11

Why do we use 3.3 V in circuits? Why not 3 V? Is there a specific standard? Or do we use it because it was 3.3 V from the beginning?

I am just curious to know.

Peter Mortensen
  • 1,676
  • 3
  • 17
  • 23
Mohammad Kholghi
  • 213
  • 3
  • 7
  • 2
    What is the context? – Dom Fomello Jan 24 '21 at 10:48
  • 1
    See https://www.eevblog.com/forum/projects/why-3-3v/ – Stefan Wyss Jan 24 '21 at 10:57
  • 2
    Because it's 67% of 5V maybe? Why 67%? In digital electronics, min. 67% of the logic voltage level (5V, 9V, 10V, etc.)is considered as logic-hi; likewise, max. 33% of the logic voltage level is considered logic-low. I really don't know the answer. Just guessing. – Rohat Kılıç Jan 24 '21 at 11:00
  • @DomFomello Just curious to know. – Mohammad Kholghi Jan 24 '21 at 11:30
  • 1
    @Rohat Kılıç: No, it is not arbitrarily divided into (equal-sized) regions. It depends on the actually technology used (transistor types, [circuit topology](https://en.wikipedia.org/wiki/Topology_(electrical_circuits)#Components_with_more_than_two_terminals), etc.). E.g., [for TTL the undefined region is 0.8 V to 2.0 V](https://en.wikipedia.org/wiki/Transistor%E2%80%93transistor_logic#Interfacing_considerations). CMOS is different for the same 5 V supply voltage. – Peter Mortensen Jan 25 '21 at 08:23
  • 1
    Possible duplicates: *[Why does much electronics operate on 5 and 3.3 volt?](https://electronics.stackexchange.com/questions/42873/why-do-many-electronics-operate-on-5-and-3-3-volts/42877#42877)* and *[What's with the operating voltages: 5 V, 3.3 V, 2.5 V, 1.8 V, etc.?](https://electronics.stackexchange.com/questions/6403)* – Peter Mortensen Jan 25 '21 at 08:44

2 Answers2

18

Why do we use 3.3V in circuits? Why not 3?

Why not \$\pi\$ V? Why not 2.8 V, a much nicer number than 3 V? The more things like power consumption or speed matter, the less you align with "human-pretty" numbers, and more with physical needs.

In this case, the physical need is actually "something slightly above 3V, but less than 5V, to save power in our new LVCMOS circuits", ca 1970.

Point is: when TTL (transistor-transistor logic) was still the dominant technology for integrated logic, supplies lower than ca 4.5 V were impossible, due to large collector-emitter voltages in the bijunction transistors used there. Hence, with a bit of headroom, 5.0V became a standard.

Now, CMOS was introduced, and it could work well with supply voltages down to ca 3 V. People wanted to give a little headroom.

So, my guess here is why it's 3.3V, and not 3.15V or 3.4 V: they picked a voltage for which there were already voltage regulators ready, in the drawers: 3.3V, which had, interestingly, already been (one) supply voltage of the Apollo Guidance Computer, so NASA and early semiconductor companies had poured in money into building these.

TL;DR: > 3 V: Need that, 3V is just a tiny bit too low for reliable CMOS logic gates at the time of invention. 3.3 V: probably because hardware for that voltage already existed in the early 1970s.

Marcus Müller
  • 88,280
  • 5
  • 131
  • 237
  • 1
    Maybe 5 volts to suit 6.3 volt capacitors and 3.3 volts to suit 4 volt capacitors (just another angle or another factor). – Andy aka Jan 24 '21 at 12:10
  • 1
    sure! Though I'd assume that capacitors and supplies kind of co-evolve; compare the availability of 2.5 V and 2 V tantalum and ceramic caps these days to potential demand for them in, say, 1999. – Marcus Müller Jan 24 '21 at 12:21
  • 6
    3.3 is also a preferred value, it's in the E series. – Hearth Jan 24 '21 at 13:57
  • @Hearth which might be the original reason why 3.3V was used in the AGC. – Marcus Müller Oct 14 '21 at 11:58
9

One useful aspect of choosing 3.3V for LVCMOS is that it allows good compatibility with 5V TTL levels due to similar VIL/VOL/VIH/VOH levels.

TEMLIB
  • 3,347
  • 1
  • 14
  • 21