Why all of the Logic Chips

[Wilksey]

A handful of logic chips is cheaper than a microcontroller, and they work in parallel execution as people have already mentioned.

An alternative which I use quite frequently is a CPLD or FPGA, but it's an expensive bit of kit (especially if you need an external PLD memory chip) for just an AND gate!  My PLD/PGA's are usually doing a lot more so I just make the logic inside of the device as there is always some spare LE's and pins left.
If I am not using a gate array or anything of the sort, and the micro won't cut it for timing / real time reasons etc then it just makes sense to put a 74 series chip on board, the only thing I find annoying sometimes is because they are usually quad devices (and are usually the cheapest) they can take up space on the board, but there are many different packages available these days that you will probably find something suitable whatever the requirement.

If you have ever read a book on digital electronics you will know that the 74 series is the holy grail even if it to build a small circuit before putting it into a PLD/PGA or ASIC, unless I have read the wrong books

[coppice]

A handful of logic chips is cheaper than a microcontroller, and they work in parallel execution as people have already mentioned.

If you don't need speed, a 25 cent MCU can replace a *lot* of simple logic chips. If you are paying, say, 6 cents a pop for 74HCxx devices, you'll break even at just 4 chips, and changes are wayyyyyyyy easier to implement cleanly in an MCU.

[Wilksey]

But you don't have to program a logic chip....

[WarSim]

A handful of logic chips is cheaper than a microcontroller, and they work in parallel execution as people have already mentioned.

If you don't need speed, a 25 cent MCU can replace a *lot* of simple logic chips. If you are paying, say, 6 cents a pop for 74HCxx devices, you'll break even at just 4 chips, and changes are wayyyyyyyy easier to implement cleanly in an MCU.

You are defending you opinion and ignoring the competing points. 
Yes at low time constraints a cheep uC could be less expensive. 
These are the points made that have been made that you seem to be missing. 
A cheep $0.25 uC can not supply <1nS latency of multiple gates. 
For a uC that can the power requirements will be much higher than That of a gate. 
Startup time is much lower with a gate than a uC.
The support requirements and costs are higher.  (In circuit programing/production programing/pre-programming all cost money) (all uC require more support passives than a gate)

The question was why all the gates.  People here gave valid reasons for opting for gates in certain circumstances.  You are countering with over generalized noise.  Do you have a point to add.  Are we completely missing your argument for using a uC for situations that can not meet the required "real time" timing requirements mentioned? 

[amyk]

I see nobody has used the reasons of serial vs parallel and synchronous and asynchronous.

Actually I mentioned that in my first post in this thread...

Another reason for using individual gates over a CPLD or FPGA: there may be various places in the design that need a few gates, but they're spread out and trying to route all the signals into one CPLD/FPGA would be nearly impossible. I see this a lot in laptops - several dozen single or dual-gate devices are spread out over the motherboard and used for various "glue logic".

[WarSim]

I see nobody has used the reasons of serial vs parallel and synchronous and asynchronous.

Actually I mentioned that in my first post in this thread...

Yes you did, apparently I missed it. 

[Dave Turner]

The 7400 series is also ideal for teaching logic. Pins can be easily probed and the effects of non-ideal timing signals demonstrated.

[tszaboo]

A handful of logic chips is cheaper than a microcontroller, and they work in parallel execution as people have already mentioned.

If you don't need speed, a 25 cent MCU can replace a *lot* of simple logic chips. If you are paying, say, 6 cents a pop for 74HCxx devices, you'll break even at just 4 chips, and changes are wayyyyyyyy easier to implement cleanly in an MCU.

And, the 7400 or 4000 series logic IC (although, several have died over the years) you will be able to buy 30 years from now, while your DS80C310-QCG/T&R-ND might not be able to buy tomorrow.

[T3sl4co1l]

To be fair, most programmable logic have life cycles much longer than you'd expect.  Altera Cyclone I and II are still around, I think the first is finally NRND or discontinued?  Altera at least seems to give a 10+ year life cycle.

And some things never go away, like PALs/GALs, which are still on Digikey no less.

Tim

[WarSim]

Funny how a single decade is considered a long time now.  A decade is even called forever by the dramatic. 

