials/
They can buy the microcontroller with the bootloader installed, you could write the sketch ( program ) once and all they need to do is upload it to each chip.
Or you could built and supply ready made board for them.
 
Status
Not open for further replies.

Latest threads

Back
Top