4:20 Thanks Dave! That's my functional diagram drawing. I was a little worried some people wouldn't like how obfuscated it was, but I felt that it really clearly illustrated the function of the device!
The 74XX138 3 TO 8 decoder is a favourite of mine. I love these 74 & 4000 series chips, they pretty much form the heart of traditional basic digital electronics. Cheers for the great video Dave 👍
HC595 has to be the most common HCMOS these days. Pity they messed up the reset functionality - resetting the shift reg but not the output reg makes the pin pretty much useless - would have been nice to be able to use it for power-up reset etc.
Dave is obviously a youngster. In the beginning, there was just the 74/54 series parts, and everything you wanted to know about them was available in TI's orange " The TTL Data Book for Design Engineers". Everything in one place. That book, some parts, a few wirewrap boards with gold plated pins, AWG 30 wire, and your wirewrapping tool (manual or electric), and you were good to go.
I bought a copy of the TI TTL Databook from Heathkit back in about 1972. I still have it somewhere. The really thick orange cover National Semiconductor linear book re-published by Radio Shack was also a hobbyist must-have back then. I’ve also got the Signetics book with the infamous Write Only Memory data sheet.
I got my copy in '85. 54/74 logic were the thing, we'll I never bothered with the 54 military specs. I have a handful of those dips packages, they seems fairly robust, some go back to the 80s. I didn't get in to the cmos series, we had access to 7400 series in high school electronics.
The CD4060 is a useful part comprising an RC or XTAL oscillator and a 14 stage ripple counter for generating long timeouts from < 1 second up to many hours, without the need for a microcontroller. The CD4521 is similar but has a 24 bit counter allowing longer timeouts and/or higher oscillator frequencies. An example usage is for a simple motorized door/gate opening/shutting controller where the timer is used to protect the motor by turning it off in the event of a failure to receive an open or closed limit signal.
Still remember how I created 4 gated oscillators with a CD4011 and operated em with a 4 bit binary counter (a CD4048 I think?). Hooked up their outputs to a summing amplifier and listened to some very interesting digital sounds
I must say this one is my favourite "top jellybean components". I'll be honest: my favourite logic type is still the humble inverter with Schmitt-trigger input for the simple reason that, with just very few passives, it allows a variety of basic digital signal conditioning. R-C makes de-bouncing of switches possible in HW and it also works as a digital delay/time-shift. It can also be seen as de-glitching stage (it only lets pulses longer than a minimum through). With the addition of a diode and a second resistor it performs an asymmetric delay of the rising/falling edges and therefore pulse-shaping (changes the duty-cycle). By swapping R and C we have now a pulse generator. I can't count the number of times in which I ended up with inverters in my designs despite having FPGAs, MCUs, processors...
When I was in engineering school and learned about 7400 logic "families" I was 14/15. Due to me being a teenager with a developing brain, I was unable to stop laughing my arse off each time the teacher said "logic family". I made drawings of "them" on a picnic, going on vacations, weddings, funerals. The teachers had to confiscate them and reprimand me each time, with me unable to stop laughing. I am 40 now, and still find the idea of anthropomorphic ICs hilarious 😂 "Dad, will I ever grow up to be low power Schottky?" "If you work hard, and do your best, you can become whatever you want, son!"
For higher current with the 'HC595, take a look at TI'S TPIC6B595 and similar. It's part of their PowerLogic family and ideal for driving higher current LEDs, relays, etc. It's only current sink, not source, but for many applications that's all you need. I was intrigued by the difference between the TI CD405x and 74HC405x data sheets. Then I remembered that TI acquired their 4000 series portfolio when they acquired Harris Semiconductor (late 90s I think). They also adopted the Harris data sheets for those parts and simply rebranded them rather than create new ones.
Also handles up to 50 V and has built-in inductive transient suppression, so can be used to drive e.g. 24V relays from 5V microcontrollers with no additional parts.
@michaelhall4626 Yes, I should have mentioned that. There's also the TPIC6A595 and TPIC6C595 with higher and lower current ratings, the latter part with a 16 bit variant for automotive use.
I remember way back in the day designing a pcb to route multiple analogue signals and with some digital lamp indicators & DVM's using nothing but 74/4000 series logic. That was the days before microcontrollers where it would have been easy! The days of pouring over the logic tables and scribbling designs on paper weren't that bad....fun times!
You keep saying the multi channel analog muxes are completely independent, but they’re not… they are ganged. They share the mux select inputs, and therefore switch together.
That pricked my ears up too. But, while it’s true that the 4052 is ganged, it looks like the 4053 muxes actually are independent. I guess it’s just down to the number of available pins on the packages.
The important ones are: LS - this was the first mature version that became the standard for many years. HC - the first CMOS version, but unlike all its predecessors, it's not a 'drop in' replacement for other types. (It can drive TTL just fine but needs a pull-up resistor when driven by TTL.) HCT - a variation of HC that is fully TTL compatible (unlike HC) - no pull-up resistor needed. F, ACT and a few others - more modern designs, which unlike earlier versions, are specific to one manufacturer. You can probably ignore the basic TTL (no middle letters), L, H and S varieties.
@@chrisengland5523cheers I remember LS and HC. Also remembered when F came out from Fairchild which was meant to be much faster than the LS. I used to grab the manufacturer’s data books. Kept most of them. They are in a box somewhere
I don't really think about it much, but I probably end up putting in a 7400 series chip into half of my designs. There are just countless times where I needed a logic gate or some really basic functionality and I didn't want to tie it into my micro for one reason or another. It really feels like a bodge that isn't actually a bodge and I have a smile every time it happens.
In July 1976 I spent £5 on a Motorola 4000 series McMOS data book. That book stayed with me for decades and was a very useful source of data for the many 4000 series projects I designed over the years.
Regarding 74HC14, for similar reasons I often favor HC4093, or the 74HC132 (equivalent function, different pinout). A single package and you cover both your Schmitt function and NAND function, and if you have gates left over, a NAND can be handy for a later bodge. Then again, those single-gate "1G" family parts are very handy too.
I just finished designing a cable checker for DMX wiring using 2 x CD4093 wired as 4 latches. In one mode it uses a CD4017 to pulse 4 conductors sequentially to show conductivity, like how those network cable checkers work. . However the sneaky part is the second mode, which I call the "wiggle tester". It uses pairs of gates as a latch to detect even the shortest glitch in conductivity in each conductor.. . So, you plug in both ends of any dmx or cannon-style cable, then run the first test to ensure all conductors wired correctly. Then switch to mode 2 and bend the cable and joints and plugs and sockets... if the red light comes on, then there is an intermittent fault. . Super simple, works great and would be almost impossible to make it work using a microcontroller. 😂
I recently tore down an Arturia Polybrute analog synthesizer...the control board assemblies were littered with CD4051 mux chips for modulation routing. I counted about a dozen STM32 micros throughout the unit as well!