[BloodyCactus]

for me, I have some logic chips in my design in the watchdog circuit because I need to rely on something other than the MCU if the MCU does something wonky, so all the watchdog circuitry needs to be stand alone.

[david77]

That's a fair call!

So it was mainly to extend the reliability and usability and take logic operations outside the micro controller.

It may be interesting to do a project that required some logic, I must confess I have used a very small amount of it in a project I didn't design but it would be good learning to try and use some hard logic. Any suggestions on a good project that would be good to use external logic?

Thanks Mate

Lee

As has been explained already there was a time when there were no cheap and easy to program micros to make things, so one had to use discrete logic. For many simple logic functions even today it's easier and cheaper to use a few logic gates.

In 2007 I made this board that only latches & debounces three pushbuttons and an up/down counter with two more buttons that can count from 1 to 8. I used it in my home brew preamp for my stereo system. The up/down counter lets me select inputs 1 to 8 and the other buttons control things like tape monitor and mute. The board controls a bunch of relays on another board via a four wire BCD connection.
Doing it with a micro would have been much much easier but I really learned something from it and am now reasonably confident in using discrete logic. Btw. this thing is still in use nearly every day 7 years later.

[coppice]

A handful of logic chips is cheaper than a microcontroller, and they work in parallel execution as people have already mentioned.

If you don't need speed, a 25 cent MCU can replace a *lot* of simple logic chips. If you are paying, say, 6 cents a pop for 74HCxx devices, you'll break even at just 4 chips, and changes are wayyyyyyyy easier to implement cleanly in an MCU.

And, the 7400 or 4000 series logic IC (although, several have died over the years) you will be able to buy 30 years from now, while your DS80C310-QCG/T&R-ND might not be able to buy tomorrow.

Perhaps. I wouldn't guarantee it, though. "Several" is a massive understatement. Most of the logic families that were common 30 years ago are no longer available - ECL, RTL, DTL, etc. Only a small part of the 74xx range and 4xxx range can still be obtained, and you can't generally get the originals. Being original might be important if you relied on a quality like output drive. Only the simpler 74xx and 4xxx parts can still be obtained. The more complex ones are no longer useful, and so have disappeared. People still need simple stuff as glue, so there is still a huge market for those.

I agree that logic tends to hang around longer that most MCUs, but you can still gets clones of some 30 year old MCUS, like the 8048, often still implemented as NMOS. I think someone is still making TMS1000 clones, which was the very first MCU in the 1970s.

[T3sl4co1l]

ECL is still alive and well: http://www.digikey.com/product-detail/en/MC10H101PG/MC10H101PGOS-ND/1478129 for instance, original 10k ECL.

Not sure what all variants used to be around that have since fallen out of favor...

I don't think RTL/DTL is any loss, as they were largely compatible with TTL(??), and simply technically inferior (relics from a time when every transistor was precious).  If you need to find replacements that meet timing and pinout, have fun with that, but the functions and signals haven't gone away.

Likewise with almost all 7400, which were inferior to LS for the most part.  Some things, like 7406, don't make sense, or can't be made, in other families (30V outputs!), so they stuck around.

Most of the esoteric 7400 have died off, and weren't remade in LS.  All the important (and compatible) 7400/LS were also pushed into HC (and the related HCT, AC, etc.), as well as many of the cooler 4000 series functions.  And then LVC and "Tiny" families and so on.

I wouldn't say it's so much a matter of things dying off, as things being streamlined.  Requirements are always changing, so the demands of glue logic vary a bit, but the basics obviously have weathered well.  The same is true in general at all levels of computing: from glue logic to 8-bit micros (4 bit, even!) and on up the ladder, the lowest level requirements will never go away.

Tim

[coppice]

ECL is still alive and well: http://www.digikey.com/product-detail/en/MC10H101PG/MC10H101PGOS-ND/1478129 for instance, original 10k ECL.

Interesting. Motorola killed them. On must have found a application which justified them being brought back.

Not sure what all variants used to be around that have since fallen out of favor...

I don't think RTL/DTL is any loss, as they were largely compatible with TTL(??), and simply technically inferior (relics from a time when every transistor was precious).  If you need to find replacements that meet timing and pinout, have fun with that, but the functions and signals haven't gone away.

