Hi, I'm just wondering how transistors are named. They're often a number like 2N#### or TIP##. Are the names systematic? How do I know which ones to use for a given application?

2N3904 seems to come up a lot. Is there anything special about it? What are some other common models?

 

There is some information near the bottom of the Wikipedia article - http://en.wikipedia.org/wiki/Transistor

And, for the 1N & 2N names - http://en.wikipedia.org/wiki/JEDEC

2N3904's are cheap and work well as switches. Sort of like 2N2222's and 2N4401's.

 

1N for diodes, 2N for transistors was recommended by Jedec; TIPxxx means Texas Instruments Plastic (body)
Numberings were chosen by original manufacturers production.

 

Why name part numbers at random?

Why both NPN and PNP transistors has to be 2Nxxxx, why can't it be 2N for NPN and 2P for PNP? Would people dies if a part is call 2PXXXX?

Or why not the JIS way of 2SA,2SB,2SC and 2SD, often shorten without the "2S" to mean high/low frequency, PNP and NPN.

The European way with three letters make senses and at least some basic information can be learned about the part from just reading the part numbers.

The naming of MOSFETs with type, current and voltage are good. For say 50N20, one knows immediately the MOSFET is N-Ch rated at 50A and 200V. You think so too? No, not good enough for some manufacturers.

They have to come up with random part numbers.

 

Hello,

Here is a page from the EDUCYPEDIA on the semiconductor codes:
http://www.educypedia.be/electronics/transistorcode.htm

This is a link from the component code page:
http://www.educypedia.be/electronics/datacomponent.htm

Greetings,
Bertus

 

Mr. I understand your question but nowadays you'll have to look at the datasheet if you want to know some basic facts. Naming methods are nothing you have to learn really.

 

I wish the naming of transistors was organised a bit better, but there are 3 main naming systems (summarised, as BeenThere mentioned, at wiki: http://en.wikipedia.org/wiki/Transistor) and they give some clues to the device from the name....

1. Pro-Electron (European) naming scheme:
(nnn=3 digits or letter+2 digits, sometimes followed by letters to specify gain groups, voltage rating, etc.)
ACnnn = Germanium audio transistor (old) (PNP often are even numbers)
ADnnn = ... power ... (AD14x usually TO3 case)
AFnnn = ... RF (high frequency) transistor (AF11x: often TO-7 case)
AA...,AY...,AZ... = Germanium diodes, rectifiers, and zener diodes

BCnnn = Silicon audio/general purpose transistors.
BDnnn = Silicon power transistors
BFnnn = Silicon RF transistors, or JFETs
BLnnn = Silicon transmitting transistors (power+RF)
BPnnn = Silicon phototransistors
BSnnn = Silicon switching transistors (and some MOSFETs)
BTnnn = Triacs, SCRs, etc.
BUnnn = Silicon very high voltage, e.g. TV Horizontal deflection
BAnnn, BBnnn,BYnnn & BZnnn are diodes,dual/varicap diodes, rectifiers & zeners

2. Everything else:
1N.... = diodes (or rectifiers), but 1N5GT is a pentode!
1S... = diodes/rectifiers
2N.... = (JEDEC) transistors, Triacs, SCRs, JFETs or MOSFETs (but not dual-gate)
PN.... = transistors etc. (Often the plastic case equivalent of 2N....)
MPS... = transistors - often the Motorola plastic case equivalent of 2N....
MPSA.. = bipolar transistors, sometimes Darlingtons, originally Motorola brand, PNP complement found by adding 50 to the 2-digit number, e.g. MPSA05 -> MPSA55
MJ.... = Motorola power transistors; numbering slightly related to JEDEC, e.g. MJ2955 is PNP version of 2N3055 (you see why I say SLIGHTLY!)
MJL.... = Motorola "linear" (hi-fi) transistors, sometimes the numeric part of the type number relates to a Toshiba 2S.... part that it copies.
2SA... to 2SJ... (see wiki article for these Japanese types; be aware that a J176 might be two different things - an abbreviation of 2SJ176 or a completely different animal!)
ZTX... Zetex brand transistors - just sometimes the number following ZTX corresponds to a BC... number
NKT... really old Newmarket transistors, just included to show there are zillions of prefixes (and all those RCA numbers without prefixes) that make it hard to follow.