This pattern goes all the way back to earliest digitally controlled analog synthesizers. Even though more and more of the synth’s functionality is implemented in software (hence the additional MCUs) the basic premise of controlling the routing hasn’t changed in 40+ years.
I looked into using 405x chips for sample+hold audio circuits, but I found that that at the maximum frequency I wanted them to operate at the on resistance was too high to charge a decently sized cap fast enough, and at the minimum frequency the leakage current was too high to prevent the voltage at the cap from drooping. After some research I built a spreadsheet full of analog switch ICs, and found the best results from: MAX461x, DG6xx, DG4xx families. Some of these also have seperate analog and digital supplies, which is handy. They’re basically a 4-terminal MOSFET with control circuitry, so maybe a CD4007 style chip would work too. The 7414 is also great for oscillators, I use them to make capacitive touch pads. They do switch slower than a 7404 though.
Not sure what made anybody choose the 595 instead of the 594. As Dave mentioned the reset for the shift registers is useless, you can just toggle the clock 8 times and the register is cleared. In most cases you just clock in your data anyway. However the output registers have an undefined state at power-on, and that does matter in many applications. To prevent random output you could use the output enable to keep the outputs off until valid data is latched in, but that needs an additional pin from your microcontroller. The 594 however, has an output register clear instead of the output enable. Just put an RC circuit on it and all outputs have a known value at poweron. It should have been even better if they had kept the output enable instead of the shift register clear, but as far as I know that version does not exist. Not mentioned in the video: these chips can be controlled by SPI! And if you add also input shift registers (165) into the chain you can also expand your number of input pins.
The output enable cab be handy at times, I once used '595s to create additional chip selects on a SPI bus. Only 5 control lines to control 18 slaves in a mix of DACs, ADCs and IO expanders.
Thanks for another great video! I would also love a quick video, as you mentioned at the start, going through the different variations of the chips (or at least some number of them) and the types of applications and design decisions which would make you choose each particular variant.
I caught that too. I think Dave meant that there are two separate signal paths, but the switching is ganged together. Or you might be thinking of the 4053, where the three switches are indeed independent, though oddly the datasheet symbol omits those three separate selector inputs.
I use tons of 74ALS245. Glancing at my parts shelf, I still have about 5 full tubes on hand. I'm responsible for repair and testing of a somewhat proprietary series of controllers and interface boards that handle everything from serial data to analog & digital servo control, audio playback, timecode, midi, lighting control, etc. in a rackmount card cage arrangement. The buss interface in practically all of these board designs originally used 74FCT2245, but the throughput would crumble with more cards on the buss (they can't handle the load), so the design was switched to the 74ALS245. Whenever one of the older boards comes in for repair or testing, if it still has the 74FCT2245's on it, I have to swap them out for the 74ALS245's. Luckily they're drop-in replacements.
I think thats one of the reasons that they have become such venerable parts, a definition of a jellybean part. as progress marches on, you can just drop in new silicon and get better performance. the architecture stays just the same, only better. and I think thats one reason they stay so popular, they are going to be available for a long time. we had a moderately successful product that ran on a Xilinx FPGA, with all the breakout on 74 series stuff. Xilinx made that particular chip unobtanium, we'd purchased the worlds entire supply of NOS, and HAD to re-write. but the output side stuff could always be fixed in a few hours, no matter how catastrophic the failure. you can fix alot of stuff in software, but your far better off making it robust in hardware, if your interested in having repeat customers. Edit; as the parts became better, we could push them further. and over the years we did. I had to set a DOSBOX environment to read some of our OG schematics. I got a workflow to port it over to modern software, but the client didn't want the most fancy doohicky, they just wanted one that worked. N i think thats where the 74/4000 stuff shines, is it either just works, or obvious enough to make work. also, id far prefer to pop a .50c chip than a $70+ FPGA
i was recently using an SDIO specific analogue MUX chip, basically identical to these 74 series ones but only ~5ohm on resistance and rated for 380mhz.
You are 100% right, I use these logic device all the time for my circuit designs and I have a lot of these part numbers memorize, as they are so useful.
Some chips I would have had on the list from the TTL family are the 74138 and 74139 decoder chips, the 7490 and 7492 dividers, and the 74574 8 bit latch. There is the 74374 but the '574 has the inputs on one side and the outputs on the other. Another chip that might only get honourable mention is the 7447 BCD to seven segment decoder.
Usually these kinds of videos scare the crap out of me, because they almost always result in a big spending spree to buy all sorts of stuff I don't yet have. But, this time it turns out, I actually have ALL of these chips in my bins. So, either Dave is losing his edge, or I'm getting old. Now, I did think that Dave went just a bit hard on the old 74HC164. I mean, I absolutely swear by the 74HC595 ... a real gem. But sometimes it's nice to be able to save that one extra latch pin. So the 164s definitely still have value.
I wanted to interface an old Commodore disk drive which used the General Purpose Interface Bus (IEEE-488) for communication with a kit-built computer. I ended up using 3 x 74-series chips to decode the address bus for the interface and two 74xx574 chips to latch the outputs (one for control outputs and one for data outputs) along with two 74xx245 chips for the control and data inputs. (IIRC, the 574 used open-collector outputs which meant I only needed to send them all low in order to receive). It worked, too. Unfortunately, it turned out that the drive itself had some kind of fault which meant that data stored on disks got "lost" after only a day. I'm guessing that maybe its magnetic bias was weak, but I had no way to fix it.
I was at College (81-83) and went to WH Smiths to get a copy of a catalogue which listed pretty much all the CMOS and TTL logic gates.... seeing the pinouts of a chip was awesome.....
When I started breadboarding things back in the early 80s, I fiddled with the 7476 dual J-K flip flop. It has Q and Q-not outputs, and J for Q and K for Q-not inputs as well as a preset and clear pin for each device in a 16-pin package. It was an early favorite for me.
@@deadmanwalking6342 it’s a linear double-slope integrating VCO, which is neat I guess, but the ability to lock it onto other waves using the phase comparators makes it much more useful in my opinion.