DTL is only compatible as a fudge. Its thresholds are different from TTL, and mixing them doesnt' always work. DTL definitely isn't a drop in replacement. RTL was not compatible at all.

Likewise with almost all 7400, which were inferior to LS for the most part.  Some things, like 7406, don't make sense, or can't be made, in other families (30V outputs!), so they stuck around.

You can't necessarily use an LS part to replace a non LS part. The drive is lower. This isn't so relevant to the 30 year period being considered, as  LS had already supplanted non-LS by then.

Most of the esoteric 7400 have died off, and weren't remade in LS.  All the important (and compatible) 7400/LS were also pushed into HC (and the related HCT, AC, etc.), as well as many of the cooler 4000 series functions.  And then LVC and "Tiny" families and so on.

Most of the esoteric stuff only existed in LS form, and has gone with the wind. An HCT or HC part may of may not be a suitable substitute for an LS part. The drive levels differ. The thresholds may differ. The switching characteristics certainly differ. Only the essential logic function persists through these families.

I wouldn't say it's so much a matter of things dying off, as things being streamlined.  Requirements are always changing, so the demands of glue logic vary a bit, but the basics obviously have weathered well.  The same is true in general at all levels of computing: from glue logic to 8-bit micros (4 bit, even!) and on up the ladder, the lowest level requirements will never go away.

The bottom line is its pot luck whether your favourite part will be around in a suitable form in 30 years. You can tell that some stuff will clearly not be available in 30 years, because it will be utterly irrelevant. Saying some form of 74xx will still be available in 30 years is like someone in the 1960s saying the 6L6 and EL34 will be around in 30 years. They actually were, but only because of one application - guitar amps. Nothing else (such as audioxfiles) would have generated the volume needed to keep even small makers interested. It was just pot luck.

[SeanB]

TTL bus buffers are very likely to still be here in 30 years, just because they are so veratile and very common in equipment, as most microprocessor bus outputs are not capable of high current drive or long line driving ( over 20cm or 5 loads), so you find a LS244/245 or the 5XX series ones in common use. HC138 is also still common as a decoder as well, it is often simpler than using part of a logic block to split out the address lines or to save pins.

[nitro2k01]

The bottom line is its pot luck whether your favourite part will be around in a suitable form in 30 years. You can tell that some stuff will clearly not be available in 30 years, because it will be utterly irrelevant. Saying some form of 74xx will still be available in 30 years is like someone in the 1960s saying the 6L6 and EL34 will be around in 30 years. They actually were, but only because of one application - guitar amps. Nothing else (such as audioxfiles) would have generated the volume needed to keep even small makers interested. It was just pot luck.

Re: Audiophools and tubes: I wouldn't be so sure. Sure, the audiophool marker may be 10-100 smaller than the guitar amp market, but those are the kind of people that would easily pay 10-100 times more than the current market price. So it might very well have been economically viable even if the guitar amp market for those tubes didn't exist.

That aside, I think key factors here are how easy something is to manufacture, and what the demand for it is. A vacuum tube is an oddball device to manufacture today. it would obviously require its own manufacturing procedure etc. A CMOS logic chip on the other hand, is "just another device", as long as other devices with similarly specced silicon are being manufactured. As for the demand, I don't see the demand for some glue logic here and there going away completely any time soon. For this to happen, something would have to replace it. In the case of guitar amps, this would be DSP amp modeling, which is prevalent, even if some people swear by the original sound. In the case of glue logic, one way this could happen is that in 30 years, even the cheapest piss and microcontrollers are expected to have a small programmable logic area, much like we expect even the cheapest micros today to have things like an internal clock driver, at least some internal memory and some UART or SPI type communication built-in.

[tszaboo]

Perhaps. I wouldn't guarantee it, though. "Several" is a massive understatement. Most of the logic families that were common 30 years ago are no longer available - ECL, RTL, DTL, etc. Only a small part of the 74xx range and 4xxx range can still be obtained, and you can't generally get the originals. Being original might be important if you relied on a quality like output drive. Only the simpler 74xx and 4xxx parts can still be obtained. The more complex ones are no longer useful, and so have disappeared. People still need simple stuff as glue, so there is still a huge market for those.

