Reasonable question and hopefully an interesting answer...
The simple lack of reasons to use TTL logic in 2026 was exactly why I didn't know what the deal was. It'd never come up, but I'd see it referenced.
I'm self-taught and in defiance of the people who insist that LLMs turn our brains to passive mush, the more things I learn the more things I have to be curious about.
LLMs remove the gatekeeping around asking "simple" questions that tend to make EEs roll their eyes. I didn't know, so I asked and now I know!
I'm actually pretty thrilled that you asked, because I think that this chat is an extremely solid example of LLM usage in the EE domain, and I'm happy to share.
I definitely led some questions to try and squeeze new-to-me perspectives out of it; for example, there could be tricks that make the active high variant more useful in some scenarios.
I think it does a good job of surfacing adjacent questions you might not realize you were eager to ask, as well as showing how it's able to critically evaluate real-world part suitability. I do find that ChatGPT in particular does better with a screengrab of the most likely parts vs a URL to the search engine.
I see the chat, but it looks like you’re not actually considering using TTL anywhere, and ChatGPT isn’t giving any explanations about TTL?
> I would definitely like to understand HCT vs HC (CMOS vs TTL) much better than I do, which currently isn't at all.
I think what ChatGPT should have explained at the beginning is that both HCT and HC are CMOS logic families, it’s just that HCT is designed to interface with TTL (receive TTL signal levels as inputs). The outputs are the same (CMOS outputs are rail to rail, which you can feed into TTL just fine).
Actual TTL logic, like the 7400 series and the variations (LS is one of the more popular variations), uses NPN transistors as inputs and to pull output signals low. It uses resistors to pull the signals high. The result is a lot of current consumption and asymmetrical output signals… maybe a good question to ask ChatGPT is “why does TTL use so much current?” CMOS, by comparison, uses a tiny amount of current except when it is switching.
I would probably choose AHC first as a logic family these days. It’s a slightly better version of HC, but it’s not so fast that it will cause problems.
Just peeking at one of the recommendations in the chat, if you search for 74HCT125 or 74AHC125 on Mouser, you’ll see that the AHC has more options available and more parts in stock. That’s a sign that it’s probably a more popular logic family than HCT, which is something I consider when buying (more popular = better availability).
Thanks so much for the additional context. You've given me more to dig into.
What I would like to know from you is:
1. On the whole, is the information you see it present more or less coherent and useful? Is it better to have this information than not have it at all?
2. Where does this land in terms of your expectations? Did anything surprise you?
It's clear from your reply that you know what you're talking about, while I'm still clawing my way up from nothing... so it makes sense that you have fewer things that you need to ask about.
I've bootstrapped my entire EE skillset over the past 2-3 years, largely with the help of LLMs to interrogate. It's helped me design and build my first product. I'm confident that without these tools, it's not a question of how long it would have taken so much as the truth: it would have died on the vine.
I asked it about the AHC family equivalent and it recommended against using it, suggesting either AHCT or sticking with HCT. For what it's worth, the reference board that I'm tracing uses an HCT, so the LLM isn't wrong.
Note that at the time I'm writing this, I have an extremely fuzzy understanding of the difference between these three... but I'm working through it.
I’m mostly just curious about how people use LLMs to learn. I don’t know what your goals are, and even if your goals were the same as mine, I don’t know how LLMs stack up against the way I learned (mostly from books). At least, not long-term. I’m not that good at electronics, I’m just a hobbyist that went through Forrest Mims mini-notebooks and later Horowitz and Hill.
What I like about information from humans is that humans are always trying to figure out how to say things that are relevant and informative. By “relevant”, I mean that we try to avoid saying things that don’t help you. By “informative”, I mean that we try to include information that you want to know, even if you didn’t specifically ask for it.
Picking on the chat for a moment—when you started out with the question, my first thought was, “This person is specifically asking about HC versus HCT, but maybe they want a broader overview of logic families, and maybe they want to understand which logic family to pick for their hobby project.” That’s an example where I think ChatGPT could have identified something that you wanted to know, but didn’t. (It wasn’t as informative as it could have been.)
Then there’s some times that ChatGPT gave you information of dubious relevance.
> Important: On the HCT125, the enable is active-LOW.
I don’t think that’s contextually important. It’s like saying, “Important: On the Honda Civic, the gas tank is on the left.” That’s contextually important when you’re at the gas station, but not when you’re buying a car.
I’m not sure why the LLM is recommending the TTL-compatible chips. IMO, the right thing to do here is probably to run everything at 3.3V, unless you have something that specifically needs 5V. When everything is at 3.3V, you don’t have to think about level shifting and you can just pick a very boring logic family like AHC. But I don’t know what you’re building. Likewise, I would lean towards using normal CMOS logic levels, unless I had a specific reason to choose TTL-compatible. The regular CMOS versions have better noise margin, because the threshold is in the optimum place—right in the middle.
I can actually clear a lot of that up. ChatGPT has accumulated a significant amount of ambient knowledge about what I'm working on and how I typically progress through asking questions, so the path isn't as blue sky as it appears.
For example, I'm working with a specialty SPCO switch IC that runs at 5V. There's never been and likely never will be a 3.3V version of the AS16M1. Being able to drive the switch (which functions like a shift register) from my ESP32 is top-of-mind.
The HCT125 being active low is directly responding to my question about why to choose it vs the 126 version; since the board I'm studying (which again, it has seen) uses 125s, it's reasonable to wonder why they'd choose one over the other.
Overall, the tone of chats on EE topics tend to be task-focused with permission to go on interesting side quests. I'm trying to get stuff done with room for relevant exploration along the way.
I think you're in the minority of people that are using LLMs for one of the best uses - for augmenting your own understanding and intelligence. Of course you have to triangulate and triple check what they say, but that's a good habit to get into anyway. Many of my teachers would repeat tribal myths all the same.
(I’m not saying it’s not used, but the only thing I’d use TTL for is building old circuits out of the Forrest Mims books.)