Why are ICs numbered?

When we take any IC, it has a certain number?
Is there a specific convention/methodology for numbering the ICs?

For ex. A Nand gate IC is referred as IC 7400. Why is the numbering like 7400?
Are they randomly generated or is there a specific rule to name the ICs?

 

It is the manufacturers' choice.

 

It is the manufacturers' choice.

Oh. Ok. But does he use some convention for naming his ICs?

 

Not that I can discern, except that a particular manufacturer may use particular letters before the number. For instance Maxim IC designations start with MAX.

 

But does he use some convention for naming his ICs?

I suspect they started with 00 and assigned the next available number to whatever part they designed later or had planned. Since manufacturers try to become alternate sources for parts, they typically use the same part number somewhere in theirs; though that wasn't always the case.

There's a Wikipedia article on the 74xx chips with some info on part numbering.

 

Why are ICs numbered?

When we take any IC, it has a certain number?
Is there a specific convention/methodology for numbering the ICs?

For ex. A Nand gate IC is referred as IC 7400. Why is the numbering like 7400?
Are they randomly generated or is there a specific rule to name the ICs?

Not totally random. "74xx" specifies the logic family of TTL devices or TTL compatible, anyway. "xx00" specifies the gate type. I suppose some organization assigns this number so everybody's xx00 is the same function. But the "00" can't be broken down meaningfully.
Analog ICs? Even less of a standard. Each manufacturer has their system, National Semiconductor, for instance ... LMxxx stands for Linear Monolithic family of IC. LM741 specifies an op amp with specific characteristics. An AD741 by Analog Devices would have those same characteristics.

 

I always thought the numbering conventions, not the suffix or prefix letters, were set by JEDEC. https://www.jedec.org/

 

I always thought the numbering conventions, not the suffix or prefix letters, were set by JEDEC. https://www.jedec.org/

The standards organization ,JEDEC, currently seems to be more concerned in areas like packaging (IC dimensions and pin arrangements), and signalling. They don't seem to "register" part number any more. You're not going to get a new 2Nxxxx transistor designation anymore.

Back, in the old days of electronics, there use to be many component manufacturers. Competition, forced them to make drop-in compatible devices. A Sylvania 6AU6 would work just as well as an RCA 6AU6 pentode. That system continued into transistors, and ICs. When the first TTL devices came out, the different IC makers did have their own numbering systems, for a while. But as the business "shook out", the 74xx series gained favor, and those IC makers fell in line. TI with its SN74xx series, Motorola MC74xx series, National DM74xx series, and several more. A 7400 from any of those worked just the same, and could be plugged in without any thought.

I started working in the industry, in the late 1970s. A board would usually have a CPU (8080, Z80, early 8088, ...) a few memory ICs, and some large peripheral (IO) chips. All NMOS. Tying it all together, were a dozen or so, TTL chips. 74LS, most common in the late 70's to early 80's period. All chips, big and small, followed a common numbering system. Most of the big NMOS chips followed the Intel 82xx scheme, even though, they were made by several different IC makers. Likewise with the memory chips. But these part numbers were never "registered". They occurred "naturally", by playing "follow the leader".

Fast forward, to the modern era, and most ICs are programmed. The business model has changed, and there is little incentive, to make a pin-to-pin compatible part. So no need for a standard numbering scheme. The individual part makers have their own schemes, but they aren't followed, as much, by competitors anymore. Yes, there are still "jelly bean" (inexpensive, low profit) parts, like general purpose op-amps, that share the same base number. For example, the LM324, is still being made by TI (which swallowed up National), and ST, which can be used interchangeably. But those parts, were introduced decades ago. For new designs, you don't see that.

 

The standards organization ,JEDEC, currently seems to be more concerned in areas like packaging (IC dimensions and pin arrangements), and signalling. They don't seem to "register" part number any more

But they did at one time, correct? That's what I was saying, don't know about now, but seems your correct about that.

 

Why are ICs numbered?

When we take any IC, it has a certain number?
Is there a specific convention/methodology for numbering the ICs?

For ex. A Nand gate IC is referred as IC 7400. Why is the numbering like 7400?
Are they randomly generated or is there a specific rule to name the ICs?

It's SEMI-sorta standardized. 7400 series we know is TTL. 5400 series is CMOS. Anything beyond there is like flipping an Oreo cookie onto a roulette wheel.

 

But they did at one time, correct? That's what I was saying, don't know about now, but seems your correct about that.

