I think you are missing the point. When people want a microcontroller, they typically want to REDUCE part count and size of end product. Suggesting to them to use 3 microcontrollers with a mickey mouse undocumented protocol handling inter-controller communication, then trying to get them to program each one separately is... insanity!

This is the digital equivalent of people from third world countries continually show up every month or so trying to make a 12VDC -> 120/220VAC Inverter out of 2 bipolar transistors and a transformer, and complaining it doesn't work. They insist they must make it work because "those are the only parts they can get".

People pay upwards of $100/project when working with Arduino, and you are saying that 3 processors at $0.40 each (plus wasting pins for inter-ic communication that is bit-banged) is "Better" than a single controller with far more functionality for twice the price of the combined cost for 3 separate controllers, when samples are FREE!

It may work for some people while it may not work for others.

I can do many interesting things with these controllers. I also know their limits well. And I have more powerful chips here which I can use if I wanted to.

There is no urge only to use these static LCDs. Absolutely not. But they actually need 24 I/O bits each. So if you have two, you need 48 I/O.

Multiplexing them is bad for the PCB layout, so I actually came up with idea to use two controllers. The PCB layout is much better all the multiplexing tracks gone.

Yes Mickey Mouse protocol which no one has ever heard of. You are absolutely correct.

I will even do something much more absurd with this I/O board soon. There is a 16f946 with 53 I/O controlling a 68000 but only using 8bit RAM.

And I will have a small USB PIC to load up programs. Optional, while also EEPROMs can be used. Speed is not very high indeed. It will absolutely be able to run 68000 code. I/O on the 68000 itself is done via unused address bits.

Maybe OP can decide- to give PICs a try to emulate digital logic. It does not have to be a 16f59- there are others indeed with serial port if so desired. It is a challenge to design a serial protocol in software which is A) useable, and B) not too complicated to program and to use such as fixed bit timing.

Much better investment of time than to memorize and study all the pecularities of flip-flops, DRAM drive and latching and who knows what else. PAL/GAL perhaps as well and EPROMs. They are indeed no longer manufactured.

The problem with digital logic is that the functionality is so much bound to the hardware, and you need a reasonable stockpile to build any meaningful circuit.

If you emulate the digital chips, you typically only use one type of chip all through.

 

So you hobble together a series of serial information, clock and ready lines just to do combinational logic!? Get a CPLD/FPGA and be done.

It is not so much hobbled actually, it is well designed for it's purpose of doing serial communication with HID and small displays.

I don't use combinatorial logic why would I want to? If I need higher speed, I choose a controller with more MHz. Not that I would not know about it. I even have a small stockpile here some 100 ICs I have a CPLD board as well.

So you understood a XOR gate the first time you heard about it?

From Don Lancaster's book- absolutely. It is a controlled inverter, kind of.

When I don't understand something properly, I read it again some day.

That is part of designing a circuit-knowing the different trade-offs of a certain device and understanding if another is better suited for the task....

Try primitive...

It is certainly not wrong to study the work of other people. What is not so good is this approach "It must be done that way there is no other method and this particular obscure chip must be used as well." Or even more absurd, to limit people only to use a certain functionality (educational). It is utterly useless as such.

 

It is not so much hobbled actually, it is well designed for it's purpose of doing serial communication with HID and small displays.

SPI is a well defined serial interface... Yours is a small implementation of it.

I don't use combinatorial logic why would I want to? If I need higher speed, I choose a controller with more MHz. Not that I would not know about it. I even have a small stockpile here some 100 ICs I have a CPLD board as well.[/QUOTE]
....What happens when you need to do a 16-bit multiplication in less than 100us?

From Don Lancaster's book- absolutely. It is a controlled inverter, kind of.

When I don't understand something properly, I read it again some day.

You said you were taught this information before you read this book, so the answer is no. By your reasoning, you have/had the

logical thinking abilities of a very young girl

It is certainly not wrong to study the work of other people. What is not so good is this approach "It must be done that way there is no other method and this particular obscure chip must be used as well." Or even more absured, to limit people only to use a certain functionality (educational). It is utterly useless as such.

... and your argument of using a few decades old chip is different how?

It is a way of thinking, that's taught, something that must be learned from the ground up.

 

...if I run into such questions, maybe I hire a CPLD expert. I am not specialized for this tech., even if I have some related components here.

It is a question of the mindset yes.

They told us, "You memorize all the pecularities, get a good degree, and get a good job".

