Best Chip Recommendation - 1 pulse / min

Thread Starter

PastorJeff

Joined Jun 13, 2020
8
I am an amateur seismologist. I currently have three seismometers running into an a/d board where the analog signal is processed and then sent to a monitor on the wall.

The software has all the times except there is no display on each trace for a minute mark

I'm wondering what people are doing today to get single pulses out at predetermined times?

Years ago, I built this circuit using decade counters and divided down to the pulse / minute

While it worked great, it wasn't very accurate over time

I know there are clock chips which have embeded crystals for accuracy, but have never used one
or know exactly what to look for. Any assistance in pointing me to a device or two would be great.

I am going to build buffer amps for each input, configuring them as summing amplifers with no
gain.

I also need to be able to change the signal from the clock output to set the pulse height on the
display. I'm guessing a pot as a voltage divider before the summing input? Or is there something
inherent with the clock chips that I can vary the pulse output?

Thanks a ton for any suggestions
 

jpanhalt

Joined Jan 18, 2008
11,087
If I needed a pulse per minute with other things being done, I would use a scaled (e.g., 1:60) timer interrupt with almost any microcontroller.
 

dl324

Joined Mar 30, 2015
16,843
Welcome to AAC!

What options do you have available? What accuracy do you require?

A crystal controlled timing device will have good short term accuracy, but long term will be lacking. This includes counters and microcontrollers, e.g. Arduino.

If you use something that polls a time server (like Raspberry Pi), you can get the best of both worlds.

Another option is using line frequency, but we don't know if that's an option.

Not sure what you're asking for regarding pulse height. What every solution you choose will determine pulse "height" (which I assume means voltage). You also need to specify pulse width.
 

MrChips

Joined Oct 2, 2009
30,707
As John has said, since you are already using an MCU there is no need for an external timer circuit. Use whatever timer feature is already available to count 60 seconds.
 

Thread Starter

PastorJeff

Joined Jun 13, 2020
8
Honestly, I did this in the past and used decade counters and op amps and it worked great. I'm really not interested in using a microprocessor or an Arduino. The responses tell me how far removed I might be in the world of current electronics.

I don't need atomic level accuracy. I would like something that would give me a minute pulse which I can put into a buffer op amp circuit and get a mixed output. The analog seimometer signal and the minute mark. Most of the time, the trace is running quiet until an event rolls through.

I am not looking for the minute mark to ramp up to cover the quake signal. I just want to be able to change the pulse height at the input of the summing op amp when I change gains in the a/d converter.

I do not have access to the a/d circuit.

I was hoping someone would say, "Ooooooh, National makes a great little clock chip that will do exactly what you want. It will be crystal controlled, you can pick out a minute pulse on pin ? and you are good to go.

If that isn't real world, then I just need someone to say, "Hey! That isn't possible!" Then I will smile and go back and build a decade counter and re-create what I had. It just wasn't very accurate over time. Hope this helps.
 

dl324

Joined Mar 30, 2015
16,843
Then I will smile and go back and build a decade counter and re-create what I had. It just wasn't very accurate over time.
You still haven't said what accuracy you were getting and what accuracy you need.

What timing "devices" do you have available?
 

MrChips

Joined Oct 2, 2009
30,707
Well now that you asked, National did have a chip MM5369 that you put a 3.579545MHz color burst crystal on it and you got 60Hz output.
 

Delta Prime

Joined Nov 15, 2019
1,311
Hello there. :) You got to forgive these youngsters pastor they got to go way back to go discreet I hope this helps.photo_1592097946896.png
R1 – R28 = 560 Ω
R29 = 10 KΩ
R30 = 1.8 KΩ
C1 – C17 = 0.1 µF Caramic Disc
C18 = 1 µF/16V Electrolytic
IC1 – IC6 = 74LS90 (Decade Counter)
IC7 – IC10 = 74LS247 (BCD to 7-Segment Decoder/Driver)
IC11, IC12 = 74LS373 (Octal D-type transparent latches)
IC13 – IC16 = 74LS85 (4-bit magnitude comparator)
IC17 = 74LS21 (dual 4-input positive AND gate)
T1 = BC547 NPN Transistor
T2 = SL100 NPN Transistor
D1 = 1N4001 Rectifier Diode
 

Thread Starter

PastorJeff

Joined Jun 13, 2020
8
I'm laughing... Thank you. My gosh. I was waiting for someone to take my Tahoe's cpu out and tap into the clock circuit.

