Do We Still Need Discrete Logic Integrated Circuits?

Hi redfred, I built systems using discrete 74F series which is almost gone now. I am collecting part to rebuild my ST2006 delay timers for education and training which is discrete 10MHz design.

Even 74HC and 74HCT are out from market. Last production of 100 such instrument likely to finish after I get parts in few months.I have built discrete state machines. I will look for replacement in VLSI for entire circuit.

1. Kinda "wasteful" to use a whole VLSI just to illustrate 'how' a half-adder circuit is built and operates.

2. Akin to wanting to abandon NUMBERS (1,2,3,4...etc.) since only ONEs & ZEROs are needed in all our computerized stuff.

Hi 70AARCuda, Yes, basic idea of binary can't disappear. However, lots of standard ICs are gone by now.

It is hard to find replacement for standard parts as obsolete list is breathing its last breathe.

VLSI ccan replace entire board on chip rather than each discrete. I have learnt few manufacturers have collected designs and can manufacture old parts on bulk order at high cost x5 to x10. Sure MOQ going to be massive numbers.

For reasons I no longer remember, other than the more flexible voltage source, I always preferred the 4000 series CMOS chips. Now that you remind me, I realize that it has been quite a few years since I ordered any. I still have a pretty good variety on hand.

Hi dkwarner, 4000 series CMOS LOGIC may be still there but soon going to be out of production. I think the original manufacturers have changed hands and sold their technologies and new owners if technology have changed the application and this has resulted in loss of the vast resource of discrete technology. Another good reason is that people no longer work with discrete anymore.

How can any individual CR4 subscriber guess, accurately, what <...we still need...> (rhetorical question - NNTR)? That is a job for the marketplace!

ICs have two main advantages over discrete circuits: cost and performance. Cost is low because the chips, with all their components, are printed as a unit by photolithography rather than being constructed one transistor at a time.

Hi kumarpradeep, discrete parts are usually deducated LOGIC or counter, registers, encoder, decoders etc. 100MHz usable speed. VLSI usable speed can be 500MHz plus.

Yes.
Next!
Del

I have a hard time believing that Discrete logic chips are obsolete. There will always be a need for these in a variety of situations. Whether it's a company customize process equipment or a tinkerer on in his garage.

Not only the tinkerer, but also the newbie/student. I've always been a very hands-on learner/doer. I have a very difficult time imagining how to learn the basics without building quite a few circuits on one or another form of protoboard.

A simulator on a computer would obviously help, but I'll remember a circuit much better after screwing up a couple of times as I build it, then finally getting it to work. The more senses involved in the learning process, the more permanent will be the memory.

I'm pretty sure that a student will more permanently remember to check the polarity on an electrolytic cap after seeing, hearing, feeling, and smelling one explode when he/she has put one in backward.

Even if no errors are made (unlikely in my experience) while building a circuit, the tactile experiences of building it will help strengthen the memory.

My first college class in this field was Digital Electronics. Great way to learn fundamentals of digital electronics, with J-K flop-flops, One Shots, Counters, Clocks and inverters.

Its a thing where I don't believe (from experience of hiring) that Colleges aren't teaching anymore,... and that is fundamentals.

And the reason is,... we have software that does it.... What?

"Its a thing where I don't believe (from experience of hiring) that Colleges aren't teaching anymore,... and that is fundamentals."

I think you need to either remove that "don't", or replace that "aren't" with "are" (currently double negative), but I'm sure you are lamenting the lack of understanding of fundamentals by many students, and I completely agree.

The same is true in other fields besides electronics. An ME who has a great theoretical background, but has never actually used his hands to construct anything, is close to worthless in my opinion.

I think you need to either remove that "don't", or replace that "aren't" with "are" (currently double negative),

thanks, You're correct.

I've have been hiring mainly in the mechanical engineering sector, is where I saw this... But I always took Machine Control electives just do to the fact it complimented my skills for process control... The students that took the Digital and Electronic courses, when they started they were quite... boisterous, we had over 70 students, class/lab size were limited to 30 so they added another class and split us up. And the instructors assume we'd have a few drop in the first weeks from their experience.

When the semester was over, there were 7 left in our class initial class of about 35. I would assume the other class had the same ahhhh 'mortality' rate.

That is what I do and teach but discrete is problem now. Many students want LEGO like solution, robots, drones, wireless sensors and contrils etc.

