Favorite BJT transistors, sot-23 and TO-92 equivalents - I

Oops! Good catch, I meant BFR93 for the NPN. Thanks!

Even with damping resistors? E.g., EF outputs look like an inductor, due to the falling beta (assuming a low source impedance). L = Rs / 2pi fT, above the breakpoint frequency, f = fT r_e / Rs. That output inductance resonates with capacitive loads, but a damping resistor can keep Q = 1 max.

--
 Thanks,
    - Win

Good, good.

--
 Thanks,
    - Win

Emitter followers, diffamps, cascodes, even plain common-emitter amps like to oscillate with hot transistors, like the zillion-GHz NEC parts for example. By the time you add enough base resistance to calm them down, you may as well have used a slower transistor.

The RF boys know how to design tuned circuits that are stable.

I've played with fast NPNs, including some 45 GHz SiGe things, as switches. They are disappointingly slow. PHEMTS, on the other hand, switch really, really fast.

John

Are you going to mention MMICS somewhere else? Some of them make radical wideband amps, even pulse amplifiers, so are options to replace discretes and opamps. They tend to be very stable.

John

OK, we're fans as well, they'll go into 2x.

--
 Thanks,
    - Win

Here's a couple of On Semi's low Vce sat bjt's.They are a little pricey but still cheaper then diodes comparable FMMT series. The only reason I have them is Newark puts them on sale time to time like right now I see.

NSS40201LT1G - NPN 40V 4A SOT-23

Complement

NSS40200LT1G- -PNP 40V 4A SOT-23

Two more of On semis low Vce sat bjt's I have are below . They are SOT-223 though. Reasonably priced to.

NJT4031N - NPN 40V 3A SOT-23

Complement

NJT4030P - PNP 40V 3A SOT-23

Very nice, I like the ON Semi parts, and the way they go out of their way to give you decent datasheets. Unlike Zetex, who often makes you do with one page.

Have you noticed the way these parts present you with Rds(on) values? Like 80 m-ohms for the NSS40200? I like that. Sure, trying to compare to a MOSFET, but hey, characterizing the high-current saturation-voltage upswing as a resistance is really the right way to go.

Doing a quick look right now, I couldn't even find a sot-23 MOSFET with Rds(on) = 0.08 ohms or less (they were bigger), let alone one going for $0.90 each, qty 100. Nice.

--
 Thanks,
    - Win

John Larkin wrote: : On 31 Jan 2010 16:10:06 -0800, Winfield Hill : wrote: : >

: > Even with damping resistors? E.g., EF outputs look : > like an inductor, due to the falling beta (assuming : > a low source impedance). L = Rs / 2pi fT, above : > the breakpoint frequency, f = fT r_e / Rs. That : > output inductance resonates with capacitive loads, : > but a damping resistor can keep Q = 1 max. : : like to oscillate with hot transistors, like the zillion-GHz NEC parts : for example. By the time you add enough base resistance to calm them : down, you may as well have used a slower transistor.

There is one place for those fast transistor, tranquilized by GHz damping circuits close to the transistor: when you need very low noise at intermediate frequencies. As an instance, the voltage noise of the SiGe BFP650 can be less than 1/4 nanovolt. An example spectrum is found in fig.4 of

(now out in AIP Conf. Proc. vol. 1185) . NESG3031 is another nice device in this respect.

Regards, Mikko

Cool.

Fig 4 is impressive, but "base grounded" implies that it will be useful only for very low source impedances. Transformers get interesting in such situations, or *lots* of opamps or jfets in parallel, which will be better at low frequencies.

John

=A0 =A0

=A0 =A0 =A0 =A0 SOT-23

=A0 =A0 =A0MMBT3906

=A0 =A0 =A0MMBT4403

=A0 =A0 =A0 =A0 BC807

=A0 =A0 =A0MMBT5087

=A0 =A0 =A0 =A0BC857C

=A0 =A0 =A0MMBTA64

=A0 =A0 =A0FMMT718

=A0 =A0 =A0MMBT5771

=A0 =A0 =A0MMBT5401

=A0 =A0 =A0MMBTA92

=A0 =A0 =A0 =A0 =A0

=A0 =A0 =A0 BFT93

=A0 =A0 =A0 =A0 =A0

The BFR92 and BFR92A are the complements of the PNP BFT92.

They all tend to oscillate, but putting a 33R "base-stopper" resistor in series with base (and close up against it) kills off most oscillations, without doing much damage to the high-frequency gain.

-- Bill Sloman, Nijmegen

I've never heard of that. I think gold and other alchemy is what keeps discretes fabs alive and kicking in ol' US and Europe.

I need to look, but I think NXP's 3904 and 3906 (complementary) are gold doped. The 3904 uses the same process as the BSR13, and I know that BSR13 has a gold doping step.

robert

Good call! NXP also have a similar line (called BISS), I don't know how they compare to ON's stuff.

robert

I'd also throw the BSS84 for P-channel into the hat. It's my jelly-bean part.

The BFT25 is NPN. Phil Hobbs should be able to give a good account on that one.

While we're at it, what other categories do you need hints for? For example, for drivers con mucho gusto the MIC4421/22 series is great but hardly any young engineer knows about stuff like that.

Oh, and tell readers not to blindly trust POR/BOR circuitry inside a uC ;-)

Can't hurt to mention some of the budget rockets, like the BFP620, but with the usual caution notes (like that it might be oscillating without even noticing it, until a government van shows up ...):

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

I think they all have low Vce sat BJT's now. For example Fairchild's got these;

FSB560 NPN - 2A; 60V; SOT-23

FSB660A PNP - 2A; 60V; SOT-23

I've never used those two but they are cheap a dime for the NPN at Newark; the PNP isn't stocked.

A quick search of Fairchild shows at least 6 low sat bjt's ranging from 1 amp to 3amps.

Another decent one from Onsemi I have used and have is the

MBT35200MT1 PNP - 2A; 35V; TSOP-6

I do prefer On semis datasheets to Zetex/Diodes, and stateing the equivalent Rdson is a good idea for quick comparisons.

John Larkin wrote in

Current noise is in the ballpark of 4.5 pA/rtHz, which makes it approximately to noise-match with a 50ohm source at an order-of 40K noise temperature. The frequency range extends at least up to

100 MHz but probably still much higher, even with the tranquilizers present.

I have had difficulties to wind really wideband transformers on low-loss cores in the past. The sporadic commercial transformers I've tried tended to have to high-enough losses to eat the noise temp.

True, done that in the past, with BF862's for instance. It just gets complicated, and in wideband designs you may create curious oscillation modes. In the sub-10kHz you can indeed do better with some other amps.

Regards, Mikko

[...]

^^^^^

Hey, Winfield, Mikko just brought another excellent one for the jelly-bean list. This one has got to be in there, very low pinch-off JFET, steep enough to start an oscillator from less voltage than a fuel cell provides. Very useful.

How could I have forgotten, oh, this will haunt me, the wrath of the old Philips engineers will come upon me ...

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Joerg wrote in news:7soub2Fc5mU1

Yes, but he asked for BJTs...

Want a working unit?

I have several each of CK722, CK760 and CK761 (if I can remember where I stashed them :-). My father was a Raytheon wholesaler from 1956 onward for quite a few years. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.