Lowest noise BJT transistor?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
My favourites are 2SC2545/2SC2546/2SC2547 and their PNP complements (2SA1085 or something, I don't remember the exact type number). SSM2210/SSM2220 (cheap version of MAT02 and its complement) are also pretty good. I don't know if any of them are still in production, or maybe in production with a different package and a different type number.

For extremely low voltage noise, the Philips/NXP BFW16A used to be a good choice because of its low base resistance, but it is out of production. You could try a more modern low-noise wideband transistor, such as the BFU760F (RB=3.32 ohm according to its Spice model). It will be difficult to keep it stable due to its high transition frequency.
 
2SA 872A (Hitachi)!.....this is the pnp version. Noise is specified between 2-5 dB maximum! I'm no expert, but perhaps someone else could through some light on this.

According to the last graph in the datasheet, the optimal noise figure is 2 dB related to 190 ohm at 1 mA collector current, 1 kHz. The noise contribution of the transistor is then equivalent to (10^(2/10)-1)*190 ohm ~= 111.1297 ohm. About 13 ohm of this is collector shot noise translated back to the input (half of kT/(qIc)). As 1 mA is the optimum according to the graph, the contribution of base shot noise must be roughly the same. That leaves some 85 ohm for base resistance thermal noise.

85 ohm is not bad at all for a discrete low-frequency transistor, but not exceptionally good either.
 
Member
Joined 2007
Paid Member
Hi Marcel,

Interesting thread ... would you happen to have advices on current production low-noise complementary BJT's? I searched for the 2SC2545 on Mouser's pages but it doesn't appear to be in their catalog ... In my context can be low voltage/current/high fT as well.

Best regards,

Jesper

P.S.: is it correctly understood that the BFU760F only accepts 2.8 VDC between collector and emitter?
 
Last edited:
Hi Marcel,

Interesting thread ... would you happen to have advices on current production low-noise complementary BJT's? I searched for the 2SC2545 on Mouser's pages but it doesn't appear to be in their catalog ... In my context can be low voltage/current/high fT as well.

Best regards,

Jesper

P.S.: is it correctly understood that the BFU760F only accepts 2.8 VDC between collector and emitter?

Hi Jesper,

Maybe there is an SMD version of the good old 2SC2545 in production somewhere? I don't know. They were always a bit hard to get.

About the BFU760F: yes, definitely if the circuit cannot handle the negative base current you get at larger collector-emitter voltages. If it can, you may be able to go higher, between 2.8 V and 10 V, although the datasheet is not very clear about that. I would try to stay below 2.8 V, just to be safe. The low voltage handling is a consequence of the small device sizes and thin base regions they use to make the fT very high.

Best regards,
Marcel
 
Member
Joined 2007
Paid Member
Hi - thanks Marcel & Dale for replying.

@Dale: Hmmm... Thanks for the tip on the KSC1845 & 3324. I actually have some of the KSC1845s so maybe I should give them a listen ...

@Marcel: Can I ask you if there are there any "side-effects" - audio-wise - with these transistors? I mean: Do the design parameters in a transistor like the BFU760F mean that they may not be the most suitable for audio design? Here I'm not thinking about their potential for oscillating but would there be something else that is less feasible - besides the size :-(

Winter greetings from Denmark - cold it is ...

Jesper

P.S.: Oups - forgot this: Do you know of any PNP transistor that compares with the BFU760F? Maybe a silicon-germanium type?
 
Last edited:
@Marcel: Can I ask you if there are there any "side-effects" - audio-wise - with these transistors? I mean: Do the design parameters in a transistor like the BFU760F mean that they may not be the most suitable for audio design? Here I'm not thinking about their potential for oscillating but would there be something else that is less feasible - besides the size :-(

P.S.: Oups - forgot this: Do you know of any PNP transistor that compares with the BFU760F? Maybe a silicon-germanium type?

No, I don't know a complementary version.

About disadvantages of an RF transistor like the BFU760F:


1. Parasitic oscillations. You will very probably need a base stopper right next to the transistor, and for absolute minimum audio frequency noise that will need to be a lossy ferrite bead rather than a resistor. For SMD boards you have special SMD ferrites, like the Murata BLM18EG series. A small resistor in the collector connection may also help. If the transistor oscillates, you will only be able to measure that directly with a very expensive RF spectrum analyser. If you haven't got one, check for excessive distortion and for bias points that change when you hold your finger or a small screwdriver close to the transistor.

2. Vulnerability to ESD and electrical overstress. Bipolar transistors are sensitive to emitter-base junction breakdown, it can cause beta degradation at low collector currents and increased 1/f noise. RF devices tend to be more sensitive than low-frequency devices; the BFU760F has an absolute maximum emitter-base voltage of only 1 V. Make sure you include an antiparallel diode to protect the emitter-base junction against reverse breakdown.

3. Uncharacterised 1/f noise. There is nothing in the datasheet or in the Spice model about 1/f noise. Chances are that 1/f noise is quite low in such a modern transistor, but it would have been nice if they had included measured data somewhere, now you will have to measure it yourself. In bipolar transistors, 1/f noise is mainly current noise, so it doesn't have much effect if you have an extremely low source impedance (like an MC cartridge).


Besides, some RF transistors have a low hFE, but this does not apply to the BFU760F. Low hFE means large base currents and, hence, large base current shot noise at a given collector current. This again doesn't matter much if you have an extremely low source impedance. Anyway, for the BFU760F, hFE is typically 340 at 10 mA collector current, which is not bad at all.
 
Last edited:
Member
Joined 2007
Paid Member
Hi Marcel ... thanks again for replying to my questions. What I'm considering using these transistors for is a wide-band analog stage prior to an A/D converter. I have no experience with how such devices sound but if they sound well their low voltage capacity may be suitable for this application - I guess.

I've experimented with listening to the 2SC2240 and to my ears it sounds at its best with a Vce of about 15 - 20 volts which would mean that to avoid destroying the A/D converter I'd have to add some kind of protection circuitry - in case there's an unforeseen high level signal on the input. That may not be necessary if the transistor itself sounds well at low voltages and the circuitry as such can be designed around low voltage rails.

So I've ordered a couple of the BUF760s and PNP versions from NXP and will try them out. Thanks also for the tip on ferrite beads & a resistor to - hopefully - avoid oscillation. I will keep this in mind.

Merry Christmas to you all ;-)

Jesper
 
BC550 is also a quite good transistor regarding to noise. NF1 is typically around 0.6dB, but then one should remember: There is a lot of mistakes to do wich has greater influence to the noise a transistor emits, wich makes the total design around it a lot more to take in consideration than the actual Noisefigure for the transistor. Just try to drive a LowNoise transistor just a tad outside the "currentwindow" where the transistor is at its most optimum regarding to low noise.

But have you considered J-FETs in this stage of the A/D-converter?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.