4069. Is basically a MOSFET complementary pair and if you give it the right feedback and polarization, it can operate like a linear amplifier (inverter). I learnt that from a very old Don Lancaster schematic for an AC detector using only a 4069, a few resistors, a capacitor, and a LED
Never apologize for your insightful commentary Dave! I love it. I first started playing with TTL back in the very early 70s with recoved chip😅 that had been sawed off some retired computer of the day. Massive fun. Learned so much. Great video!
Yeah, built some projects with those. Did not use sockets, which I later regretted because a bunch of them failed during a lightning strike a km away, and I had to de-solder and replace them all...
There's an interesting 4000 series alternative to 74*595, a 4094 shift register. This register has a nice feature of an alternative serial output synchronized to phase inverted clock signal, this feature solves one of major downfalls of 595 - a desynchronization in long chains. Unfortunately, it has no reset pin to clear the register explicitly. So we basically had to add this piece of internal 4094 schematics to 595 register externally with a discrete logic thus turning it into a hybrid (4094 with reset / 595 with phase inverted sync output).
I'm currently recreating a ludlum gieger counter and slowly realizing that half of the chips inside the thing can be functionally replaced with a plain old cd40106 hex schmitt inverter. super useful for both digital and analog signals.
Still have my Do Lancaster TTL Cookbook, from high school, in the 1970s. Later in college picked up the TI TTL Databook, in hardback. This was pre-HC/HCT. Haven't used the 7400 series, in decades. Still use the 4000 series, though. Slower, but less picky. If fast logic is needed, that's PALs, CPLDs, and gate arrays are for ! 4051 - 4053 analog switches are gold !
I expected the super handy "Quad tri-state buffer" 74x125 in there. Use it all the time. I used the 74x254 a lot in the past, but apparently not anymore.
FlashcatUSB XPORT uses 3 in sequence 74HC595 in order to generate the lower 24 bits of the 28-bit address output and it uses the SPI from the AT90, as a simple and fast (8MHz) output expander.
It's quite hard to make a pick in this extremely wide field. When recently fixing an old 8-bit machine, it had a bad 74LS04. As I did not have the exact replacement in my stash, I used a 74LS14 instead. Worked like a treat, of course. The 7414 has a larger delay than the 7404, but this did not really matter in this case. Together with proper R+C at the inputs, the 7414 also makes sensible process interfaces to automation circuits running on 24 V or so. I just used a 74HC74 to generate a 1 MHz clock with well defined 50% duty cycle. I had a 4 MHz crystal oscillator at hand and one 74HC74 was just right to divide this by 4. So, I am well covered with these jellybean picks.
I was surprised that some of the configurable multi-function parts weren't mentioned. LVC1G97, 98, 99 can be great to have in-stock to get you out of a bind.
Designing a ISA bus VGA card and those 244s and 245s are really coming in clutch multiplexing the address and data bus for the main video chip. Even uses two more 245s and an 04 just for the VGA feature connector header!
My bad! It's actually two 244s since they're split in half. One full 244 for the 8-bit color bus, the other 244 split for the sync and clock lines, and the 04 to generate the three enable signals.
I love the 74xx125 for switching and / or driving a 3 or 4 wire I2S digital audio bus. I like the 74xx541 as a general buffer. they both do level translation, three state output etc. I'm just putting my fanboy love for these parts out there 🤓, and your list was spot on. A jack of all trades list, but master of a few also!
The 4046 is an awesome little chip. Too weird to be jellybean but a lot of fun - especually if you're into making noise. Its a PLL but the phase comparator and VCO sections are split apart so there are a ton of interesting circuits that can be built out of it.
Another thing about slow rise times is power dissipation, as both transistors of the chip are switching on, so the chip power dissipation can be a lot higher than expected, and current use can rise up to high levels. Enough slow inputs and you can cook the chip over time. 7474 I replaced many of them that got deaf, they started to have high input current, and thus no longer worked with TTL input currents, but would still work if you used a short to ground. As they were being used for clock and data recovery, kind of important, so those were always first suspect if you got garbled data coming in on the differential data bus. Some were kind of a pain to replace, seeing as the designers wanted to get as many TTL devices onto a card as possible, so placed them both sides, with the leads of one in the middle of the package on the other side. Replacing involved the use of a dremel, and also grinding the leadframe of the new IC down, and removal of part of the epoxy moulding, so as to fit the leads closer to the package, and soldering to the old leadframe. boards were vapour phase reflowed, then conformal coated, so the old one was often simply ground out to remove it without damaging the board traces. 1970's design and build 15 layer boards as well, with those ultra fragile traces that would lift at the slightest hint. Standard repair kit was superglue, and 40SWG enamelled copper wire.
I have built walking robots with nothing but 74HC14 inverters, and some RC pairs to create very rudimentary "neurons"... I use that term very loosely, but they worked.
595’s are bread and butter for me, I used the heck out of them in several of my projects, in combination with cd4511 7 segment decoders in some cases. You can get 2 BCD digits per register. Or if you want direct segment control some uln2803s or similar
I always keep a cache of the SN74AHC1G series (e.g. 00, 02, 04, 08, 14, 32, 86, 125, usually in DBVR / DCKR pkg's), they come in handy when a quick bodge job is required…
Apparently my old wacom mouse has some sort of low power 7400 logic in it. It gets AC from a coil inside the mouse from the pad. I still use them because they are on all the time and do not have cords. Every now and then I have to solder new buttons in because they stop working.
I think a video about the most popular family types and when to use one over another would be useful for beginners, maybe what family's you would consider "jellybean"
29:54 Nothing wrong with single-serving jellybeans. A shoutout for the 3-input configurable logic gate, which performs the function (Y or Y')=AC' + BC and can be connected to perform one of 9 different functions. TI, Diodes, Nexperia, and others offer the 74xx1G97 and 1G98 configurable logic gate with true resp. inverted outputs.
another terribly useful logic dhip is the 4046 PLL that also has a 74HC4046 version. so useful for PWM optical communications, closed loop speed control, clock syncronization and more.
I'm an analog switch fanboy you would say. On harmonic distortion- the delta Ron over Ron is the THD. You can derive this if you remember your math well enough.
I've got a parts bin chock full of 74-series logic for an 8-bit computer project I need to get round to building. I got into audio though, so the computer project got put on the back burner.
What is the predominant distortion in those muxes - compression or crossover? If the former, these have a lot of potential for audio switching; if the latter, fugheddaboudit. Non-schmitt trigger inverters like the 4049 are frequently used for guitar distortion. Low frequencies and fairly complex waveforms; don't want Schmitt mucking that up.