Which was a lie. No one is asking about schoolbook digital logic and MSDOS softwares these days as well antiquated rack computers.

Without knowing microcontrollers seriously as well CPLD, Verilog and things like that, you won't get any job in microcontrollers engineering.

If you mention the technology used in college, your letter will not even give you an interview nor a reply.

Never mention homebrew projects either especially TTL CPUs.

Don Lancaster more maintains an approach- take a look the technology, build something wonderful. The limit is your creativity.

I have not often seen such a knowledgeable person.

And also what he maintains if you are not that fancy of a design, do something new or something else or do it differently.

In college you must exactly answer the obscure questions to the point- which was instilled literally previously. All that knowledge is for the mental dustbin.

Could you memorize 500k pages PDF that way? I don't think.

 

Could you memorize 500k pages PDF that way? I don't think.

No, but as I said, they teach you how to approach a problem. If you know where to look, you don't need to memorize things. If you know how to think, you can adapt to the change in technology.

Without a solid understanding of the basics and why/how they work. There is no hope for someone that doesn't get that. They are limited to the technology they know and will not be able to grasp the concept of the new.

 

Maybe 74xx will stay with us forever- including hordes of eager students putting large efforts into fabrication of a plug board, maintaining drawers as well, using sockets, labelling all these ICs as well.

I never use sockets, have thrown all 74xx into one drawer- that's it pretty much.

When I am asked to build a meaningful circuit from Don Lancasters book, or some circuit related to it, I am forced to take a look all these flip flops. Then I choose one that suits me.

The end result should be a person able to design meaningful circuits independently- not even using books at all.

If there is someone telling me he's a C++ programmer but is not able to design a complex digital circuit- my thought is that person is not really a C++ programmer. It is someone who maybe changed a few lines within some scripts, and who made some C++ stuff working from a book, and who maybe even wrote some small programs independently.

I saw some programs written by college students. Embarassingly primitive. (My BASIC stuff when I was a teen certainly was not better, only thing is I was much younger, and I pretty much programmed what ever I fancied).

The only way to learn all this software and technology is to work hard with it, and to build many circuits, and to read many documents.

Without hard work, nothing will result. What you actually do is really secondary, as long as you are constantly challenged, and you constantly try new things, and you always try to improve.

It's like using wall adapters + jacks + 330 Ohms LEDs, which is often OK, sometimes it is just dumb, using a 3v button cell + 4.7k LEDs is smarter. Or no resistors at all perhaps

Yes there must be some good colleges and schools, some knowledgeable people, and some good teams.

 

I think you are missing the point. When people want a microcontroller, they typically want to REDUCE part count and size of end product. Suggesting to them to use 3 microcontrollers with a mickey mouse undocumented protocol handling inter-controller communication, then trying to get them to program each one separately is... insanity!

I use them as replacement for serial buffers. Which for instance are needed for static displays or if you need a lot of I/O.

I don't use any additional chips. Using individual I/O is the only way I know to drive static LCD.

So, insanity? I am reducing chip count. My point is to replace stockpiles of discrete logic with 16f59.

If you need high speed use a CPLD correct.

At first I only had one 16f59, multiplexing it for 2 displays. Many PCB tracks. They are so much low cost, a smaller PCB with less tracks is better than to save that one extra controller.

Dealing with one large TQFP100 is a different thing, they also can drive all that stuff with one chip, but aren't available in that price range. They are usually quite sophisticated. And not so easy to deploy as a TQFP44.

It depends if you have a large production run, and you want to optimize it, or if you want to produce a small batch economically.

True is that also small TFTs are cheap now- 6 dollar. And a 2 dollar 18F can certanly display fonts and small graphics as well.

But- think a little. These TFTs are more sensible. They can and will fail some day statistically. User-replaceable? These static LCDs are, and they are also quite rugged.

If you have two controllers and one fails- one still works?

In a cold environment, you would maybe use LED displays. If you don't know if and how often it may become serviced.

All devices have their special niches where they prevail and predominate.

Why would you want a wireless network for a small office complex, where a 2-wire line and two cheap handsets will be all what's needed?

I am certainly never saying, only use this or that technology, and I never refuse new technology.

 

It may work for some people while it may not work for others.

I can do many interesting things with these controllers. I also know their limits well. And I have more powerful chips here which I can use if I wanted to.

There is no urge only to use these static LCDs. Absolutely not. But they actually need 24 I/O bits each. So if you have two, you need 48 I/O.

Multiplexing them is bad for the PCB layout, so I actually came up with idea to use two controllers. The PCB layout is much better all the multiplexing tracks gone.