What would you use for the 1hz source that might be more accurate than a 555?
 

nsaspook

Joined Aug 27, 2009
13,079
So you need to mix a analog pulse (how long of a duration? 250ms, 500ms, ?) every 60 seconds with the analog seismo signal so the rolling strip display from that signal will show a tick every 60 seconds. It's trivial to do that with a MCU and external clock chip. I have a few left-over modules with a pic18f1320 and 10MHz clock. It would be simple to divide the clock down to for a 60 second tick pulse of x ms width.
 

Thread Starter

PastorJeff

Joined Jun 13, 2020
8
So you need to mix a analog pulse (how long of a duration? 250ms, 500ms, ?) every 60 seconds with the analog seismo signal so the rolling strip display from that signal will show a tick every 60 seconds. It's trivial to do that with a MCU and external clock chip. I have a few left-over modules with a pic18f1320 and 10MHz clock. It would be simple to divide the clock down to for a 60 second tick pulse of x ms width.
Sounds great! How do I make this happen?
 

jpanhalt

Joined Jan 18, 2008
11,087
Honestly, I did this in the past and used decade counters and op amps and it worked great. I'm really not interested in using a microprocessor or an Arduino. The responses tell me how far removed I might be in the world of current electronics.
Has that plan changed?
I was hoping someone would say, "Ooooooh, National makes a great little clock chip that will do exactly what you want. It will be crystal controlled, you can pick out a minute pulse on pin ? and you are good to go.
"On September 23, 2011, the company formally became part of Texas Instruments as the "Silicon Valley" division." (Google) Have you checked Texas Instruments?

Later you mention wanting something more accurate than a 555. A 555 is as accurate as any RC oscillator. That implies using a crystal or ceramic resonator as the clock source. Here's one way with a crystal:
1592130264580.png

Of course, any MCU with which I am familiar makes dividing down a clock to 1 second easy and converting that to minutes is just as easy. You can even chose a crystal with a frequency that is divisible by 2^n to give seconds.
 

atferrari

Joined Jan 6, 2004
4,764
The chip inside most of common alarm clocks, gives a 2 Hz pulse that could be good if you go the "components" way. Verified that in many in the distant past.
 

MrChips

Joined Oct 2, 2009
30,707
Just ask and there are any number of members here who can program and send you an 8-pin MCU chip that outputs your desired pulse width and repetition rate. You can have the no-crystal version if you do not need high accuracy.
 

KMoffett

Joined Dec 19, 2007
2,918
I've used the guts of 1.5V wall/desk clocks for 1Hz pulses, The little circuit puts our two 0.5Hz pulses to drive the clock motor. If you sum these you get 1Hz pulses. Some circuits put out 1.5v>0v pulses. Some put out 0v>1.5v pulses. The attached is how I use them.1 Second Clock.gif
 

BobTPH

Joined Jun 5, 2013
8,808
A microcontroller solution is by far the simplest hardware you could get. An 8 pin chip, one capacitor, and a crystal if you want accuracy of better than 1%. Compare that to your decade counter solution.

The reason you don’t see single purpose chips for things like this is the fact that a microcontroller IS the single chip solution.

Bob
 

nsaspook

Joined Aug 27, 2009
13,079
Wow... 1 pulse / minute With the pulse width 1 sec. Wow...
I have the circuit built and tested but need a little time to calibrate (the software constants) the cycle time to 60 seconds ticks (120 second clock).
A older PCB version of this with the PWM analog filter replaced with a coupling capacitor.
https://github.com/nsaspook/ihc_mon/blob/re20a/ihc_monb.pdf
https://github.com/nsaspook/ihc_mon/blob/re20a/ihc_mons.pdf

IMG_20200614_175840.jpgIMG_20200614_175937.jpgIMG_20200614_153620.jpg

https://github.com/nsaspook/ihc_mon/blob/tick60/dale_10mhx_osc.pdf

Rewritten existing software code for the PIC18F1320
https://raw.githubusercontent.com/nsaspook/ihc_mon/tick60/ibsmon.h

ISR routines. I just reused old code so the names don't related to timing generation.
https://raw.githubusercontent.com/nsaspook/ihc_mon/tick60/ihc_vector.c

Setup, main code.
https://raw.githubusercontent.com/nsaspook/ihc_mon/tick60/ibsmon.c
 
Last edited:
Top