How to build a pulse truncator?

I don't know about the designer, but a lot of users miss this option.

But there *are* sure a lot of 555 users. Googling "555 timer" gets 1.4 megahits, including a lot of obvious worshipers.

I guess 555's and LEDs are good for getting kids into electronics. The circuits are simple, they blink and flash, and they provide positive-feedback encouragement.

John

Thanks for the very thorough advice. It looks like a 74221 is the way to go unless I choose to use a microprocessor. I've written quite a few programs in C++ and scripts in PHP in the past, but I've not done any programming of this kind. Using a microprocessor in this case would probably replace seven discrete ICs. If this was a commercial project, I'm sure a microprocessor would be the way to go, but it's a one-off hobby project. The seven ICs are pretty cheap and I don't want to spend a lot of money on interfacing equipment for programming the microprocessor (if the software is available for Solaris, that is - I don't have a Windows box handy). I'm also unsure how long it would take me to learn to write the code. I'll think it over.

Thanks for all the advice.

Chris

--- Sometimes, in the bipolar 555, it's not an option at all since you have headroom and tailroom constraints because the output can't swing rail-to-rail

---

--- What's not to like?

You get a nice, stable ratiometric voltage divider with a tap on it that you can get to, a couple of comparators, an output stage that can source or sink a couple of hundred milliamperes, and an open collector output that can be handy.

---

--- There are a hell of a lot of 555's in equipment out there that wasn't designed by kids getting into electronics, but around here

-- John Fields Professional Circuit Designer

I don't often design stuff that's simple enough to justify a 555. Most of my products have FPGAs and microprocessors with 16-channel TPUs and lots of inputs and outputs, so all we usually need is input RCs and a row of power drivers, SOT-23 mosfets or ULN2003's or whatever; the "computing" ability of the 555 is available in bulk elsewhere. And I guess PIC-type small things can do a lot of the timing internally, too.

OK, I did use a 555 once, as part of a low-power charge-pump, to boost a +12 rail up to +15. CMOS 555 off +5, driving a voltage doubler sitting on +12, 6 parts total. In that case, the single RC feedback made sense, although I could have just as well used one HC14, I suppose. Heck, the HC14 could have been a ring oscillator...

John

It depends what you mean by stable. The first - and as it turned out - the only time I took a serious look at a 555 as a timer, it got rapildy rejected precisely because it wasn't stable.

The infinitely better LM322 wasn't stable enough for our job either, and I spent a couple of weeks finding out that an emitter-coupled multivibrator got screwed up enough by the temperature dependence of the Early effect that that wasn't an option either, so we enede up going back to cyrstals (which were a bit better than we needed).

--- Well, hello Bill! :-)

How nice to actually engage in a technical discussion with you instead of what we usually seem to get caught up in here.

I've posted the first two pages of Signetics' 1976 data book's data sheet for the 555 to abse for your perusal.

Note that on page 158, under "FEATURES", the temperature stability is given as 0.005%/°C, which is consistent with the 50ppm "typical" timing error attributable to temperature on page 159 for the NE device.

Also note that on page 159, for the one-shot, the drift with temperature is given as 100ppm/°C worst case for the "SE" device, which is, I assume, the device you would have chosen for testing.

Further, note that the drift with supply voltage is given as 0.2% per volt of supply voltage change for the SE device.

Since the chip itself seems to have been specified as able to perform nicely with changes in temperature, I wonder if the disappointing performance you found wasn't due to the tempcos of the timing resistor and capacitor and/or the stability of the DC supply.

---

--- Good decision, IMO.

I'll _always_ throw in a few pennies extra to keep the kids from showing up, uninvited.

-- John Fields Professional Circuit Designer

We will tell you THE way to go, you contemptible little snooty retard troll, the '121 is rarely it, and no one is interested your summary of the application.

You can use the same arguments for the 2N2222A, all PICs, 8051s, STAMPS, MSI logic, 1K ohm resistors and quite a few other parts. But the 555 and flashing lamps are passe, it is the microcontroller applications that have the kids' interest these days.

Right- well when they ask for something called a "pulse truncator" that is a giveaway there's not much there- and then they have the gall to rate responses and tell us what they have decided...too much.

Does _anyone_ use TTL anymore?

...Jim Thompson

TTL isn't even competitive as glue logic anymore. ECL is alive, well, and thriving though.

Which isn't to say that knowing about TTL is not essential. A few months ago I spent four days going through literally 300 pages of the old time gate level schematics of a 1980's mostly TTL product for purposes of reverse engineering the control and timing for upgrade to a power PC implementation. There had been an intermediate upgrade to that Atmel mux based logic in the 90's but the performance was intermittently faulty because they never went into the detail I did, and I know where they screwed up. That's the difference between hardware and software orientation. Also, they paid little attention to some seriously complicated interlock logic involving about 50 different failure conditions under which you would want to shut a certain something down.

Does a 74F38 count?

John

I found that for simple, not overly time critical applications, selecting the same part of the same vendor yields very reproducable results. Built a series of simple RC model PPM applications around the Philips HEF40106, which all worked without tuning at all. Switching to another brand would require tweaking indeed.

So I would consider this a nice solution for hobby projects. Obviously not for industrial mass production. (where I would hate to use 555's as well :-)

Already gone when I lloed, but the 1987 Naitonal Semiconductor Linear Databook 3 has their data sheet for the same part. See also

which is dated February 2000. The Philips data sheet is also available on line

and - since Philips bought up Signetics early on - this is probably close to your original.

Not a chance. Supplies if military-rated devices were very erratic. I was working for Kent Instruments in Luton at the time, where the accounts had chosen to try to save money by not paying the distributors as fast as the distributors wanted, which made even widely stocked parts hard to get hold of - I wasn't allowed to introduce a new component if there were fewer than four manufacturers.

Since I was using 15ppm/C precision film resistors and some +/-30ppm/C silver mica parts we had in stores, you can stop wondering. Stability against supply voltage was explicitly checked by varying the supply voltage +/-5% - the acceptable range at the time - while monitoring the voltage with a DVM (not the 6-decimal-digit HP monster, unless there was nothing lighter handy).

Temperature stability was similarly checked by shoving the whole circuit into a temperature controlled oven (we had several of them - a luxury that I've rarely had since then).

Not mine. The same - very bright - boss who launched me on this wild-goose chase also decreed when we stopped. I was too wet behind the ears at the time to have had a useful opinion (but not so wet that I didn't know that my opinion would have been ill-informed).

--=20 Bill Sloman, Nijmegen