2N2222A: they're all backwards, Jim! (P2N2222AGD)

The PN2222 datasheet at ON semi has the leads exactly in the usual EBC order. There's LOTS of makers for PN2222 and PN2222A, and all use EBC order.

Your part (P2N2222AGD) is the wierdo. Hopefully, it won't confuse you or the local grad students for long. You have 'a drawer full', eh? How did that happen?

Seems to be consistent with their datasheet.

Jon

Yikes, the OnSemi PN2222 is CBE, and everybody else seems to be EBC!

Lots of people used to make BCE transistors.

We use one diode that has the band on the anode! And a surfmount 1206 LED with a corner bevel on the anode. And one surfmount SOT143 transistor that has the index dot on the *bottom* of the part.

Don't get me started on electrolytic cap markings.

John

You have to be cautious, the pin order is package style dependent.

Package 92 is E-B-C, package 97 is C-B-E ...Jim Thompson

Interesting. I recently received a bunch of the OnSemi parts. But I haven't yet used them.

Looking at the datasheet for the P2N2222A from OnSemi, and orienting the BJT's TO-92 plastic so that I'm staring at the flat face with the leads pointing downward, the datasheet says that the pins are numbered, left to right, as 1, 2, and

1. And that 1=C, 2=B, and 3=E. Or CBE.

Which is what you observe. So it seems the datasheet does, at least, say so.

Fairchild's PN2222A datasheet, using the same observational orientation, shows EBC. So again, just as you say, OnSemi has it different in their P2N2222A devices.

I suppose this says "read the datasheet?" Also, note the slight difference in nomenclature. OnSemi marks these as P2N2222A -- the '2' between the 'P' and the 'N'. Chances are, there was a reason for that.

You know? I bought a bunch of these P2N2222A from OnSemi for "real cheap" (about 4 to a penny) and it is possible that this pin difference was the real reason I found them priced so well.

Jon

Yeah. I first had to get used to the BCE orientation. Then I had to get used to the EBC. Then the CBE. It's a pain.

Any idea what motivates these variations? What's in it for the manufacturer -- captivity?

Jon

I think C in the middle was physically sensible, the middle lead being the chip paddle. Power transistors and power mosfets are mostly still that way, BCE or GDS.

EBC may have been for easier stuffing into TO-5 footprints?

CBE makes no sense at all.

I like gaas/phemt power transistors, because the middle lead/case is the source, which usually wants to be both heat sunk and electrically grounded.

John

))))))

u/wbeaty/

A lot of the Japanese transistors have that pinout in a TO-92 looking package...

Tiny plastic drawer

Digikey was out of stock on PN2222A, but their search page showed stock on P2N2222A. I couldn't see any obvious diff in datasheet. (This was a couple years ago, and I just noticed it now!)

Hmmm. NTE crossref search on 2N2222 shows "P2N2222A(CBE)." But their actual replacement part is EBC.

((((((((((((((((((((((( ( ( (o) ) ) ))))))))))))))))))))))) William J. Beaty Research Engineer beaty, chem washington edu UW Chem Dept, Bagley Hall RM74 billb, eskimocom Box 351700, Seattle, WA 98195-1700 ph 206-543-6195

I can imagine some OEM picking the wrong one and having to send them back. (Datasheet values seem the same, so P2N2222A equals PN2222A, right? DOH!) Suppliers probably can't re-sell, since customer returns aren't guaranteed to have avoided ESD exposure if the packages aren't sealed.

What good are huge numbers of NPN discretes? Well, this guy Dieter build a DIY computer, the "MT15," not from TTL, not from gates, but from NPN parts in SMT:

In Deutch it's "sauerkraut" wiring not "spaghetti:"

OK now let's see some discrete transistor RAM. Put LEDs on each flipflop and clock it real slow, like the "computer" in Voyage to the Bottom of the Sea:

Front:

Earlier version, better sound effects:

((((((((((((((((((((((( ( ( (o) ) ) ))))))))))))))))))))))) William J. Beaty Research Engineer beaty, chem washington edu UW Chem Dept, Bagley Hall RM74 billb, eskimocom Box 351700, Seattle, WA 98195-1700 ph 206-543-6195

It is PACKAGE dependent.

If you were not just a lay person taking a stab at the industry, you would know that.

The metal cans are yet another configuration.

Only idiots cry about because they don't know, and then they typically show even more stupidity about it after being told because their common sense factor hovers near nil.

Same type of folks that claim to be civil, but flip like a switch into a little gang boy retard and assault someone because they think "they'll get away with it". Too stupid to even understand criminality.

ON conventional lead-out part numbers are PN2222A or MPS2222A.

The P2n2222A and P2N2907A both have the reverse of convention.

Fairchild, MCC, DI, STM and others also all make parts with your expected lead-out.

RL

I'm guessing, cheaper lead frame.

Cheers! Rich

Nah - just attach a note to the assembly instructions saying for this particular unit, install it the other way around. (i.e., rotate it 180 degrees), or if you're going to do more than, say, a thousand units, change the silkscreen.

I would do this, but people will probably jump all over me for using the KISS (Keep It Simple Stupid) principle. ;-)

Cheers! Rich

Some decades ago, our apartment took a direct lightning hit. According to the manager, it caused upwards of $3000.00 damage. Luckily, I had unplugged the modem; but it killed the answering machine, and the transistor in the TV remote's on/off circuit failed short - I couldn't turn the TV off!

So I dragged it to the dealer (Sears, or Monkey Ward or something) and asked the tech to simply diagnose it, and tell me which transistor it was (they've got the schematic and all.) I had a pile of NPNs with a different pinout than the original one, so I took a piece of teflon sleeve, and physically swapped the two leads (I think one was EBC and the other was BCE or something.) The diagnosis was $30.00, the repair would have been about $75.

I was happy. :-)

Cheers! Rich

Think circuitry.

The mil spec stuff was in cans.

So, the epoxy stuff was consumer/commercial grade.

Now think about computers and logic circuits built from discreet components, which many were way back when. Now think of circuit layout complications.

They were probably made to make the building of certain logic circuitry easier, so GE or someone somewhere along they way made a new package to specifically allow it to be made without getting banged for not complying. Or Motorola. It could have been the master manufacturer that decided to do it, or it could have been customer demand driven.

It really only comes down to form factor. Form factors get "designed" by folks way back when they had to worry about circuit real estate and a host of other little nasties. Even simple via count was important, and a different pinout package may well have made a certain company build hundreds of thousands of boards with millions of these in them with less hassle than the original package gave them.

Not likely lead frame at that scale. Three pins is three pins is three pins. Pretty much period.

Scrape the old silk screen and give the assemblers a deviation instruction (without the silk screen, it would not be a deviation).

That could be what happened. But I'm ignorant about it, so no real idea except that I got a nice price.

You trying to get me to use the damned things myself instead of giving them away? ;)

Uh, too much work in these days of FPGA. I'm old, but I'm not that retro. But don't forget this (it was built a few miles from me):

Had to use the wayback machine as I think the regular site is not functioning, anymore.

Jon

Fairchild caught me out on 2N7000, only a packaging addendum had the correct pinout, I found it by looking for the source drain diode.

Stuff happens.

Grant.