Industrial electronics is no longer a fun. I consider basic LOGIC level work very important though simulation tools help working on VLSI.

Electronics require knowledge physics, chemistry, mechanical, electrical, electronics and software engineering all in one. It isn't easy. China can do it much better. Like a ready packed food for others.

Good point. I am teaching.

Shyam, you've mentioned several lines of discrete logic IC's that have gone obsolete, but these have typically been replaced by newer lines that are at least as good and sometimes better.

74F series logic was always a heavy power draw series and typically ran hotter than the CMOS devices that are still available at similar or better speeds.

I will admit to having seen an application where a 74LS device would work, but an equivalent CMOS device wouldn't simply drop in. If I recall, it was being used as an oscillator, so the slew rate and propogation delay of the equivalent CMOS part required a change in RC values as well. BUT, the majority of simple logic IC's will cross from Schottky to CMOS just fine.

As for discrete logic in general, there will always be a need for limited circuit complexity that individual gates will fill very well.

I'll also highlight the individual gates that are available today, i.e. a single inverter in a package, typically a six pin SMT device. This enables the designer to avoid putting a a larger 74xx series device with 4 or 6 gates when only a single gate is needed.

Hi Lo_Volt,

I did use DTL obsolate, 74 general, 74L, 74LS, 74S, 74F, 74C, 74HC, 74HCT, 74ALS, 74LVS, 74HCF etc and then ECL too.

Speed and power are two basic concerns other than operating voltage range and signal voltage ranges.

Both TTL and CMOS were easy to use. However, ECL and new differential LOGIC are great at speed, but very difficult to design with on breadboards.Most of the manufacturers have soldout entire companies now. I find that Rochester is stcking as well as manufacturing on order. Rest of them either have old parts, pulled out parts, or even fake dummy printed number parts.

I have stocked large number of larts for students to get first hand idea of LOGIC on breadboard to play. VLSI are only for experts even to figure out exactly which way to design.

This reminds me of my semiconductors class that I took in 1986. At one point we were studying the Hall Effect. The accompanying lab had us use a small chunk of germanium to illustrate the effect. However, the lab instructors were finding it increasingly hard to purchase it. It was widely used for transistors in the 1960's and 1970's, but advances with silicon pushed germanium out of the market. With no one using it commercially, the school couldn't find sources for it. I think that the school dropped that particular lab after their stock of it ran out.

Two things that I'd like to ask of you as an instructor:

First, don't forget to highlight the fact that any logic signal needs a ground reference. I've seen a few newbie engineers design a discrete output with one wire! They've completely forgotten that they need a signal reference as well. Working on breadboards, they never realized how necessary the ground is.

Second, stress the importance of de-coupling caps for logic circuits. Again, I've worked with engineers that are fresh out of school who have never heard of de-coupling caps. With the edge rates of today's high speed parts, capacitors are even more important.

First, don't forget to highlight the fact that any logic signal needs a ground reference.

good point,... I forgot about that.

Yes, good clean power soyrce and ground reference are needed for LOGIC boards though they do tolerate marginal noise.

There is another problem I face is called ground bounce current when too many outputs from one IC draw current simultaneously from ground pin. Signal bounces as high pulse current forms significant voltage pulse that mat exceed the allowed limit. This happens due to AC current in stray capacitive loads.

I have collected good number of LOGIC IC that I can give to students and looking for more in series to have a variety for teaching experimental electronics.

I also have fuse type PROMS but these require old types programmers to blow fuse inside. I am not sure I can get any of them now in the market. I need to make one. 74S472 type etc.

Texas Instruments still manufactures a HUGE number of discrete logic chips. If you only need a few gates to accomplish your desired function, why would anyone use a machine gun to hunt mosquitoes?

In fact a recent design used CD4000 family of gates because they would take care of the basic logic functions needed and operate up to 20 VDC and meant we didn't have add another power supply bus voltage.

There is another world out there that doesn't do everything in the digital domain.

Hi Brave Sir Robin,

I hhave tested Texas and National Semiconductor parts and they are inferior to Signetics, Phillips, NXP, Motorola parts. IDT was the best.

74F I use have 3ns-5ns delays. 4000 series has delays of 50ns-100ns are good for limited use. 4000 series is good for low power and high range of operating voltages. I am using only few ICs from 4000 series.