Yes Mickey Mouse protocol which no one has ever heard of. You are absolutely correct.

I will even do something much more absurd with this I/O board soon. There is a 16f946 with 53 I/O controlling a 68000 but only using 8bit RAM.

And I will have a small USB PIC to load up programs. Optional, while also EEPROMs can be used. Speed is not very high indeed. It will absolutely be able to run 68000 code. I/O on the 68000 itself is done via unused address bits.

Maybe OP can decide- to give PICs a try to emulate digital logic. It does not have to be a 16f59- there are others indeed with serial port if so desired. It is a challenge to design a serial protocol in software which is A) useable, and B) not too complicated to program and to use such as fixed bit timing.

Much better investment of time than to memorize and study all the pecularities of flip-flops, DRAM drive and latching and who knows what else. PAL/GAL perhaps as well and EPROMs. They are indeed no longer manufactured.

The problem with digital logic is that the functionality is so much bound to the hardware, and you need a reasonable stockpile to build any meaningful circuit.

If you emulate the digital chips, you typically only use one type of chip all through.

I'm glad that is working for you. You are taking my first comment in the wrong context, however. If you have built it, it's your system, it works, and you are happy with it, good job. It may not be optimal or production ready, but you completed a digital hobby circuit. You enjoy that, and I am not bashing you for that.

I am ONLY referring to suggesting to OTHER PEOPLE that ask on this forum what to use for "Starting with PIC", to suggest a "General Purpose" PIC, such as 16F690 or 18F2550 (the latter is an Excellent uC). The 16F59 is NOT a useful recommendation for other people new to PICs. That is all I'm saying.

You can do what you like connecting whatever you want, but keep in mind that just because it works, doesn't mean it is good design practice. (Just like driving LEDs without Current Limiting)

We are here to help other people get started in fun and interesting worlds in electronics. Once a user tries all the published examples for the common general purpose chips, they THEN have the knowledge to decide if they need either a less or more powerful controller model.

Knowing how to add a FPGA and sometimes CPLD is a fundamental today. Sometimes a few very fast responding Combinational Logic Block (CLB) can save a TON of controller time, especially in end use applications that are essentially Finite State Machines.

Verilog is fairly easy to learn, try it out, download Xilinx ISE and look at the examples, which will also simulate without hardware. There are even drop in IP cores for microcontroller (8 bit picoblaze to 32 bit microblaze), where other IP blocks, such as I²C, SPI, DDR, etc. can be added, then you do all you can with combinational logic, as it runs at gate speed. If you come to something that needs processing, the picoblaze/microblaze soft core controller can interface. The code for the controller is written in C separately, just like programming a PIC.

Altera uses Verliog and AHDL rather than VHDL/Verilog that Xilinx use. Altera have the NIOS II soft core that can be "dropped in" to a design.

The "Game Shield" for Arduino is based around a small FPGA with embedded controller. Same for the $50 32 bit USB Logic Analyzer on Seeed Studio. They aren't ALL BGA or other hard to use packages, many are available in flat packs with leads.

If you know digital logic, and a programming language, learning FPGA isn't a huge learning curve to build a DSP system of arbitrary sized bit width (Provided you have/understand the math/formulas for the DSP itself)

CLB are becoming more common, to the point Microchip integrates a couple in their controllers as a concession to reality.

 

Yes I have 16f1503 here with CLC, is that what you talk about?

On a sidetrack your information is also worth reading and useful for me. I only write what is in my mind instantly, which means this is never statements or points of view set into stone.

The 16f59 is so easy to use you don't loose a big effort if you do, and having a box of these around does not cost you much.

My circuit is actually a prototype for a real PCB which in my opinion would be production ready, once all open questions are answered and all issues are tested on the prototype.

My circuits often need some kind of display, but the infrastructure that comes with modern graphic displays is not always required or even wanted.

The problem which I have not solved until now since I have this prototype is to be able to use any kind of static LCD or multiplexed LED display with one PCB, bury the decoding into the firmware, not into the PCB tracks. Which are really connected to the display port 1:1 as to reduce complexity.

It is sorted out via a staggering scheme of decoding tables.
The data for the segments or columns or whatever is combined from several tables for each phase, the ORed together for each port, then the relevant ports are updated.

For 3-digit LCD it is not that much of data tables. I think I could pack data for at least 3 or 4 different display into the 2K, in addition to the serial interface.