Yes, they did. Transistors and thyristors (TRIACs and SCRs) with numbers started with 2N. Diodes and rectifiers with 1N. Early MOSFETs with 3N, and optocouplers with 4N.

 

It's SEMI-sorta standardized. 7400 series we know is TTL. 5400 series is CMOS. Anything beyond there is like flipping an Oreo cookie onto a roulette wheel.

The 5400 series was the mil-spec version of the 7400 series. Parts like 74LSxx, 74Sxx, 74ALSxx, used bipolar transistors, and true TTL. 7400 parts, with the middle letters like 74Cxx, 74HCxx, 74HCTxx, and 74ACxx are CMOS, and use complementary P and N channel MOSFET transistors. The original CMOS series was the RCA 4000 series.

4000 CMOS operated a lot slower than 7400 TTL. For example a 4000 series counter could only clock at ~3MHz max, while a 7400 counter could do close to 30MHz. CMOS selling point was much reduced power usage. Speed versus power is were the middle letters come into play. It started with no letters, the original series. Then they played with sacrificing power for speed, and the other way around. 74H00 and 74L00. The 74H00 was faster, but burned more power. The 74L00 used less power, but at the sacrifice of speed. Some bright guy got the idea of sticking Schottky diodes across the base and collector, that clamped saturation, and made it real fast. Then 74S00 and 74LS00 came out. 74LS00 was as fast as the original 7400, but only used a fifth of the power.

74C00 was basically 4000 CMOS with 7400 pin outs, and made by National Semiconductor. By ~1983, semiconductor processes came into use that allowed for faster and higher powered MOSFET transistors. FETs that could pass 10s of amps came into being. But you could also make smaller ones for logic ICs. The 74HC00 and 74HCT00 series were introduced at this time. They could switch as fast as 74LS00, but were CMOS. Later this technology was also applied to large chips, like the processors. An older NMOS Z80 would draw ~120mA, while its CMOS version only drew 15mA, and that was at the same clock speed.

 

The 5400 series was the mil-spec version of the 7400 series. Parts like 74LSxx, 74Sxx, 74ALSxx, used bipolar transistors, and true TTL. 7400 parts, with the middle letters like 74Cxx, 74HCxx, 74HCTxx, and 74ACxx are CMOS, and use complementary P and N channel MOSFET transistors. The original CMOS series was the RCA 4000 series.

4000 CMOS operated a lot slower than 7400 TTL. For example a 4000 series counter could only clock at ~3MHz max, while a 7400 counter could do close to 30MHz. CMOS selling point was much reduced power usage. Speed versus power is were the middle letters come into play. It started with no letters, the original series. Then they played with sacrificing power for speed, and the other way around. 74H00 and 74L00. The 74H00 was faster, but burned more power. The 74L00 used less power, but at the sacrifice of speed. Some bright guy got the idea of sticking Schottky diodes across the base and collector, that clamped saturation, and made it real fast. Then 74S00 and 74LS00 came out. 74LS00 was as fast as the original 7400, but only used a fifth of the power.

74C00 was basically 4000 CMOS with 7400 pin outs, and made by National Semiconductor. By ~1983, semiconductor processes came into use that allowed for faster and higher powered MOSFET transistors. FETs that could pass 10s of amps came into being. But you could also make smaller ones for logic ICs. The 74HC00 and 74HCT00 series were introduced at this time. They could switch as fast as 74LS00, but were CMOS. Later this technology was also applied to large chips, like the processors. An older NMOS Z80 would draw ~120mA, while its CMOS version only drew 15mA, and that was at the same clock speed.

Ahh...thanks for the clarification...you are correct. I'd forgotten about the mil-spec designation....we had such a mixed bag back then....yikes..how long ago was that! We also had a lot of ECL....but totally forgot the designation for that.

 

...yikes..how long ago was that! We also had a lot of ECL....but totally forgot the designation for that.

I think ECL came in 10K and 100K series (?). That was a long time ago. You had to ask Seymour Cray, when he was alive. He used a lot of it. Packed in tight, in his computers. All soaking in Fluorinert. A coolant related to freon. Not liquid nitrogen, like some think. I worked with a guy, that serviced Cray computers at one time.

 

The first Cray computers were massively air cooled and cold water conduction cooled. You are referring to the Cray 2, affectionately known as "Bubbles".

ak

 

We also had a lot of ECL

in the 1970's I worked for a company that made a synthesised (PLL) signal generator which worked up to 350MHz and the first stage of division used ECL. I was employed repairing these and the ECL really threw me the first time I came across it. It used strange voltage levels and the oscilloscopes we had struggled to display the frequencies involved.