How inaccurate is a 555 or 7555 REALLY?

By unit volume, I'd say the percentage of "serious" electronic designs -- where there are critical component values, fancy discrete analog topologies, etc. -- has been on a steady decrease since the early '80s when everything started doing digital...

Well, everything's become much more integrated these days. I remember a design I did some years ago with a 74LS123 that got sucked into an FPGA when we made a 2nd version of the product with expanded functionality...

They're temperature-controlled by body heat.

Set one on the nightstand for a few nights, and see how accurate it is. ;-)

Cheers! Rich

Not at all. Legacy designs, and land occasional egacy designer, like John Fields, are an entirely sufficient explanation.

It would be interesting to see the shape of the buying curve - the number bought per year every year since the part was introduced. I suspect that that curve peaked a number of years ago, and part is now selling predominantly into the continueing production of old designs.

I've got a copy of Camenzind's book, and I've not only read the chapter on the 555, but I've also commented on it here - Camenzind can now see how he might have done a better job on the original 555, not that this would have made much difference to its usefulness or sales.

No case to answer.

....... Phil

It is an impressive number - one for every six people on earth every year. One has to wonder where they all go - washing machines and hair-dryers and the like, I guess.

A closely reasoned response

Leading to a self-serving, but unsupported conclusion.

The book doesn't make a great fuss about the long term sales numbers - though the initial market response was pretty spectacular.

You have provided one specific sales number, presumably derived from Hans Camenzind's oral history , and no year-on-year sales numbers, so the significance of your "sales numbers", whatever they are, can only be appreciated in the echoing confines of your own head.

You do remember that he thinks Maxim has only sent out samples and has never shipped an IC needed for a production run?

The designs could be smaller, cheaper, faster and more current-efficient with more modern designs and components, but manufacturers are extremely nervous about disrupting an existing production process. When I - briefly - worked for Chessell Recorders in the U.K. in 1979, their 100,000-ish per year pen recorder still used a six-transistor servo amplifier that blew up whenever the pen carriage stalled. Development had been wanting to replace it with something cheaper, smaller and stall-proof involving an op amp for years, but production wouldn' risk any change to their smooth-running production.

Eventually, the Eurotherm board shook up the Chessell management team - four of us resigned from Chessell's on the same day (purely by coincidence) and all got to talk to the chairman of the Eurotherm board. I hadn't been there long enough to say anything interesting, but others had been there longer ..

Even if a slightly more broadly educated designer might have made another choice.

Yes. I've seen them. All very nostalgic. Your recent exercise with a

In terms of components, you should have been able to squeeze your design into a single CMOS 22V10 programmable logic chip. With something bigger but pin-compatible, like the ICT 7024 PEEL part - which is the most recent programmable logic part that I've had a chance to play with

- you could proably have also fitted in Arlet's suggestion of dividing by 79 once in a while to get the average frequency closer to 420Hz, or divided down from David L. Jones 10.752MHz crystal and got exactly

A more modern programmable logic part - like the Xilinx Coolrunner series - with in-system programming, 1.8V supply and some really compact packages - looks as if it would be even nicer, but I've not yet been able to contrive an excuse to play with one of these parts.

Do you by any chance work for some kind of museum of technology, exhibiting your skills alongside the guys who make flint arrowheads?

On Tue, 05 Dec 2006 15:40:30 GMT, Robert Baer Gave us:

Just like a Usenet IDIOT to mouth off like an ass instead of giving the explanation.

I know this really must be a BEAR for you to cope with, but he is right.

WHO is it you think is right, me or Baer? I can't tell from your attributions.

...Jim Thompson

I couldn't find any numbers to supplement his single number - which, of itself, isn't all that informative.

I notice that you haven't any information to add on the historical trends in 555 use, apart from your personal conviction that it is the greatest thing since sliced bread.

National has a CMOS part (LMC555) in a rather compact package ~ 1.4mm square(ish) by 0.5mm thick with 0.5mm pitch balls.

Best regards, Spehro Pefhany

Of course I knew how. But the resistors I would have had to use were

A PIC may be programmable, but it isn't programmable logic, and can't do a lot of things that are easy in programmable logic.

Programming programmable logic can be handled through a variety of interfaces, depending on the manufacturers programming software (usually available free for everything except bleeding-edge new parts, which the Xilinx CoolRunner isn't - Xilinx bought it from Philips).

Most of the ones I've run across have included a graphical option, which makes the process similar to hooking up logic gates and bistables.

I prefer to write out a seres of logic equations.

I didn't find any of these approaches difficult to learn - nothing remotely as difficult as mastering a computer language, where it took me a week to learn Fortran 4, and nearly a year before I could produce a well formatted page of output (including printed graphs). The MACRO-8 assembly language for the PDP-8 was much easier - mainly because the teletype output made more sense.

There is a certain learning curve in learning how to use Boolean logic, and you need extra time learn how to hook up CMOS parts to realise that logic - most people would be better off putting the exra time into learning how to program programmable logic.

The choice of which programmable part should probably be referred to comp.arch.fpga, but one thing is for sure - you will be able get all the logic into a single modern PLD.