I agree that logic tends to hang around longer that most MCUs, but you can still gets clones of some 30 year old MCUS, like the 8048, often still implemented as NMOS. I think someone is still making TMS1000 clones, which was the very first MCU in the 1970s.

Well, maybe the TMS1000 will be still copied, but the things made after, I wouldn count on it. The types of MCUs jumped from a dozen to several thousand, so there is just not time and effort to copy all. If you see, every time Dave opens something, he immediately finds the logic chips and knows what they are, while with MCUs, lot of times even the manufacturer is not around anymore.
The RTL appeared in 1961-2 and made it in space in 1969 (or around). By that time, even if this wasnt said, it was NRND, as there were several better families. So I would say that 30 years ago it was common, I would say 30 years ago designing something with RTL or DTL would have been a large mistake.
I'm not against MCUs. They are good when it comes to reducing cost, size, complexity often times.

[22swg]

The 74 TTL era was a great time to learn and  develop ideas , with a strip board and iron (fe)  a hand full of TTL / 4000 anything was possible . Having seen David77's  board picture thought I would post my 1980's interface,  I was learning to play an Elka E19 organ but fancied some better sounds from a TG100 midi box so I designed and built this board, it would still work but i removed the MAX232 for the next version an mcu ! no longer have the E19 ...  got a Tyros with my redundancy cash ! 

[westfw]

One of the nice things about the 74xx and 40xx series was that they were sufficiently popular and long-lived that many chips with the same functionality and pinout have been released in newer technologies.  You may not be able to find a viable source of 7400 chips, but you can certainly find other 74xxx00 quad nand gates in the same 14pin package(s.)  And the larger functions that have survived have been "proven" to have useful bells and whistles (presets, clears, combinations of enable inputs, etc.)

It's pretty much the nature of "logic" that it doesn't need to "stay around" in exactly the same form.  Most of the circuits you build with 74ls logic will do just fine converted to 74lvhc or whatever.  Yeah, you occasionally see circuitry that relies on electrical behavior other than boolean algebra, but it is relatively uncommon.  (Particularly, compared to a microcontroller design that relies on a certain instruction set (all of them?) or peripherals...)

[Zero999]

I felt really weird the first time when I had to use discrete logic. But after a while I find it that I'm using more and more. And surprisingly, most of the times when if is near analog circuits.
Like I had a board which had the LEDs on the front panel, MCU on the other side, precision analog in between. To keep the nosie down I had to buffer the signals driving the LEDs, because after several request not to, they still decided to make PWM for the LEDs.
When the NOT gate was there, it was only microamps flowing next to the analog instead of ten miliamps. This was just one example.
I've seen a (recently designed)power supply, where they had a CD4046 PLL to make the phase shift between phases. Genius usage. If you would go with a microcontorller, it just wouldn't fit on the board, because you need a power supply, level shifting back and forth... While the 4046 just run from the input ~12V.

I agree, one of the nice thing about the 4000 series is the wide power supply voltage range. Most MCUs will work from 5.5V (6V at a push but not recommended) so you need an additional IC for that, plus logic level shifting if it needs to interface with higher voltage circuitry.

[T3sl4co1l]

Not to mention the CD4007 (transistor array -- who needs gates in a logic family?), CD4049U (CMOS pairs -- inverters, unbuffered, no input protection diode to VDD -- vs. the '69 which is the normal CMOS design), CD4046 (PLL, osc with reasonable linearity, little or no dead zone in the one phase detector -- apparently unmatched by the botched variants 74HC4046 and 74HC7046) and more.  Needless to say, analog switches are cool, too.

[22swg]

Really useful are the analogue switches. 4067, 4051, and 4016  ... beginning to miss them already 

[SeanB]

4049/50 is great as a level translator to run inputs into 5V logic from 15v or so inputs, or from 12V from external switches. Also great on circuits where you depower parts and do not have to worry about them being powered via logic inputs, or dragging the logic levels down.

OTOH, who needs a MCU when you can use a board full of CMOS to do the logic work, like this Sony controller in a Umatic deck.