My AV receiver uses a CD405x to switch between inputs so they're definitely suitable for audio switching. If you look at the test notes the listed distortion figures are for driving a 1k load. Here the chip's on-resistance is forming a noticeable voltage divider. If you are driving a high impedance load I'm sure the distortion is much lower.
I'll use 4053 for all kinds of stuff in a pinch, but if CV is logic-level while signal has a higher voltage swing then sometimes it makes more sense space/cost-wise to go with a fancier "proprietary" IC with internal level shifting. "Proprietary" in quotation marks because even some of those are somewhat standardised and made by several manufacturers. BTW, BU405x (discontinued) and MAX405x have nicer specs than the standard CD/HCF series in areas such as THD and R-on. Some (professional) line-level equipment has a max output of 20-26dBu or so, ie. 22-44Vpp, so I have to disagree that 7400 is better in that instance. Maybe not so much for consumer line level. If you're dealing with such levels then odds are impedance is low anyway and R-on is less of a concern.
I still think the 7400 series are some of the finest of all time. Hell, I still use them. Need an inverter? Done. “Clock-itty-doo-dah”. Totally using that in all future documentation. 😂 Side note, Bob actually is my uncle, Robert, so I’m always successful.
I tried all the other jellybean components from your previous videos and they all tasted terrible! Maybe I just don't understand Australian candy, but I'll stick to the jellybeans I get from my grocery store.
An overview video of the family / types and applications would be great, maybe if possible timewise only go in the datasheet for the important specs/traps. Once you know the parts exist you can always read in the datasheets yourself..
The Nintendo NES actually has a CD4021 or a 74HC165 (the reverse of the 74HC164, cant remember well now) on the gamepad to read all the button presses of the controller and send them in series to the console.
Used 595s, but experienced somewhat frequent random state changes and switched to i2c port expanders. Any ideas on what’s a typical cause for that? Great videos, keep them coming!
Have used HC595s in many projects and have never seen this behaviour, except from when doing breadboard prototypes and having a wire come loose. Are you powering the 595 from 5V and interfacing with 3.3V logic?
There are, in the wild, cheap chinese 595s, which do latch output register on front of the signal, so output keeps changing with changes in shift register.
I've only diy learned "circuit" electronics, while i'm licensed to wire up BigBang stuff.. Your jellybean videos are great because they also show me devices i never knew/used, because i always had focused problems/solutions, while there are multiple other alternative or better solutions out there. If you dont know the name or number you wont find stuff. If you only have a hammer everything looks like a nail ;-)
I remember watching your jellybean transistors video a while ago and reading a comment under it that said something along the lines of "oh, I was just looking for this and Dave just uploaded a video about it" And I've just had that happen now, I was looking for multiplexers yesterday and wished this video to exist, but didn't check UA-cam because I thought that there wasn't any. Well, guess what? Dave uploaded it yesterday and I should've checked xdd
Thanks a lot for this one! I still haven't remembered all those CMOS/TTL sub-families and designations, used some of these. BTW play Nandgame, people! A fun introduction to computer engineering. I remember a Soviet counterpart of 74245 combined with 74ALS374 used in a '80s/'90s Lell PSR rhythm synthesizer (drum machine) for communicating between the CPU and program ROM. Address out, data in. Pretty clever! 74164 + 74374 = 74595... well, we're talking frontiers of math here, LOL (Oh... and where's that unobtainium 74141 when you need one? If I was Sam Zeloof or Jeri Ellsworth, I could have a go at re-creating these.)
4:20 Thanks Dave! That's my functional diagram drawing. I was a little worried some people wouldn't like how obfuscated it was, but I felt that it really clearly illustrated the function of the device!
The 74XX138 3 TO 8 decoder is a favourite of mine. I love these 74 & 4000 series chips, they pretty much form the heart of traditional basic digital electronics.
Cheers for the great video Dave 👍
Yep, a single 74XX138 can be used to decode up to 8 bit address into a chip select! Powerful
HC595 has to be the most common HCMOS these days.
Pity they messed up the reset functionality - resetting the shift reg but not the output reg makes the pin pretty much useless - would have been nice to be able to use it for power-up reset etc.
I’ll bet that it has the highest volume of what little remains in production.
Dave is obviously a youngster. In the beginning, there was just the 74/54 series parts, and everything you wanted to know about them was available in TI's orange " The TTL Data Book for Design Engineers". Everything in one place. That book, some parts, a few wirewrap boards with gold plated pins, AWG 30 wire, and your wirewrapping tool (manual or electric), and you were good to go.
I bought a copy of the TI TTL Databook from Heathkit back in about 1972. I still have it somewhere. The really thick orange cover National Semiconductor linear book re-published by Radio Shack was also a hobbyist must-have back then. I’ve also got the Signetics book with the infamous Write Only Memory data sheet.
Huh, I still have my copy!
I got my copy in '85. 54/74 logic were the thing, we'll I never bothered with the 54 military specs. I have a handful of those dips packages, they seems fairly robust, some go back to the 80s. I didn't get in to the cmos series, we had access to 7400 series in high school electronics.
Was my bible at school. 🙂 (my physics teacher gave me a copy of data book)
@@EEVblog Never got the Heathkit one, got the Indian clone one, still have it, plus some TTL data books, around.
The CD4060 is a useful part comprising an RC or XTAL oscillator and a 14 stage ripple counter for generating long timeouts from < 1 second up to many hours, without the need for a microcontroller. The CD4521 is similar but has a 24 bit counter allowing longer timeouts and/or higher oscillator frequencies. An example usage is for a simple motorized door/gate opening/shutting controller where the timer is used to protect the motor by turning it off in the event of a failure to receive an open or closed limit signal.
I love this Jellybean videos and would like a poster containing all of them with their pinout and main characteristics.
4017 counter was my introduction IC into electronics as a teen. I also remember doing some sound experiments with a 4011 IC.
I believe mine was a 555 driving the clock of a 4017 to make a sweet LED chaser. Man that brings back some memories.
@@jrobmccoy Were you making Knight Rider lights?
Still remember how I created 4 gated oscillators with a CD4011 and operated em with a 4 bit binary counter (a CD4048 I think?). Hooked up their outputs to a summing amplifier and listened to some very interesting digital sounds
I must say this one is my favourite "top jellybean components".