Good idea.

the education systems still use and need discreet logic chips. some are getting very hard to find at a reasonable price.

It will always come down to simple economics. When the demand drops to a low enough level, the parts will no longer be made.

On another view from economics, why employ a PLC when some simple discrete logic will solve the problem at a fraction of the price and minimal complexity? Paying more for a solution than necessary has never been popular.

In principle, I agree. However your simple economics does not capture many key points. Often the one choosing between simple or complex technology is not the one paying for that technology choice. Also the one paying for the technology rarely initially understands every critical attribute so a fixed hardware logic can be effectively implemented without expensive future modifications.

It's not only the one paying for the technology that rarely understands every critical attribute. All too frequently the one who sets out to program the PLC technology fails to understand the inherent short comings. At least with wire wrap techniques, fixed logic can quickly be altered. Been there, done that.

Usually I'd say that changing a program is easier than doing a little rewiring but that would fail to take into account the pain in the brain that is necessary just to get to the point where one can write or edit a program. Wait, are we saying the same thing?

Rewiring required maintenance of wiring diagram and modification in the hardware itself which meant that two sets of hardware finally became different after each modification though they were same initially.

While software change keeps the hardware design identical. One stocks same hardware but programs in a file to be copied.

I won't like users to make changes in my hardware accept for few sets of jumpers listed in the catalog that can configure hardware at best.

I did wirewrap boards for quick assembly and these featured like multilayers as too many wires one could wrap around logic. If rat bytes a wire then it was hell.another problem was maximum usable frequency to about 50MHz. I did make many such wirewrap boards from 1974-1990.

Now PCB design software are much easier and cheaper and OCB are reliable and for this I prefer PCB.

While DIP IC sit in sockets and can be reused, SMD IC are generally for final board space saving purpose.

Makes great sense.

What is the alternative way?

CPLDs?

Not sure but if there is need then production can restart as design are just design files.

https://www.youtube.com/watch?v=OpLU__bhu2w

Cleaner Large Scale Distributed Amplification and Saving Wire and Power

There may be zero people doing this but... The jellybean circuits may be a great way to implement distributed amplifiers. Distributed amplifiers can present a really tough wiring problem if one uses a centralized boatload of logic with loads of pins all in one spot. Distributed amplification can boost the frequency and bandwidth at which an amplifier can operate relative to the inherent speed of the technology. The number of active circuits used enhances the effect. Wiring to centralized points is counterproductive since it explodes the wire count and decreases the density of the transmission line(s) possible in a specific space. The extra wiring also will be profoundly likely to couple electrically with the energy in the transmission lines prompting a circuit designer to resort to shielding, differential twisted pairs, etc. to minimize coupling thus compounding the problem of too much other stuff near the transmission line(s).

Cellphones and other devices operating at frequencies high enough to integrate the transmission lines on chip can use lots of circuits geographically distributed on chip and minimize the excess wiring issue with those techniques. A tinkerer wanting to build a Beverage antenna with an integral distributed amplifier might be better off building with jellybeans as opposed to pure discrete. It also seems like a potential way to experiment with energy gleaning techniques to power the jellybeans to astound the QRP community.

Does anyone have the foggiest idea what I am talking about or am I the only only one who spent a lot of his childhood reading ARRL and owns the three volume set of Oliver Heavyside's books ? Has anyone out there considered running a toroidal Tesla Coil with a geographically fixed Shah standing wave ? I fully expect crickets to these questions but being the eternal optimist there they are. I considered trying to explain how all this relates to jellybeans but everyone who does not already understand would just go, "Huh?" and give me an undeserved off-topic.

If anyone understands everything I just said, I challenge you to expose your nerdiness to the other readers by contrasting lossless transmission lines with distortion-less transmission lines and how that might relate to distributed amplifiers. Then propose a multiple 74HC595 and multiple 74HC14 (ferrite free) toroidal coil design to support a binary chain code in one direction with adaptive{microprocessor(s) ok here} techniques to keep the code from moving each time it circumnavigates the toroid. The prize goes to the lowest power achieved using (obviously ) adiabatic circuit techniques on the longest chain code stable for, say, a day in a stable setting. Civil engineering, Geology, and Archeology can make practical use of this circuit. Honorable mention to those who most clearly explain why.

thewildotter

911