I'm a HVAC tech who is interested in electronics. I have 3 kids and not a ton of time to really get involved with the data sheets or a real understanding of electronics. I just want to make some simple electronic devices. Some things I've done are a solid state water make up control for the Keurig and a peltier heater / cooler. My problem is that I don't really understand transistor ratings and abilities. Is there a, or a small set of, go to transistors that have a wide range of applications?

  • 2
    \$\begingroup\$ Welcome. Shopping questions are off topic on this site, but maybe someone could toss you a few of the most common we see all the time. They also have many manufactures and vendors, so they are trusted to be around for a while. \$\endgroup\$
    – user105652
    Feb 16, 2019 at 4:04
  • 1
    \$\begingroup\$ For analog stuff, I tend to select the 2N3094/2N3096 BJTs. For power applications, whenever I can get away with a single n-type FET, I would select any logic level nMOSFET that is easily accessible to me, unless some special need dictates a different model. \$\endgroup\$
    – nanofarad
    Feb 16, 2019 at 4:16
  • \$\begingroup\$ That link actually shows examples of the types of things I'd like to try. Thanks Andrey \$\endgroup\$
    – Joe Fala
    Feb 16, 2019 at 4:29

3 Answers 3


Yes, there are some common general purpose transistors, based on current handling. These do not take into account bandwidth (e.g. for RF appications), or high voltage (>30V).


  • 2N3904/2N3906 (SMD MMBT3904/MMBT3906) (upto about 100 mA)
  • BC547/BC557 (higher gain than 3904/3906, lower current)
  • 2N2222/2N2907 (upto 250 mA)
  • BD139/BD140 (upto 1A)
  • TIP122/TIP127 (darlington; upto 3A)

Current limits are conservative, about half of what datasheet says.


  • 2N7000 (SMD 2N7002)
  • BSS138 (logic level)
  • AO3400/AO3401 (not that common, but very useful)
  • IRFZ44/IRLZ44(logic level)/IRF4905 and IRF540/IRF9540
  • 1
    \$\begingroup\$ I just wanted to add that the pairs of BJT part numbers separated by a / are the NPN and PNP transistors that are often used together. For example, if you find that a 2N3904 is a good NPN transistor for your application then the 2N3906 is probably a good PNP transistor for the same application. \$\endgroup\$ Feb 16, 2019 at 16:08
  • \$\begingroup\$ @ElliotAlderson Yep. I left that out hoping that whoever reads my answer will be forced to go look up the datasheets before plugging in the transistors. (although I should mention that the 2n2222/2n2907 datasheets for the TO92 are a bit difficult to locate). \$\endgroup\$
    – Indraneel
    Feb 16, 2019 at 16:48
  • \$\begingroup\$ @Indraneel I don't know about the 2907, but the 2222 in a TO-92 (Plastic) package has a separate part number; PN2222. \$\endgroup\$
    – Hearth
    Feb 16, 2019 at 16:55
  • \$\begingroup\$ @Hearth 2N2222 and 2N2907 both exist also as TO92. At least ST makes them next.gr/components-datasheets/2N2222.pdf although I don't know who made the ones I have. Also, I don't know how different they might be from the 3904/3906. They are priced the same, do they really skimp on the bond wire? \$\endgroup\$
    – Indraneel
    Feb 16, 2019 at 17:01
  • \$\begingroup\$ @Indraneel Honestly, "what makes these parts different" seems like it could be a very good question for this site. Why not ask it and see if you get some answers from people who know more than I do? If you don't, I will, because now I'm curious about the details too. \$\endgroup\$
    – Hearth
    Feb 16, 2019 at 17:21

All transistors have both an electrical resistance as a switch Rce=Vol/Iol or RdsOn @ Vgs ( well above Vgs(th)) and a thermal resistance based on package size which is added to heatsink thermal resistance to compute junction temp rise or case temp which you want < 45'C so it does not burn fingers.

E= old schoole Voltage , = U = EU school = V American school term for voltage
Rce ( collector -emitter ) incremental resistance as a switch
Vol = V output low for NPN on collector-emitter
Iol = I output low for NPN ( for some input current typ 5~10% of Iol)
RdsOn = Drain-Source On resistance ( like collector-emitter) for some Vgs >2Vgs(th)max
Vgs(th)max= the threshold conducting as a weak switch (1mA) for the Gate to source voltage

Power loss from Ohm's Law is then enter image description here

FET's can have much lower resistance and have almost no input DC current when On but since capacitance increases as RdsOn lowers, the switching current can be large so a gate drive resistance affects slew rate. T=R*Ciss

Transistors have high current gain hFE yet when saturated as switches tends to reduce to 10% of max hFE >1 stage is often needed. specs give Vce(sat) @ Ic for Ic/Ib=10 to 20.

SMD packages are more common and if you learn to solder these with thermal pads and ground plane heatsink at 1W/sqin as a rule, you can make/buy small PCB blank cards or with logic level FETs the drivers can be driven from 3.3V CMOS logic. Use ESD protection when handling and filter interface wires for same protection.

Digikey & Mouser etc have over 50k different transistors to choose from with searchable parameter filters

  • \$\begingroup\$ +1 for the Ohm's law wheel twice in one day \$\endgroup\$
    – user156429
    Feb 16, 2019 at 9:01
  • 1
    \$\begingroup\$ +1 if you were laughing out loud \$\endgroup\$ Feb 16, 2019 at 9:05

Along with your go-to transistor types request, here is a guaranteed-to-work amplifier circuit, self-centering bias, that works on 1.5 volt batteries and higher


simulate this circuit – Schematic created using CircuitLab

  • \$\begingroup\$ Guaranteed to work for a very low frequency input? \$\endgroup\$ Feb 16, 2019 at 16:10
  • \$\begingroup\$ dah caps prevent dc operation \$\endgroup\$ Feb 16, 2019 at 17:45
  • \$\begingroup\$ So not guaranteed to work. Just wanted to clarify that. \$\endgroup\$ Feb 16, 2019 at 19:05
  • 1
    \$\begingroup\$ Unlikely to work at 10.453Ghz either. Might that be an issue for you? \$\endgroup\$ Feb 17, 2019 at 22:43
  • \$\begingroup\$ I just think that you shouldn't say that a circuit is "guaranteed-to-work" if it has significant limitations (at dc, not 10 GHz) ...not for my sake, but for the sake of newcomers who might take you at your word. Things that seem obvious to you and me are not so obvious to those who are just starting out, and a bad experience due to a circuit that doesn't work as advertised might cause them to give up on learning electronics. I think that would be a shame. \$\endgroup\$ Feb 18, 2019 at 2:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.