Having many standards is just another way of saying there is NO standard! Some time I feel I should explain how the transistor naming scheme could be made a bit better organised - although, realistically, there is a limit to how far people should try, so I somewhat agree with peranders saying you have to look at the datasheets.

 

Just a postscript to the Pro-Electron naming standard:

many (but certainly not all) transistors whose names follow the pattern BC..[789][ABC]* have a bit of a pattern that aids recognition of several attributes:

1. These are silicon general purpose transistors (as the "BC" implies")
2. If the last digit is: 7/8/9 the transistor is high voltage (well, 40 volts or so is high, isn't it?) if 7, cheap 20Vceo if 8, low noise low voltage if 9; then...
3. the first 2 digits give the case and the polarity (2nd digit >=5 means PNP in some, but not all, of this subset of the BC... range). So BC107/8/9 is NPN TO-18, BC177/8/9 is PNP TO-18 case; BC147/8/9 is lockfit (obsolete) NPN, BC157/8/9 is lockfit PNP; BC547/8/9 is plastic TO-92 NPN, BC557/8/9 is TO-92 PNP
4. a following A means low gain, B means medium, C means high gain, so a BC549C would be a high gain, low noise, low voltage TO-92 transistor.

The vast majority of the BC... range don't follow that pattern, but the ones that do are those that have been the most popular over time. The BC546 to BC560 are popular TO-92 transistors, of which most follow the pattern, except note: the BC550 is NPN (basically a higher-spec BC549, like a BC560 is a high-spec BC559).

 

In past years R.C.A. had yearly tube and transistors manuals that
gave detailed Information.Since overseas brands started hitting the markets
and withdrawing Information about there products.

 

Every so often I also wonder whether transistors could still get a true standard for naming, after all these years. Is it too late? Is there no hope for any scheme other than a chronological sequence number plus little else? The European and Japanese naming systems are much better than the JEDEC (2N....) series and most proprietary naming systems, but still it is hard to spot valuable information - like what is the complement - just from their numbers.

Just as a thought experiment - let me know what you think - I came up with a possible naming system that is both useful, extensible (future-proof), compatible with old systems to a reasonable extent, and not very complicated to learn (I think):

Transistors and diodes would start with a 2-letter prefix based on the Pro Electron standard (e.g. "BD" for a power transistor), then a letter or letter/number combination to give the case+pinout type, e.g. "K" for TO-3 with the standard CBE pinout (or DGS for FETs), then "N" for NPN or "P" for PNP, or even NNPP for a package with two complementary pairs. Then would come the "base type number", which would be 4 (or less) digits if it is referring to an old device (e.g. a 2N3055 would be given the name BDKN3055 in the new system), or a 5 (or more) digit number assigned to new devices, which would convey more information, and I would expect that every time a 5-digit number is allocated under the new scheme the corresponding 2N.... JEDEC code and a 2S..... JIS code would also be reserved for the same device. The 5-digit numbers could have a digit to indicate the approximate current rating (e.g. 0=under 100mA, 1=0.1A to 0.9A, etc); another to give the voltage rating roughly (0=under 10V, 1=10-19V, 2=20-29V...9=1000V and above), and another to indicate special properties (e.g. low noise, resistor-input, etc.). There would still be digits for sequence numbers. The option of more restriction on characteristics should be handled by optional suffixes, e.g. a BCS23N3904-H150T10 might be a 2N3904 in a SOT-23 package with an hFE of 150 plus or minus 10%, if selection/matching is important. Although 18 characters, in this case, is quite a lot to identify a device, if specially selected devices are not required the length is quite manageable, and the redundant information reduces the chances of a single misread character in a transistor's marking causing great problems. I could describe my ideas for a system in greater detail if anyone is interested.