Take a blank PCB 2 dollar cost, take 2 TQFP44 2 dollar cost, FLASH via ICSP, take 2 displays 2 dollar cost. 6 dollars and the work of minutes and can use any of these displays. Even LED matrix is possible.


Character LCDs yes- small, take a lot of current, don't light in the dark until you power the backlight and and and.

So it is not a circuit or a design irked together for a hobby fun. It is the current stage of years of building such circuits and improving them. And spending a lot of money just on this display technology.

If I wanted to build a plugboard for a student, I would use FLASH controllers certainly. Why limit experimentation to 4 gates with 2 inputs each?

If they move on to microcontrollers then, they maybe already have experienced the way how a microcontroller is flashed, how it is clocked, and all this.

And why don't get rid of these big power supplies from the school lab altogether? Do it with small batteries.

For sure the 16f59 is pretty dated and does not have much capabilities (except the software which you program).

But what do you loose? Looking at 50 pages PDF? It is certainly obvious that more powerful PICs do exist.

Why do people always ask these questions again about ANSEL and they don't actually read the information and can't do it right?

See? For such people the 16f59 is a solution making the first contact a more pleasant experience.

 

If they move on to microcontrollers then, they maybe already have experienced the way how a microcontroller is flashed, how it is clocked, and all this.

That is THE rub.

CLB/CLC/LAB (Xilinx/Microchip/Altera names for combinational logic blocks) allow essentially instant change of outputs based on inputs. No clock is required, which allows some things to be done VERY quickly relative to a clocked microcontroller.

Once you fully understand the difference between State Machine that can be hardwired in logic, and something needing a computational module (uC or uP w/Clock), you'll "get" what I'm trying to tell you. It is amazing what can be broken down into combinational logic when using verilog to modularize the "System", so when inputs go a certain way, the results of a function show up on the outputs in nanoseconds rather than milliseconds. Speed and low power (efficiency) are the future, learn it now as you can, it's never too late to start.

Most all of what I work with now didn't even EXIST in College, some was theoretically possible, but the rest was unknowable. If you've had told me I'd have a 1.5Ghz phone that worked globally, had a 32 bit computer, and was 1/8" thick and 3"x5" with a 1080p display and 64GB of storage, I'd have told you that you've read too much sci fi. 16Mhz and 128k of RAM was "Fast computer with lots of memory", so Giga- was used as a joke, literally.

 

Yes all correct. But normally the teacher would present the board already flashed with firmware for certain logic gates.

So the content would be to learn combinatorial logic, not the microcontroller as such. True once it is there for sure it is mentioned on a sidetrack.

It is cost saving- do away with these 74xx stockpiles.

Or maybe it must be maintained on the effects coming from poor students. For a thousand pound you can get a decent amount of big 1000pcs bags with electronic components, a big box with controllers, a soldering station and a good DMM as well. Why use it just for digital logic? Money is tight these days and why not use it so you get the most out of it.

Why pay 10x the price to people who don't add actual value to stuff?

Rich college students maybe don't bother, but a 1000 pound is still a lot for an average student, and it can be invested far more effectively.

The question if 74xx will be maintained forever is not upto me to decide. For sure some common parts will continue to live on, while many already have disappeared. Only a few large distributors still offer a relatively large selection of these- and virtually none of them is offering HEX decoders.

Digital circuits which can not display HEX are stupid.

I saw a DELL PC with tons of small logic ICs buried on the PCB not too long ago, a Windows 98 or WinXP PC, not interesting to keep, but still used professionally in the late 1990s or even early 2000s.

But it is only a subset of these chips which still has a meaningful customer base, the rest is more or less manufactured on purpose and stockpiled.

If I look at eBay and see some vintage stuff for moon prices, I sometimes do not know who is buying this, or what should I think about such an economy. For sure in America there is freedom to maintain any capitalist enterprise and another one's business never must be disrupted or commented negatively.

Just there is so and so as much money and oil reserves and it could be used in way getting much more out of it.

As for the simple GALs, Lattice stopped manufacturing them alltogether a few years ago.

The semiconductor industry for sure can afford to continue to offer 74xx- just for 1000 pounds there is a better use than to bury it into a metal frame logic board.

Also mind it that cutting cost has become a key technology these days.

 

Yes all correct. But normally the teacher would present the board already flashed with firmware for certain logic gates.