I'll be honest: my favourite logic type is still the humble inverter with Schmitt-trigger input for the simple reason that, with just very few passives, it allows a variety of basic digital signal conditioning.
R-C makes de-bouncing of switches possible in HW and it also works as a digital delay/time-shift. It can also be seen as de-glitching stage (it only lets pulses longer than a minimum through). With the addition of a diode and a second resistor it performs an asymmetric delay of the rising/falling edges and therefore pulse-shaping (changes the duty-cycle). By swapping R and C we have now a pulse generator. I can't count the number of times in which I ended up with inverters in my designs despite having FPGAs, MCUs, processors...
When I was in engineering school and learned about 7400 logic "families" I was 14/15. Due to me being a teenager with a developing brain, I was unable to stop laughing my arse off each time the teacher said "logic family". I made drawings of "them" on a picnic, going on vacations, weddings, funerals. The teachers had to confiscate them and reprimand me each time, with me unable to stop laughing. I am 40 now, and still find the idea of anthropomorphic ICs hilarious 😂 "Dad, will I ever grow up to be low power Schottky?" "If you work hard, and do your best, you can become whatever you want, son!"
Chip designers always did that, there are many famous ones whose doodles actually got included in the National Semiconductor data sheets as well.
TTL and CMOS are the Hatfield's and the McCoy's.
@@SeanBZA And they can sometimes be found on the actual die
For higher current with the 'HC595, take a look at TI'S TPIC6B595 and similar. It's part of their PowerLogic family and ideal for driving higher current LEDs, relays, etc. It's only current sink, not source, but for many applications that's all you need. I was intrigued by the difference between the TI CD405x and 74HC405x data sheets. Then I remembered that TI acquired their 4000 series portfolio when they acquired Harris Semiconductor (late 90s I think). They also adopted the Harris data sheets for those parts and simply rebranded them rather than create new ones.
Also handles up to 50 V and has built-in inductive transient suppression, so can be used to drive e.g. 24V relays from 5V microcontrollers with no additional parts.
@michaelhall4626 Yes, I should have mentioned that. There's also the TPIC6A595 and TPIC6C595 with higher and lower current ratings, the latter part with a 16 bit variant for automotive use.
16:23 I appreciate Dave's Signing for audiophiles.
I remember way back in the day designing a pcb to route multiple analogue signals and with some digital lamp indicators & DVM's using nothing but 74/4000 series logic. That was the days before microcontrollers where it would have been easy! The days of pouring over the logic tables and scribbling designs on paper weren't that bad....fun times!
You keep saying the multi channel analog muxes are completely independent, but they’re not… they are ganged. They share the mux select inputs, and therefore switch together.
That pricked my ears up too. But, while it’s true that the 4052 is ganged, it looks like the 4053 muxes actually are independent. I guess it’s just down to the number of available pins on the packages.
A breakdown of all the different 74 types would be great. HC L LS F etc
The important ones are:
LS - this was the first mature version that became the standard for many years.
HC - the first CMOS version, but unlike all its predecessors, it's not a 'drop in' replacement for other types. (It can drive TTL just fine but needs a pull-up resistor when driven by TTL.)
HCT - a variation of HC that is fully TTL compatible (unlike HC) - no pull-up resistor needed.
F, ACT and a few others - more modern designs, which unlike earlier versions, are specific to one manufacturer.
You can probably ignore the basic TTL (no middle letters), L, H and S varieties.
@@chrisengland5523cheers
I remember LS and HC. Also remembered when F came out from Fairchild which was meant to be much faster than the LS. I used to grab the manufacturer’s data books.
Kept most of them. They are in a box somewhere
I don't really think about it much, but I probably end up putting in a 7400 series chip into half of my designs. There are just countless times where I needed a logic gate or some really basic functionality and I didn't want to tie it into my micro for one reason or another. It really feels like a bodge that isn't actually a bodge and I have a smile every time it happens.
In July 1976 I spent £5 on a Motorola 4000 series McMOS data book. That book stayed with me for decades and was a very useful source of data for the many 4000 series projects I designed over the years.
Do McMOS circuits use Scottish diodes? 😊
@@copernicofelinis HaHa oh yes definately
Regarding 74HC14, for similar reasons I often favor HC4093, or the 74HC132 (equivalent function, different pinout). A single package and you cover both your Schmitt function and NAND function, and if you have gates left over, a NAND can be handy for a later bodge. Then again, those single-gate "1G" family parts are very handy too.
I just finished designing a cable checker for DMX wiring using 2 x CD4093 wired as 4 latches. In one mode it uses a CD4017 to pulse 4 conductors sequentially to show conductivity, like how those network cable checkers work.
.
However the sneaky part is the second mode, which I call the "wiggle tester". It uses pairs of gates as a latch to detect even the shortest glitch in conductivity in each conductor..
.
So, you plug in both ends of any dmx or cannon-style cable, then run the first test to ensure all conductors wired correctly. Then switch to mode 2 and bend the cable and joints and plugs and sockets... if the red light comes on, then there is an intermittent fault.
.
Super simple, works great and would be almost impossible to make it work using a microcontroller. 😂
I agree, I wanted to post that the 4093/HC(T)132 is my favorite jelly-bean device too.
0:57 Yes! I want a video about all the families!
seconded.
I recently tore down an Arturia Polybrute analog synthesizer...the control board assemblies were littered with CD4051 mux chips for modulation routing. I counted about a dozen STM32 micros throughout the unit as well!
You tore it down?!? Those are worth around $8,500.
@@hxhdfjifzirstc894 More like $2700 USD.
This pattern goes all the way back to earliest digitally controlled analog synthesizers. Even though more and more of the synth’s functionality is implemented in software (hence the additional MCUs) the basic premise of controlling the routing hasn’t changed in 40+ years.
Nice one Dave, using 595's at present to expand my PIC's output to drive 16 leds.
Seen all of these out in the wild. And these are good ones to know because they are often in designs that may not be full of logic chips.
I looked into using 405x chips for sample+hold audio circuits, but I found that that at the maximum frequency I wanted them to operate at the on resistance was too high to charge a decently sized cap fast enough, and at the minimum frequency the leakage current was too high to prevent the voltage at the cap from drooping. After some research I built a spreadsheet full of analog switch ICs, and found the best results from: MAX461x, DG6xx, DG4xx families. Some of these also have seperate analog and digital supplies, which is handy. They’re basically a 4-terminal MOSFET with control circuitry, so maybe a CD4007 style chip would work too.