Why name part numbers at random?

Why both NPN and PNP transistors has to be 2Nxxxx, why can't it be 2N for NPN and 2P for PNP? Would people dies if a part is call 2PXXXX?

Or why not the JIS way of 2SA,2SB,2SC and 2SD, often shorten without the "2S" to mean high/low frequency, PNP and NPN.

The European way with three letters make senses and at least some basic information can be learned about the part from just reading the part numbers.

The naming of MOSFETs with type, current and voltage are good. For say 50N20, one knows immediately the MOSFET is N-Ch rated at 50A and 200V. You think so too? No, not good enough for some manufacturers.

They have to come up with random part numbers.

 

Good luck with that, you would have better luck herding cats than 100 or more transistor manufacturers.

Truth, no standard will win. You might be able to come up with a convention that has the important specs, such as unity gain, DC gain, type, manufacturing process, and whatever else is important embedded in a part number, but you would have to start from scratch. People will want their system to be on top. Politics shows in odd places.

At least this way you could have a built in cross reference, matching a standard part (or parts) to the new system.

The other thing you will fight is tradition. I like the 2N2222. This is not a rational thing, it is what I've always used, so I want to continue using it because I know it. The military (whose deep pockets has equally deep influence) also likes what they like, and from what I've seen it isn't always rational either. It is a powerful thing to overcome.

 

The following information is added mainly for historic completeness since Maitchy has done such a good job already.

When diodes and transistors were first manufactured the principal UK manufacturer was Mullard.
They were a highly successful valve manufacturer and they followed the valve naming conventions.

So the first character was anumber denoting the heater voltage = 0
The next character denoted material usually A = germanium
Then followed a type serial number

so for instance the OA81 germanium diode was introduced before the OA91.

Diodes were introduced before transistors.

For transistors, called crystal triodes at the time, the non existant heater was dropped and the letters meant

First letter A=germainium B=silicon
Second letter C = audio/general purpose F = radio Y or S = switching

Again a serial type number was added

So an AC128 was a germanium audio transistor
A BC109 was a silicon general purpose transistor
An ASY69 was a germainium switching transistor
A BSY65 was a silicon switching transistor
A BFY90 was a silicon RF/switching transistor

And so on

The Odd man out was the OC series of power transistors which were germanium power transistors. There are many stories about the progeny of their numbering but the truth is lost in the mists of time.

 

Many distributors' websites will start with tens of thousands of transistor models and you (the customer) gradually whittle it down by "filtering" the results by transistor type, bias voltages, gain, power limits, and so forth. That allows you to find the type of transistor you want. The part names and numbers are arbitrary. Mouser and Digikey have this system in their web catalogs.

 

Hi all,
very interesting posts on Transistors...

talk about prefixes & codes to the Transistors.. say a Circuit diagram, needed a 2N3702<old) and a newbie looked at it as E,B,C ... and connected in that manner.. well would not work unless like the above people said look at data-sheets.. because the collector of the above transistor is in the Middle!!! going e,c,b.... now thats confusing is it not for the newcomer..
eyes rolling here..!!!

 

Even better would be to rename electrons as positively charged. It would make the thought processes and math so much easier! Then there is the symbol for an NPN transistor with the arrow pointing at where the electrons come from instead of the way they flow...and Mosfets! One end is called the drain, regardless of whether it is draining electrons or supplying electrons!

I'm kidding. No, I'm serious. No, I'm confused. Irrelavent or stupid looking numbering and naming schemes seem to be the rule, not the exception. It's just one of the things we have to deal with. History happened, and it happened when people were a lot less knowledgable than they are now. I have no expectation that all the stupid naming conventions will ever be "fixed".