Okay so I posted an earlier post regarding using a 555 timer to create a clock timer for a midi controller, but have decided to abandon the idea and retweak the idea with my teammates.

How the initial stage of the midi controller is gonna work is that a 5V supply will be connected to my timer stage's input. As soon as power is applied, a 32kilobit signal is sent out each second to a next-stage shift register. This keeps going until the unit is powered off.

Any ideas of how I can go about constructing this? Something simple yet effective is preferred.

 

Hello,

What do you want to do with the timer?

Here is a page with some midi projects:
http://www.ucapps.de/index.html

This is a link from the EDUCYPEDIA page on the midi standard:
http://www.educypedia.be/electronics/midi.htm

Bertus

 

Do you mean a clock frequency of 32,768 Hz?

If so, why not use a watch crystal and a couple of caps driven by a 4093 NAND gate?

 

Do you mean a clock frequency of 32,768 Hz?

If so, why not use a watch crystal and a couple of caps driven by a 4093 NAND gate?

That's not a bad idea, I never thought of that idea.

And yes I am certain that they mean 32,768 hertz as a frequency output, according to their conversions that they gave me.

 

To make the oscillator more stable, make it an OCXO.
Check out Roman Black's OCXO page:
http://www.romanblack.com/xoven.htm
Cheap and effective.

 

What would the ideal capacitors for this be? something really low or something high?

Or go with what the 4093's data sheet says for voltages 5V-15V (50pF)? This is found on table 10 of the 4093's datasheet.

 

Hold on a second there. What exactly does "a 32kilobit signal" every second mean? 32000 bits sent (at some unspecified data rate, but presumably faster than 32K/sec) every second? Or a signal (of unspecified size) sent at 32000 bits/sec, once a second?

If it's MIDI, the baud rate is 31.25KB, so don't fool around with 32.768KB. The odd-looking 31.25KB rate is generally produced by dividing down the clock of a microprocessor running at some integral number of MHz.

 

Good point, definately 31.25KHz
http://en.wikipedia.org/wiki/MIDI_1.0
.
.
Thats 1MHz divided by 32 - perfectly possible with a few extra logic chips.

 

OK, then get a CD4060B/HEF4060/MC14060, a 1MHz crystal, and build the circuit as shown in the datasheet (look at the MC14060 datasheet, figure 6 on the bottom of page 5, and take your output from Q6 (pin 4).

Note that you'll need to make a circuit board or "dead bug" it. Using a breadboard probably won't work very well; too much parasitic inductance and capacitance.

You'll need an 82pF and 20pF ceramic cap. For Ro, a 22k resistor should work OK. Instead of 18 MEG, use a 10MEG resistor. The exact value isn't critical; it just needs to be in there to prevent the clock from "stalling".

 

Maybe using a processor to process the signal will be better.

 

Well, a uC would likely save a lot of fooling around with individual logic IC's.

In any case, they're going to need an accurate timebase.

Our OP really hasn't expounded much on what they are trying to do; vague pretty much covers it.