The 7414 is also great for oscillators, I use them to make capacitive touch pads. They do switch slower than a 7404 though.
I like TPIC6B595 which is open-drain variant of 74x595 (this one has 50V clamping voltage and 5 Ohm Rds on and over 100mA continuous current).
Also it will work with the Arduino library for driving the 74HC595.
Not sure what made anybody choose the 595 instead of the 594. As Dave mentioned the reset for the shift registers is useless, you can just toggle the clock 8 times and the register is cleared. In most cases you just clock in your data anyway. However the output registers have an undefined state at power-on, and that does matter in many applications. To prevent random output you could use the output enable to keep the outputs off until valid data is latched in, but that needs an additional pin from your microcontroller. The 594 however, has an output register clear instead of the output enable. Just put an RC circuit on it and all outputs have a known value at poweron. It should have been even better if they had kept the output enable instead of the shift register clear, but as far as I know that version does not exist. Not mentioned in the video: these chips can be controlled by SPI! And if you add also input shift registers (165) into the chain you can also expand your number of input pins.
The output enable cab be handy at times, I once used '595s to create additional chip selects on a SPI bus. Only 5 control lines to control 18 slaves in a mix of DACs, ADCs and IO expanders.
Thanks for another great video! I would also love a quick video, as you mentioned at the start, going through the different variations of the chips (or at least some number of them) and the types of applications and design decisions which would make you choose each particular variant.
6:52 - 4052 I don’t think the switches are independent, there are only two select lines. What am I missing?
I caught that too. I think Dave meant that there are two separate signal paths, but the switching is ganged together. Or you might be thinking of the 4053, where the three switches are indeed independent, though oddly the datasheet symbol omits those three separate selector inputs.
@@Graham_Wideman 4053 has A, B, C marked as individual selectors in the drawing, at the switch.
@@niclash Right, but the symbol omits the actual input pins, unlike the 4051 and 4052 symbols.
I use tons of 74ALS245. Glancing at my parts shelf, I still have about 5 full tubes on hand. I'm responsible for repair and testing of a somewhat proprietary series of controllers and interface boards that handle everything from serial data to analog & digital servo control, audio playback, timecode, midi, lighting control, etc. in a rackmount card cage arrangement. The buss interface in practically all of these board designs originally used 74FCT2245, but the throughput would crumble with more cards on the buss (they can't handle the load), so the design was switched to the 74ALS245. Whenever one of the older boards comes in for repair or testing, if it still has the 74FCT2245's on it, I have to swap them out for the 74ALS245's. Luckily they're drop-in replacements.
I think thats one of the reasons that they have become such venerable parts, a definition of a jellybean part. as progress marches on, you can just drop in new silicon and get better performance. the architecture stays just the same, only better. and I think thats one reason they stay so popular, they are going to be available for a long time. we had a moderately successful product that ran on a Xilinx FPGA, with all the breakout on 74 series stuff. Xilinx made that particular chip unobtanium, we'd purchased the worlds entire supply of NOS, and HAD to re-write. but the output side stuff could always be fixed in a few hours, no matter how catastrophic the failure. you can fix alot of stuff in software, but your far better off making it robust in hardware, if your interested in having repeat customers. Edit; as the parts became better, we could push them further. and over the years we did. I had to set a DOSBOX environment to read some of our OG schematics. I got a workflow to port it over to modern software, but the client didn't want the most fancy doohicky, they just wanted one that worked. N i think thats where the 74/4000 stuff shines, is it either just works, or obvious enough to make work. also, id far prefer to pop a .50c chip than a $70+ FPGA
i was recently using an SDIO specific analogue MUX chip, basically identical to these 74 series ones but only ~5ohm on resistance and rated for 380mhz.
380 millihertz seems a little restrictive...
You are 100% right, I use these logic device all the time for my circuit designs and I have a lot of these part numbers memorize, as they are so useful.
Brilliant run through as always Dave, nice one, I hadn't used the 405x series and learned a lot from this
Some chips I would have had on the list from the TTL family are the 74138 and 74139 decoder chips, the 7490 and 7492 dividers, and the 74574 8 bit latch. There is the 74374 but the '574 has the inputs on one side and the outputs on the other. Another chip that might only get honourable mention is the 7447 BCD to seven segment decoder.
Usually these kinds of videos scare the crap out of me, because they almost always result in a big spending spree to buy all sorts of stuff I don't yet have. But, this time it turns out, I actually have ALL of these chips in my bins. So, either Dave is losing his edge, or I'm getting old.
Now, I did think that Dave went just a bit hard on the old 74HC164. I mean, I absolutely swear by the 74HC595 ... a real gem. But sometimes it's nice to be able to save that one extra latch pin. So the 164s definitely still have value.
I wanted to interface an old Commodore disk drive which used the General Purpose Interface Bus (IEEE-488) for communication with a kit-built computer. I ended up using 3 x 74-series chips to decode the address bus for the interface and two 74xx574 chips to latch the outputs (one for control outputs and one for data outputs) along with two 74xx245 chips for the control and data inputs. (IIRC, the 574 used open-collector outputs which meant I only needed to send them all low in order to receive).
It worked, too. Unfortunately, it turned out that the drive itself had some kind of fault which meant that data stored on disks got "lost" after only a day. I'm guessing that maybe its magnetic bias was weak, but I had no way to fix it.
An interesting application for CD4502 is digital IQ mixer (in practice use high-speed alternatives like cbt3253).
I was at College (81-83) and went to WH Smiths to get a copy of a catalogue which listed pretty much all the CMOS and TTL logic gates.... seeing the pinouts of a chip was awesome.....
I'll have to admit, for messing around my fav chip is the quad nand. Can do anything with enough of them.
When I started breadboarding things back in the early 80s, I fiddled with the 7476 dual J-K flip flop. It has Q and Q-not outputs, and J for Q and K for Q-not inputs as well as a preset and clear pin for each device in a 16-pin package. It was an early favorite for me.
17:33. I’ve used the 4066 with +/-7.5V before in my modular synth. Or am I misunderstanding something?
As Dave said you need to bias accordingly. If you drive the digital input pins negative instead of just to zero it will work just fine.
Fastest way to get a 2 phase variable frequency oscillator
using 1 resistor and 1 capacitor? CD4046 and HC4046
jelly beans!
Woo, PLLs are awesome!
@@Scrogan ? 2 phase variable oscillator not PLL.4046 can do a lot more then PLL.
@@deadmanwalking6342 it’s a linear double-slope integrating VCO, which is neat I guess, but the ability to lock it onto other waves using the phase comparators makes it much more useful in my opinion.
