Pulse Circuit Needed

Plenty of ideas. But first you will have to clarify your output signals a little better. Like how wide of a pulse width do you want to happen that is spaced one second apart? How precisely do all of the values have to be? Since you haven't specified anything but the 1 second interval, let me offer the perennial 555/556 timer circuit collection as an initial answer. You'll also find a nice collection of cookbooks for this popular chip.

Oh it also works quite well with the atypical supply voltage range you mentioned.

I am looking for a circuit to pulse a lamp dimmer circuit. I want them to turn on with half or 1/3 brightness, thus the initial pulses. Timing accuracy is not important at all. I would think 1/2 second pulses spaced 1/2 second apart would be adequate. I was hoping that you would recommend the 555 timer. I just don't know of an easy way to generate on 2 pulses or 3 pulses.

Thanks for your comments.

David

Well as you point out, your timing precision will be easily met by a 555 timer circuit. So now you need a circuit that can count to two or three to turn the clock circuit off. I wonder if any of the cookbooks has a circuit that can count?

Good morning redfred.

After reading your two responses here, I checked the CR4 FAQ and found "Good answer votes should be given only to comments that provide a valuable technical contribution".

You have done exactly that here, and I vote 'Good Answer'. Nicely done.

CR4 member babybear does not come around too often, and I miss watching his gentle paw guide people down the path to better understanding and clearer thinking. He has some big shoes to fill, but you seem to be filling them very well.

I've used this part to perform similar functions.

http://www.kpsec.freeuk.com/components/cmos.htm#4060

It should be able to provide the pulses you require.

Additional info: After reading your post #2 I suggest you just to put the value of 1μF (or 10μF) for all caps of the circuit. [In my application I wanted to get two small sequential pulses (and that's why I putted 100nF) with a distance between them of 1sec.] Then you can "play" with the resistors' values (i.e. by trimming).

You can use a small linear regulator (e.g. 78L05) in order to get the 5V and power this circuit. And don't forget to place decoupling caps (100nF) close to the ICs.

Thank you ... this looks like just what I need. I will breadboard it this weekend.

David

I suggest you to use "Schmitt Trigger NOT" (i.e. an 74HC14 instead of an 74HC04) due to the slow charging/decharging time durations (1sec) of the caps. In this way, there will be no "bouncing" at the moment where a NOT changes the state of its output.

The circuit is simple and it uses (essentially) just one chip, i.e. an 74HC14 (which contains 6 gates). [Just consider that the OR gate could be replaced by [two diodes and a pull down resistor] in an "ORing connection" . So, you use just one 74HC14 for the "two pulses generator" (and two 74HC14 for the "three pulses generator") ]

You got a G.A. from me. This is the simplicity of analog electronics.

But just as I'm away from this field, I was considering whether one could use a 555 and a counter or even a shift register and "program" not only two or three but as many pulses as he wishes, by manipulating some switches that initialize the counter. It's certainly more complicated, but also more flexible to changes. I wonder, what will be the use of such a device, so I drop this idea too. What do you think G.K.?

Yes tkot. Redfred presented this idea (555 & counter) in his post #3. By using (e.g.) the 74HC4060 you can have 1 to 12 sequential pulses (by just selecting the proper counter's output that will stop the operation of 555).

A shift register (taking its clock by an e.g. 555) could, also, do the job by having 1 & 0 alternatively to some of its parallel inputs (e.g. 74HC165 could give 1 to 4 sequential pulses). However, in this case, you don't have variable duration of the pulses and their separation.

Which solution is better (concerning the number of chips needed for my solution and the [555 & counter] solution) depends on the number of pulses. E.g. in the case of 2 pulses I need just 1 chip (better) while in the case of 4 pulses I need 3 chips (worse). If multiple outputs are needed the [555 & counter] solution is better than mine for more than 4 outputs (out1:2 pulses, out2:3 pulses, out3: 4 pulses, out4: 5 pulses): I need 4 chips while the other solution needs 3 chips (i.e. 2 dual timers 556 and 1 counter).

Of course the more "universal" (and more expensive) solution is to use a small PIC controller which could give you multiple outputs of several sequential pulses with variable durations by just using the proper software.