Uh, No. Students in Digital Design get a high-midrange FPGA board, with Lots of peripherals (Such as DE2-115 by TerAsic), for around $250. It has a self-test built in. If it is supposed to do anything else, the students code it from scratch and build on that, some classes making fractals with VGA display, some DSP, and many other functions far out of range of a microcontroller. It's not just speed, it's the ability to have massive parallel processors working together, with 4 different clocks, and multiple phases of those clocks, resulting in extremely high throughput.

FPGAs lose their program every time power is lost. They are programmed from NOR Flash at power on, which doesn't last as long as NAND Flash (Flash Drives), only a bit over 10,000 cycles. This makes getting the design as close to correct as possible in simulation before using a bit more important, the next step is simply loading the bitstream onto the FPGA while it is powered on, and it will run the logic given until power is removed, or a new bitstream is dumped in.

 

Yes I am aware some people use amazing stuff. Does not mean it is worldwide standard because it is common somewhere locally.

We were exposed to crap 10 years to 20 years back in time at a slow pace, even shown some stuff which never was even powered up just long explanation and paper questions.

Not even independent projects back then.

If I am a student in Cowabunga I would think twice to save for years for 1000 pounds or to use something small scale and which costs 6 dollar. Be it ready made from Texas Instruments, or self designed like my LCD circuit.

I saw some execptional bad PIC stuff over the years, some really bad, some just burying a lot of functionality which is never used just looks nice.

If we talk about amazing technology, I want to mention parallel computing on 3D cards- there is a lot of software, downloads, and whitepapers. A brand new world!

Question of time until some clever person builds a standalone board for graphics cards. Cheap since they are mass produced. Maybe there is no market now but the capabilities are amazing.

Maybe this was what this Raspberry PI guy was thinking. Use the existing technology from 3D cards and scale it down to a single board computer.

What's the point to take a chip which costs less than $1, and flash it with some lines C, and use it as combinatorial logic? Threshold fear maybe. You don't have much to loose neither time nor money trying out PICs for this purpose.

Even if you only learn C programming, you also learn digital logic- a fundamental of programming as such.

 

Maybe this was what this Raspberry PI guy was thinking. Use the existing technology from 3D cards and scale it down to a single board computer.

Huh? The Raspberry Pi was originally designed around an ATmel Microcontroller, then moved to the ARM "System on Chip" for production, same CPU/GPU used in some smartphones and other systems. The ARM CPU/GPU is designed in Verilog, then made into masked hardware after FPGA prototypes are verified.

Just because in the past, technology wasn't available to hobbyists doesn't mean one should ignore it now that it is in easy reach. There are several online courses for Verilog on YouTube, no hardware needed. Icarus Verilog runs on a PC and is completely free, capable of whatever one can come up with, which can later be modified to be synthesized on an FPGA.

 

Huh? The Raspberry Pi was originally designed around an ATmel Microcontroller, then moved to the ARM "System on Chip" for production, same CPU/GPU used in some smartphones and other systems. The ARM CPU/GPU is designed in Verilog, then made into masked hardware after FPGA prototypes are verified.

Just because in the past, technology wasn't available to hobbyists doesn't mean one should ignore it now that it is in easy reach. There are several online courses for Verilog on YouTube, no hardware needed. Icarus Verilog runs on a PC and is completely free, capable of whatever one can come up with, which can later be modified to be synthesized on an FPGA.

So it is single core only? Doesn't it support OpenGL some kind of? I was thinking this is all multicore based like modern DirectX. I had to learn a lot about recent developement and DirectX changed constantly from version to version.

What does a noname entry graphics card cost? What is the computing need of most users most of the time? So why entertain a complete PC with parallel bus and all that.

What chips to buy (OP). If you want to build anything else than trivial and useless circuits, you need maybe 200 ICs or so, together with a storage rack. Only to figure out having them sitting around after a while and not that useful.

If you wanted to, download the old WinCupl from Atmel and use that to simulate digital logic.

 

There are multi-core ARM, but the Pi is a single core ARM processor. The GPU/Graphics Processor is on the same chip as the CPU, bus, DSP, USB, and RAM (System On Chip).

The 3D Support is a subset of OpenGL that doesn't have the toolkit or library functions onboard. It's the same OpenGL that mobile phones use for 3D graphics, which is still fast. The full OpenGL spec supports a ton of stuff, but the extra complexity needs more power and isn't an economically effective gain.

 

Yes interesting. I have a STM32 board since a few weeks with 170 MHz.

Mind it the 16f59 chips on my PCB are just commodity processors- they actually sit under the displays.

I am actually considering to change over to PIC32 with a few exceptions. It is more complexity- they need capacitors, a USB socket normally and a small switcher IC.