Two independent 74HC04s made one of the best ps high noise environment power VCC buffer circuits from my experience.
4069. Is basically a MOSFET complementary pair and if you give it the right feedback and polarization, it can operate like a linear amplifier (inverter).
I learnt that from a very old Don Lancaster schematic for an AC detector using only a 4069, a few resistors, a capacitor, and a LED
Never apologize for your insightful commentary Dave! I love it. I first started playing with TTL back in the very early 70s with recoved chip😅 that had been sawed off some retired computer of the day. Massive fun. Learned so much. Great video!
Lived through the _extremely static sensitive_ original CMOS 4000 AE DIPS
The trusty, 7474 D-triggered FF was my earliest favorite.
I remember those coming in metal rails sealed with metal tape - only way to be sure.
Yeah, built some projects with those. Did not use sockets, which I later regretted because a bunch of them failed during a lightning strike a km away, and I had to de-solder and replace them all...
There's an interesting 4000 series alternative to 74*595, a 4094 shift register. This register has a nice feature of an alternative serial output synchronized to phase inverted clock signal, this feature solves one of major downfalls of 595 - a desynchronization in long chains. Unfortunately, it has no reset pin to clear the register explicitly. So we basically had to add this piece of internal 4094 schematics to 595 register externally with a discrete logic thus turning it into a hybrid (4094 with reset / 595 with phase inverted sync output).
I'm currently recreating a ludlum gieger counter and slowly realizing that half of the chips inside the thing can be functionally replaced with a plain old cd40106 hex schmitt inverter. super useful for both digital and analog signals.
I named 3 of the same chips on your list at the start of your video. Just used 2 in my last PCB design.
Still have my Do Lancaster TTL Cookbook, from high school, in the 1970s. Later in college picked up the TI TTL Databook, in hardback. This was pre-HC/HCT.
Haven't used the 7400 series, in decades. Still use the 4000 series, though. Slower, but less picky. If fast logic is needed, that's PALs, CPLDs, and gate arrays are for !
4051 - 4053 analog switches are gold !
I expected the super handy "Quad tri-state buffer" 74x125 in there. Use it all the time.
I used the 74x254 a lot in the past, but apparently not anymore.
As a homebrew computer nut my favorites would be the 74F350 which do shifting of 4 bits by 0-3 bits
ooooh that's a good one!
FlashcatUSB XPORT uses 3 in sequence 74HC595 in order to generate the lower 24 bits of the 28-bit address output and it uses the SPI from the AT90, as a simple and fast (8MHz) output expander.
I made a 6502 ROR test board (the first batch of 6502s didn't have ROR instructions) and i used a 74ls14 and 74ls74 to drive everything.
Oh yeah, this brings back the good old days of repairing industrial circuit-boards. It was the cat's meow for that era. 😎 Thank you.
They came up in exactly the order I thought of them when I read the title. The analog switch is in so many of our designs.
It's quite hard to make a pick in this extremely wide field.
When recently fixing an old 8-bit machine, it had a bad 74LS04. As I did not have the exact replacement in my stash, I used a 74LS14 instead. Worked like a treat, of course. The 7414 has a larger delay than the 7404, but this did not really matter in this case. Together with proper R+C at the inputs, the 7414 also makes sensible process interfaces to automation circuits running on 24 V or so.
I just used a 74HC74 to generate a 1 MHz clock with well defined 50% duty cycle. I had a 4 MHz crystal oscillator at hand and one 74HC74 was just right to divide this by 4.
So, I am well covered with these jellybean picks.
I was surprised that some of the configurable multi-function parts weren't mentioned. LVC1G97, 98, 99 can be great to have in-stock to get you out of a bind.
XOR fanboy here too. Shout out to the 74163, the 4 bit synchronous counter with preset, carry and clear. None of that ripple counter crap 😄
Designing a ISA bus VGA card and those 244s and 245s are really coming in clutch multiplexing the address and data bus for the main video chip. Even uses two more 245s and an 04 just for the VGA feature connector header!
My bad! It's actually two 244s since they're split in half. One full 244 for the 8-bit color bus, the other 244 split for the sync and clock lines, and the 04 to generate the three enable signals.
I love the 74xx125 for switching and / or driving a 3 or 4 wire I2S digital audio bus. I like the 74xx541 as a general buffer. they both do level translation, three state output etc. I'm just putting my fanboy love for these parts out there 🤓, and your list was spot on. A jack of all trades list, but master of a few also!
The 4046 is an awesome little chip. Too weird to be jellybean but a lot of fun - especually if you're into making noise. Its a PLL but the phase comparator and VCO sections are split apart so there are a ton of interesting circuits that can be built out of it.
Thanks!
Thanks, much appreciated.
Another thing about slow rise times is power dissipation, as both transistors of the chip are switching on, so the chip power dissipation can be a lot higher than expected, and current use can rise up to high levels. Enough slow inputs and you can cook the chip over time.
7474 I replaced many of them that got deaf, they started to have high input current, and thus no longer worked with TTL input currents, but would still work if you used a short to ground. As they were being used for clock and data recovery, kind of important, so those were always first suspect if you got garbled data coming in on the differential data bus. Some were kind of a pain to replace, seeing as the designers wanted to get as many TTL devices onto a card as possible, so placed them both sides, with the leads of one in the middle of the package on the other side. Replacing involved the use of a dremel, and also grinding the leadframe of the new IC down, and removal of part of the epoxy moulding, so as to fit the leads closer to the package, and soldering to the old leadframe. boards were vapour phase reflowed, then conformal coated, so the old one was often simply ground out to remove it without damaging the board traces. 1970's design and build 15 layer boards as well, with those ultra fragile traces that would lift at the slightest hint. Standard repair kit was superglue, and 40SWG enamelled copper wire.
One of my favorites on this channel, love it.
I have built walking robots with nothing but 74HC14 inverters, and some RC pairs to create very rudimentary "neurons"... I use that term very loosely, but they worked.
30:35 "you can tell it's bettah, it's called a Shmick Triggah'!" Hahaha
595’s are bread and butter for me, I used the heck out of them in several of my projects, in combination with cd4511 7 segment decoders in some cases. You can get 2 BCD digits per register. Or if you want direct segment control some uln2803s or similar
I miss this Dave, good info. Loving it.
