7400 Quad 2-Input NAND Gate: Neglected Survivor from a Pre-Microprocessor World | Hacker News Reader

7400 Quad 2-Input NAND Gate: Neglected Survivor from a Pre-Microprocessor World

hackaday.com

56 pointssohkamyung20 comments

Surprising they didn't mention the PO74G00A and its friends.

http://www.potatosemi.com/datasheet/PO74G00A.pdf

Note before you try to make a 1 GHz CPU using 1970s design with those chips, for current limit reasons those chips can only switch in the GHz range with output capacitance below the input capacitance of a single gate input, so your redesigned PDP-11 is only gonna run at a mere 500 MHz not in the GHz range.

I'm a little unclear where you buy their products other than their ebay store. Digikey doesn't seem to stock. They DO seem to be real, not a meme.

I've never screwed around with the PotSemi products but I've seen the youtube videos where some dude makes a classic ring oscillator out of an inverter that runs at some high UHF frequency like seven hundred MHz or so. I thought I was pretty cool making ring oscilators running at 20-something MHz back in the 80s. If you're really bored or don't have a tape measure you can measure a length of cable by inserting it into a ring osc and recording the resulting frequency and applying some correctional math based on the speed of the inverter... I suppose with a GHz class inverter the correction factor would be round to zero for any cable longer than a couple feet and could be ignored. You can also make a toy TDR using what amounts to a ring osc and a very fast (expensive) o-scope.

Power supply noise decoupling caps must be "fun" with those chips, what happens if you put a very old 0.1 uF cap that self-resonates at a freq lower than the switching freq of the chip? Nothing good, I imagine. I know you can buy 0402 SMD caps that resonate in the higher GHz range so thats OK, but you know some goofball is going to try an 0.1 uF electrolytic for decoupling LOL.

In the "Old Days" you used strange fast high power logic sparingly to eliminate race conditions, obviously the 74F would always latch up before the 74LS or whatever so careful use of ttl family could eliminate race conditions. I would imagine that, and strange interfacing tricks, would be the purpose of modern TTL. If you wanted a slow arduino to intelligently monitor a 30 MHz signal, a little TTL support might make it possible in certain very limited cases.

That datasheet is unclear about what exactly is required to get that 1GHz+ performance; my naive reading is that the CL=2pF case is half of another gate's input capacitance, which seems implausible.

OTOH, if CMOS compatibility isn't a problem you can go waaaaay faster with things like this:

https://www.analog.com/en/products/hmc746.html

Single gate per package, but supports 10+Gbps data rates. 290mW per gate power consumption o_O

At those frequencies an electrolytic cap is best modeled as an inductor with a parasitic capacitance ;)

You can do TDR (cable measurement) with 74LVC series easy enough. I built a nice little pulse generator fixture for this and use it with my Tek 475A (1.4ns risetime). Also 74AUC has Tpd of 300ps at 2.5V. I don't get why Potato even have these on the market.

At my work we use a COM Express module that has a Potato Semi part on it. Gotta wonder what the backstory is on that company.

If you want to get experience with 7400 series, i had a great time building an 8bit computer following this YouTube series https://eater.net/8bit/

I've had a lot of fun over Christmas assembling a Gigatron TTL - a computer built almost entirely out of 74-series chips. It's amazingly easy to follow exactly how it works just by looking at the schematics. (More build photos coming on my blog in a few days: https://rwmj.wordpress.com/)

I looked over the few solder joints that are visible and I think your soldering technique needs some work. First make sure your iron is actually hot enough, a hotter iron works quicker and the result is less, not more heating of the joint. Then, try to work like this: first heat up the joint with just a tiny bit of solder on the tip of the iron, much less than you need for the joint, just enough to transfer heat, do that for 1 to 2 seconds, then feed in just enough solder from the spool to make the joint, withdraw the spool and then the iron and let it cool without moving at all. That way your joints will flow nicely instead of beading above the PCB and you won't have too much solder in the joints. Good luck!

Yup I'm totally sure my soldering sucks. In fact I started out my career as a hardware engineer, and there's a few reasons why I'm a software engineer now :-)

(Edit: To be fair the picture shows the first few joints where I was basically getting into the swing of things. Things improved after that.)

I'm delighted to see this make the front page. A couple of years ago I submitted "50 years of TTL" and several comments were about confusion of "time to life" with the logic. Which is not to make fun of anyone; what I found interesting as the primary technical plaything of my early teen years has long been displaced.

It's easier than ever before to get started playing with circuits, but in many cases why would you? Just whack a mpu or even a RPI or Arduino and get going It's a similar issue with ham radio: this is a great time in many ways to be a ham, but there are so many alternative and less fiddly ways to talk to others.

> there are so many alternative and less fiddly ways to talk to others.

That's the main reason I never became a ham, there was a ton of interesting tech and a couple of very knowledgeable people around the place where I lived as a teen but the majority of them were engaged in hour long chat sessions with their buddies rather than to work on even more interesting tech.

So, now with the internet you can chat to people halfway around the globe without even knowing what EME or Heaviside layer means.

Neglected? Not so much I'd say

It's the basic component for digital electronics experiments (pre microprocessor ones)

Agreed. I think it’s probably more famous than anything mentioned, except maybe the 555.

741. When I was starting it was easier to make simple waveforms with the 555 so I have a warm spot in my heart for it and the SSI TTL parts, even if the 741 has ended up handier over the years.

There was a nice front page post on the 555 a couple of months ago: https://news.ycombinator.com/item?id=18190765

Discrete logic chips are still in use all over the place, I've seen them in reset circuitry. Modern versions come in smaller packages like 4-pin bgas, or sot-23 and are available on Digikey.

The caption under the picture of the Xero Alto states:

>"We could have featured a minicomputer such as a DEC PDP-11 as an example of a CPU built from 74 series logic. But the 74-driven Xerox Alto makes a greater point about 74 logic as the progenitor of modern computing devices."

Can anyone say how many discrete logic chips the CPU for the PDP-11 or the Alto used in the design of their CPUs?

Might anyone have some links or suggested literature regarding these pre-microprocessor CPUs?

Plenty of literature, including schematics, can be found here:

http://www.bitsavers.org/pdf/dec/

http://bitsavers.org/pdf/xerox/alto/

Wow this is a real treasure trove, thanks!

O yeah, good old K155LA3 (the Soviet clone of 7400). Was very popular among the electronics enthusiasts.

TI and Fairchild make a 6 pin “tiny logic” part that can be configured for any gate. I use them frequently for glue logic.

Reply on news.ycombinator.com