While other PICs nearly don't need any extra parts at all.

Just one coil is totally different than to add 15 or so extra components each time.

Or even a PIC which actually does not need any external components.

Maybe I can not judge the question properly because I am not actively using 74xx (not that I wanted to).

The new PICs are so powerful it is far more interesting that to deal with all the different datasheets and individual quirks of digital 74xx logic. Got a few 14pin 16F PICs here which don't even need a crystal for USB. They are cheaper than USB to serial adapters.

I saw a few years ago a project on the net, a guy who built a HDD controller from discrete 74xx. And I was thinking, a large size Whopper. At least. Stay away from that technology if it results in projects like that.

While I also think it may result in great learning experience for some individuals.

The bad thing was, when I remember right, it was not really working properly as originally intended. Maybe I don't remember right anyway.

Investing 1000 pounds just for a 74xx system is a knock-knock joke.

 

All logic needs a 0.1uF bypass capacitor on the power to ground, they are almost free when bought a reel at a time, and maybe 5 cents for 25 lots. I'd suggest Monolithic Ceramic caps (mono 0.1uF), they are the most effective for bypass and among the lowest price.

uCs can be run without the caps, but ADC readings won't be stable, and there will be occasional glitches. Stick with 8 bit until you need to work with more than 8 bits of data at a time. When you go to 32 bits, the effective memory is reduced due to word size doubling, though larger numbers are easily handled, and the uCs are made to run at higher frequencies. They are also 3.3v or less in most cases for heat reasons, so they can be passively cooled without a heatsink.

It takes a lot to "max out" an 8 bit controller running at 16Mhz, unless you want to generate a VGA signal. Monochrome VGA is possible with overclocking and some tricks, as VGA needs a properly formatted signal a bit over 25 Mhz. These are the cases where the higher speed/lower voltage 32 bit devices are needed for a full color display, but the 3.3v-2.5v makes interfacing tough at hobbyist level where 5V is the common voltage for peripherals.

 

I have 4.7uF caps here, SMD, don't cost much.

I don't know why you suggest I wanted to fiddle an 8bit controller to interface with VGA. I'd have to use assembler I guess.

I have used 18F already together with TFT display and serial FLASH.

This was a serial TFT, but I also have 16bit parallel display here.

I have recently ordered 20 pcs. PIC32- they have a price a little more than $2.
It makes little difference except they need the capacitors everywhere.

Also 50pcs. USB mini sockets are here. I have figured out I can mount them on a TQFP44 adapter, and cut off the unneeded pads (it is a cheaper type epoxy).

I have PIC32 here indeed for years but have not yet found time to make even one circuit.

When I want to use TFT there are many reasons not to stick to 8bit, even if it is possible.

Look at what's up for sale as PCBs and kits and the prices and the small powerless 8bit controllers buried. Often the price is 30 dollar to 50 dollar.

It makes no difference to use a 8bit controller or a small PIC32. The latter is just more powerful.

When I split up circuits over more than one PIC, this is for the reason to re-use modules, and to have to deal with smaller source files. It is also possible to integrate everything into one larger chip.

I have here for instance MC68SEC000 and small SMD memory chips. Again I have not yet found the time to complete a PCB layout. And what is the use and would it be more powerful than a PIC32?

A few days ago I have taken apart one 68000 prototype, removed all the wires and 74xx logic chips (about 10 or 12 of these). And installed a TQFP64 PIC.

I am sacrifying speed against to get rid of as much 74xx as possible. Actually. If I want speed I take a PIC32 or a STM32 ready-made board.

This was not as much clear to me some years ago. I could have saved the money for all that 74xx stuiff

What I have done wrong with PICs was to use assembler, I wrote programs using upto 4K memory. I learned a lot, but it is not productive enough. So I made the transition to C a while ago.

OPs question- why learn 74xx? If you need them, simply use them. The digital logic you can learn via C programming. Most C programs contain flip-flops, decoding tables, logical operations, and all this.

Why learn in hardware when you can learn in software?

 

Why learn in hardware when you can learn in software?

That is the Arduino "Just Play" idea. When one gets to a point the software cannot do what they want, the user has no clue how to make a faster circuit, without "getting a faster processor".

This site teaches electronics and circuits, digital and analog. Microcontrollers are a part of it by necessity, as are full digital solutions and simulation.

Throwing more monkeys on more typewriters to get Shakespeare to appear faster is NOT sound design philosophy.