The CD4060 was suggested in post #4. Quick review of the data sheet shows that the CD4060s built-in oscillator and numerous divide-by-2 outputs can be used as a gated pulse generator. Another viable option would be the CD4017 and an external oscillator.

_________________________________________________________________________

"Give someone a fish and you feed them for a day. Teach someone to fish and you feed them for a lifetime."

My apologies: I didn't remember that the 4060 has a built-in oscillator. In fact, in my previous post, I meant 4040 (and not 4060) as it has 12 outputs... Just a small correction...

You could use one of 4060 outputs to reset (i.e. stop) the chip itshelf. However you can't use it for this kind of application. Remember that we want a sequence of 2 or 3 or 4 or ... pulses. As the outputs of a counter (e.g. 4060) are divided by 2, using one of these outputs to "stop" the chip will give you an even number of pulses (at its output Q4) i.e. you can get 2 or 4 or 8... pulses. You can't get (e.g.) 3 pulses.

So I have, also, to reconsider my previous post (i.e. to "improve" Redfred's suggestion): The proper solution is a [555 & shift register] (instead of [555 & counter]). You can use (e.g.) the 74HC194: you "fill" its outputs with 1 in every clock pulse. Then you can use one of these outputs to "stop" the 555. That's the proper way (just to be accurate).

[Moreover the use of a 555 gives you more "flexibility": you can have variable pulses duration.]

No apologies necessary, no offense taken

"Single" chip solution to get 2 and 3 pulses is not easy. PIC micro (like the tiny 8-pin soic version) would work great, but cost would be higher and the effort requires some programming skills and tools. Obviously the CD4060 will not do 3 pulses without an extra logic gate. Two chip hardware solution (analog, digital, or combo) is fairly easy. Oscillator(555) and counter(or shift register), or counter with extra logic gate can get the job done per OPs request.

Whenever possible, I prefer to guide towards answers instead of just handing them out. The learning experience can be more rewarding than the end result.

Whenever possible, I prefer to guide towards answers instead of just handing them out. The learning experience can be more rewarding than the end result.

I tottaly agree with you. Especially in the case of a class: The teachers must not just provide knowledge to their students. They, also, have to "learn them how to learn" and guide them to find the answers and solutions by themselves. This is the appropriate way of teaching.

(However -in CR4- sometimes it's easier just to give a clear solution rather than trying to guide another guy to a solution.)

GK... could you email the circuit to me. Hard to read and print from this post.

Thanks to all for the ideas and comments.

David

Contrary to my own suggestion of not providing a direct solution, see rough sketch below. You will still need to review CD4093 data-sheet to determine approximate timing values and some breadboard tweaking will be necessary. Good luck!

Nice solution. You got a GA from me. In this way you can have 2 & 3 sequential pulses (i.e. 2 outputs) at the same time by using just one chip.

There is a drawback though: The "gate time" is not independent from the "push time" of the push button. I.e. you have to push the button momentarily in order to get the proper "gate time" (and this depends on the user's skills). If, for example, the user pushes the button for 1sec (instead of a momentary push) then he will get -at the 1st output- 3 pulses (instead of 2 pulses). So, you should add another stage right after the push button: a network {[C & R//D] + gate} (like the one I use at the 2nd stage of my schematic) (and, of course, you have to connect the button to gnd instead of V+). In this way, right after this stage, you get a narrow, standard pulse (high) and then you use this pulse as the "trigger pulse" for your "gate timer" networks. As a result, the behaviour of your circuit is always predictable (i.e. independent from the "push time" of the button).

Another improvement is to use a somewhat more complex network for your oscillators: Instead of an R, you can have [R & D: —ι>ι—ΛΛΛ— ] // [R & D: —ι<ι—ΛΛΛ— ] combined with the C. In this way you can get variable "on" and "off" time durations of the output pulses.

Yes tkot. Redfred presented this idea (555 & counter) in his post #3.

Oops, indeed. Sorry redfred. Still, the idea to use a shift register (or more than one in cascade mode) can give you the chance to configure the number and the width of the pulses (as soon as they are multiples of the clock period).

Just for your information, I once used an 8-bit ROM filled with certain patterns and a 555-clocked counter which addressed the ROM sequencially. That way, I could get 8 data outputs configured to produce any pulse sequence I desired!

OK, that one was a more complicated problem, so don't hammer me...