How to build a pulse truncator?

A small series cap with pull up or down resistor ( cant remember which ) to the trigger I/P of the cmos555 will do it.

Taff.... ( Thunder Ridge PA, Bristol, UK )

----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----

----= East and West-Coast Server Farms - Total Privacy via Encryption =----

Sn74HCT123 one-shot, or an r-c differentiator followed by a schmitt trigger.

John

Thanks for the two suggestions. Taff, I think it must be a pull up resistor. John, you're thinking along similar lines as me. Were you thinking of having a 555 monostable driven from the Schmitt trigger, or adjusting the pulse length using the values of components in the RC differentiator? Any thoughts about which is the best solution?

Many thanks for the help.

Chris

Sorry, I am deathly allergic to 555s. Please don't mention them in my presence.

Yes.

Well, I suppose a simple R-C differentiator might make the, um, XXX device fire on one edge. Perhaps someone else could advise you on that.

John

Sometimes it can be this simple - 2 components:

Vcc---------+ | +----> out to 555 Q ----+ | | | | | [Relay] +-->|

Are you happy writing software? Consider using a PIC.

They come in 8 pin DIPs. You don't need any extra R/Cs, but you do have to write some software. I had a similar problem a while ago. It took 100 lines of code in a 12F629.

Atmel/AVR has similar chips.

Hear Hear! I thought I was the only one.

:-))

Try and get away from using a 555 for every job. There are better monostables - nobody seems to have mentioned the 74121 and its dual, the 74221, which give very stable pulse widths, at the price of being non-retriggerable (and you don't seem to need retriggering).

Building your own monstable in TTL is a waste of tinme and effort. With regular inverters and gates, the tolerance on the threshold voltage gives you a large range in pulse width - something like ten to one when I last looked. Schmitt trigger input parts are better, but still pretty horrible. HCT parts aren't as bad as regular TTL - the outputs at least swing from rail to rail, even if the thresholds are still loosely defined and in the wrong place. Regular CMOS, with threshold voltages between Vcc/3 and 2.Vcc/3 is better than TTL, but still pretty sloppy.

--- This'll work; view in Courier:

_____________ _____________ JKQ ___| |_____________| |_________ _ ___ _____________ _________ JKQ |_____________| |_____________|

______ ______ OUT2 __| |____________________| |________________

______ ______ OUT1 ________________| |____________________| |__

You can get two 7555's in a single package, a 7556, but if you wanted to save 6 parts you could easily switch to a 4538 or an HC123. If you'd prefer to use those parts I have some schematics I can post, if you like.

-- John Fields Professional Circuit Designer

For your application wherein the triggering pulse is longer than the output pulse, there is a particularly simple positive going edge triggered 555 monostable configuration as shown below. The purpose of the RC delay to the TRIG input is to allow sufficient recovery time for the 555 internal latch while holding TRIG below the internal threshold. The internal latch then SETs and puts OUT high on the leading edge of the EXT trigger input. The RtCt then times out and RSTs after 1.1RtCt. View in a fixed-width font such as Courier.

. . EXT TRIG ALWAYS LASTS LONGER THAN PW OUT . . 5V . .------+----------------. . | | | . === [RT] | . 100P | | . | | | . | | | . .----+ | | . | | | | . | | | | . | | | | . | [100K] | | . | | | | . | | | +-------------+ . EXT >--------+----------------|RST V+ | . TRIG | | | OUT|------>PW=1.1xRTxCT . | | | | . | +---------|THRESH | . | | | | . | +--[100]--|DIS | . | | | | . '---------------------|TRIG | . | | GND | . === +-------------+ . CT | . | | . ---------------+----------------+------------------

Trigger is active low so the resistor needs to pull up.

Do we have to spin around 3 times and swear?

How about, to allow small parts, we use a CD4060 in the circuit. An RC could make a short reset pulse and Qn could be hooked by a diode into the oscillator. The only problem is there has to be an inverter to make a positive going pulse.

In article , ehsjr wrote: [...]

It can be even simpler. Use a 1 second thermal delay relay.

With both these ideas you need R^2L to drive it so you have to make a JK out of parts.

In article , Hal Murray wrote: [...]

Use an 8051. Lets see how many line off the top of my head

mov P1,#0FEH ; LSB is the output, init low

loop1: jnb P1.7,Loop1 ; MSB is the input wait for high setb P1.0 ; Output high mov R2,#OneSecLSB ; Load the delay mov R3,#OneSec2SB ; .. three bytes mov R4,#OneSecMSB ; .. long

Loop2: djnz R2,Loop2 djnz R3,Loop2 djnz R4,Loop3

clr P1.0 ; End output

Loop3: jb P1.7,Loop3 ; Wait for low sjmp Loop1

Well, maybe you don't want to use one as a monostable, OK. OTOH, a

To the OP: a nonretriggerable monostable is a good match to the problem. The 123 and 555 are retriggerable, meaning that unless you do something fancy, they'll give you a pulse at least as long as the trigger pulse. Nonretriggerables will time out even if the trigger hasn't gone away.

Cheers,

Phil Hobbs

--- "Retriggerable" means that the length of the output pulse can be extended by triggering the input before the output pulse times out.

The 123 is either retriggerable or not depending on how it's wired and there's no necessity for its input pulse to be shorter than its output pulse in prder for it to time out properly.

The 555 isn't retriggerable (its output pulse width can't be extended by repeatedly pulsing its input) without external components, and its input pulse must _always_ be shorter than its output pulse in normal operation as a one-shot.

-- John Fields Professional Circuit Designer

I wonder why most everybody uses the silly charging resistor/discharge resistor, when a single RC from the output is so obvious?

John

The 555 designer missed the obvious ?-)

...Jim Thompson

Makes the frequency sensitive to heavy output loading to ground, assuming the original bipolar version and a 5V supply.

With the modern CMOS versions and light output loading, of course, it's the ideal way to get a 50% duty cycle.

Best regards, Spehro Pefhany