I always keep a cache of the SN74AHC1G series (e.g. 00, 02, 04, 08, 14, 32, 86, 125, usually in DBVR / DCKR pkg's), they come in handy when a quick bodge job is required…
Apparently my old wacom mouse has some sort of low power 7400 logic in it. It gets AC from a coil inside the mouse from the pad. I still use them because they are on all the time and do not have cords. Every now and then I have to solder new buttons in because they stop working.
I think a video about the most popular family types and when to use one over another would be useful for beginners, maybe what family's you would consider "jellybean"
I learned all these parts back before time it seems. Still a nand fan boy. Fun video.
Gold. Absolutely gold. Thanks Dave!
I was introduced to these chips in 1975. Realizing that was almost 50 years ago, I feel old. Sigh!
29:54 Nothing wrong with single-serving jellybeans. A shoutout for the 3-input configurable logic gate, which performs the function (Y or Y')=AC' + BC and can be connected to perform one of 9 different functions. TI, Diodes, Nexperia, and others offer the 74xx1G97 and 1G98 configurable logic gate with true resp. inverted outputs.
another terribly useful logic dhip is the 4046 PLL that also has a 74HC4046 version. so useful for PWM optical communications, closed loop speed control, clock syncronization and more.
I'm an analog switch fanboy you would say. On harmonic distortion- the delta Ron over Ron is the THD. You can derive this if you remember your math well enough.
I've got a parts bin chock full of 74-series logic for an 8-bit computer project I need to get round to building. I got into audio though, so the computer project got put on the back burner.
SPOT ON, my exact pick, and crazy crazy, IN THE SAME ORDER!!!!!!!!!!!!!!!!!!! .....love it Dave!
What is the predominant distortion in those muxes - compression or crossover? If the former, these have a lot of potential for audio switching; if the latter, fugheddaboudit.
Non-schmitt trigger inverters like the 4049 are frequently used for guitar distortion. Low frequencies and fairly complex waveforms; don't want Schmitt mucking that up.
My AV receiver uses a CD405x to switch between inputs so they're definitely suitable for audio switching. If you look at the test notes the listed distortion figures are for driving a 1k load. Here the chip's on-resistance is forming a noticeable voltage divider. If you are driving a high impedance load I'm sure the distortion is much lower.
I'll use 4053 for all kinds of stuff in a pinch, but if CV is logic-level while signal has a higher voltage swing then sometimes it makes more sense space/cost-wise to go with a fancier "proprietary" IC with internal level shifting. "Proprietary" in quotation marks because even some of those are somewhat standardised and made by several manufacturers.
BTW, BU405x (discontinued) and MAX405x have nicer specs than the standard CD/HCF series in areas such as THD and R-on.
Some (professional) line-level equipment has a max output of 20-26dBu or so, ie. 22-44Vpp, so I have to disagree that 7400 is better in that instance. Maybe not so much for consumer line level. If you're dealing with such levels then odds are impedance is low anyway and R-on is less of a concern.
4046 for the win! PLLs make all kinds of fancy things possible. And they're fun to play with. 😁👍
I still think the 7400 series are some of the finest of all time. Hell, I still use them. Need an inverter? Done.
“Clock-itty-doo-dah”. Totally using that in all future documentation. 😂
Side note, Bob actually is my uncle, Robert, so I’m always successful.
I tried all the other jellybean components from your previous videos and they all tasted terrible! Maybe I just don't understand Australian candy, but I'll stick to the jellybeans I get from my grocery store.
Add salt or garlic.
You gotta chuck em on the barbie for a few mins first, then wash them down with 8 tinnies.
Top 5 jellybean vacuum tubes would be nice :)
An overview video of the family / types and applications would be great, maybe if possible timewise only go in the datasheet for the important specs/traps.
Once you know the parts exist you can always read in the datasheets yourself..
The Nintendo NES actually has a CD4021 or a 74HC165 (the reverse of the 74HC164, cant remember well now) on the gamepad to read all the button presses of the controller and send them in series to the console.
Love this classic stuff, it’s like LEGO for engineers :)
Thanks for posting Dave. Question have you ever used the CD4024be ic chip? It’s a seven stage counter chip.
4052 could be for switching stereo sources in an amp for example?
I bought some assorted 7400 and 4000 ICs and just go down the line looking up datasheets for them.
XOR fanbois! 🤣 love it. Thanks for great rundown, handy stuff
👍 thanks for this one Dave! I've got to place an order now ;-)
74LVC245 / SN74LVC4245 also useful for interfacing 5 V to low voltage logic.
I'm wondering how did they come up with the numbering?
Why is a 04 an inverter, or a 74 a flip-flop and so on?
Used 595s, but experienced somewhat frequent random state changes and switched to i2c port expanders. Any ideas on what’s a typical cause for that? Great videos, keep them coming!
Have used HC595s in many projects and have never seen this behaviour, except from when doing breadboard prototypes and having a wire come loose. Are you powering the 595 from 5V and interfacing with 3.3V logic?
As far as I remember it was powered by 3.3, using Raspberry Pi.
There are, in the wild, cheap chinese 595s, which do latch output register on front of the signal, so output keeps changing with changes in shift register.
I've only diy learned "circuit" electronics, while i'm licensed to wire up BigBang stuff..
Your jellybean videos are great because they also show me devices i never knew/used, because i always had focused problems/solutions, while there are multiple other alternative or better solutions out there. If you dont know the name or number you wont find stuff.
If you only have a hammer everything looks like a nail ;-)
I remember watching your jellybean transistors video a while ago and reading a comment under it that said something along the lines of "oh, I was just looking for this and Dave just uploaded a video about it"
And I've just had that happen now, I was looking for multiplexers yesterday and wished this video to exist, but didn't check UA-cam because I thought that there wasn't any. Well, guess what? Dave uploaded it yesterday and I should've checked xdd
I love these videos Dave!
As opposite to the digital output expander 74xx595, i missed the digital input expander 74xx165
Thanks a lot for this one! I still haven't remembered all those CMOS/TTL sub-families and designations, used some of these. BTW play Nandgame, people! A fun introduction to computer engineering.
I remember a Soviet counterpart of 74245 combined with 74ALS374 used in a '80s/'90s Lell PSR rhythm synthesizer (drum machine) for communicating between the CPU and program ROM. Address out, data in. Pretty clever!
74164 + 74374 = 74595... well, we're talking frontiers of math here, LOL
(Oh... and where's that unobtainium 74141 when you need one? If I was Sam Zeloof or Jeri Ellsworth, I could have a go at re-creating these.)
A video about the different families would be helpful