BCX70J is my go-to gumdrop NPN transistor, and they cost 2.4 cents each. I was mildly offended when you listed it as the dead last, absolute worst, highest noise transistor on page 501.
Not that I use many small-signal bipolars any more.
Hah, 2N4401 (to-93) is my first choice.
George H.
Last year I bought a whole pile of TO-92 parts because they're going away. So I'd say that for hand-wired protos mine is either an MPSA18 or an MPS5179.
Cheers
Phil Hobbs
Hey, at least it made our table! It's got the highest r_bb' at 760 ohms. It's a small-die part, not well suited for many types of muscular jobs.
The well-performing ZTX618 / FMMT618 is a symmetrical transistor, reducing the distinction between the C and E terminals. (In old days we used such parts as precision switches, i.e. a bipolar 16-bit ADC with 10-volt reference might require a 0.3mV switch. We enjoyed sub-mV accuracy for our precision-switched signals.) Yes, it costs $0.30.
I remember dual-emitter chopper transistors. Top that!
2N7000.
Less base current. Saves energy. Reduces global warming.
"Collector" is kind of poorly named, in regards to general (bidirectional) transistor operation. It's still an emitter all the same; just usually not a very good one. More of a... tinkler, than an emitter?
. T . o . | . | / . |/ . B o---| . |\ . | V . | . o . E
Tim
I bought a bunch of those too, in both protected and unprotected versions. That salvage attempt was one of the better 200-buckses I've spent recently.
Cheers
Phil Hobbs
Hah, Hah. I measured a 2n4401, and found ZERO inverted beta! BTW, the three quasi-symmetrical Zetex transistors I tested had symmetrical Vce(sat) voltages for current in or out of the collector. The 2N4401 had a rather high current-in (positive) saturation voltage, +50mV at +5mA (sat beta =5), but was OFF for current out of the collector. Sheesh! The collector diode stole all of the base drive.
If that's reliable, it's potentially useful as well.
Long ago, Widlar published an app note about a "temperature compensated breakpoint amp" circuit that used one or more emitter followers hung on the summing junction of an inverting amp, with resistors from the overall input to the emitter, and from the output to the collector.
When the transistor saturates, the collector resistor gets connected to the SJ in parallel with the main feedback resistor.
If the inverted beta were zero, the breakpoints in the two quadrants would be usefully different.
Cheers
Phil Hobbs
Given that anybody can make and sell 2Nxxxx parts, I wonder how similar different parts will be. The processes could be all over the place.
I remember when RCA 2N3055s were great parts, then Fairchild started selling 2N3055s that were 5x faster, 1/4 the die area, and
1/1000000000 the lifetime.
Right, I didn't even check to see who made my part.
OK thanks, I've never tried running them inverse, (Except by mistake :^)
George H.
Oh good idea! All my 2n4401's are from fairchild. I should order some others and look at them.
George H.
Yes, IIRC they used an epitaxial process instead of the slow and lower beta but more robust homotaxial process.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required