EMRFD Message Archive 10081

Message Date From Subject
10081 2014-05-30 17:31:03 Glen Leinweber Low-noise audio amps
Have been measuring noise figures of devices that might be used
in direct-conversion receivers as post-mixer audio amplifiers.
When Googling for "low-noise transistors", 2N5087 & 2N5088 & BC549
show up. When faced with low-Z sources like 50 ohms, noise figure
for these devices falls well above 5 dB. SSM-2220 is about the best
low voltage noise bipolar transistor I've found, but $$
Ultra-low-noise op-amps have voltage noise in the 1nV/rt(Hz) ballpark.
Noise figure can't be better than 3 dB, only worse.

Have come to believe that a discrete transistor beats the pants off
any op-amp, so have been testing PNP transistors for devices yielding
very low Voltage noise (current noise is less important when source
resistance is 50 ohms).
It seems that some rather ordinary power transistors come out nicely:
MPSW51A
NTE129P
2N4033
Oddly, data sheets for these devices don't even bother mentioning that
voltage noise is far lower than so-called "low-noise" transistors or ANY
op-amp.
Like EMRFD warns: don't trust lore....MEASURE
10082 2014-05-30 19:41:01 Dave Re: Low-noise audio amps
Glen:
  Try a 2N4401.

Dave - WB6DHW


10083 2014-05-31 05:11:46 David Re: Low-noise audio amps
The 2N5087 and similar transistors are optimized for low noise (and high hfe) at
low currents where voltage noise will be higher anyway. They need to be run at
high currents with an associated high input bias current for optimum noise with
a low source impedance but I believe they are limited by relatively high base
spreading resistance.

I agree with Dave about testing the 2N4401 but what I think would be more
interesting are the various "super" transistors pioneered by Zetex (now Diodes
Incorporated). Fairchild and On Semiconductor now have similar lines of
transistors intended for high efficiency high current density switching
applications. Most of these transistors are going to be NPN instead of PNP
though.

Zetex wrote a short application note discussing what you have run across:

http://diodes.com/_files/design_note_pdfs/zetex/dn11.pdf

Besides audio driver transistors like the BC327/BC328 (PNP) and BC337/BC338
(NPN), I would also consider large area JFETs like the LSK389. They have the
advantage of low bias current even at high drain currents.

On Fri, 30 May 2014 20:27:26 -0400, you wrote:

>Have been measuring noise figures of devices that might be used
>in direct-conversion receivers as post-mixer audio amplifiers.
>When Googling for "low-noise transistors", 2N5087 & 2N5088 & BC549
>show up. When faced with low-Z sources like 50 ohms, noise figure
>for these devices falls well above 5 dB. SSM-2220 is about the best
>low voltage noise bipolar transistor I've found, but $$
>Ultra-low-noise op-amps have voltage noise in the 1nV/rt(Hz) ballpark.
>Noise figure can't be better than 3 dB, only worse.
>
>Have come to believe that a discrete transistor beats the pants off
>any op-amp, so have been testing PNP transistors for devices yielding
>very low Voltage noise (current noise is less important when source
>resistance is 50 ohms).
> It seems that some rather ordinary power transistors come out nicely:
>MPSW51A
>NTE129P
>2N4033
>Oddly, data sheets for these devices don't even bother mentioning that
>voltage noise is far lower than so-called "low-noise" transistors or ANY
>op-amp.
>Like EMRFD warns: don't trust lore....MEASURE
10084 2014-05-31 05:45:19 Dan Mills Re: Low-noise audio amps
Large geometry medium power transistors actually have a long history as microphone amp input stages, see for example some of the old neve designs.

The win is in the low intrinsic emitter resistance at least at reasonable current levels.

The other thing you sometimes see is multiple parallel devices for much the same reason.

I have wondered about the possibility of a common base stage in this position (You basically never see these in audio applications).

73 Dan M0HCN


10085 2014-05-31 06:10:31 John Levreault Re: Low-noise audio amps
10088 2014-05-31 15:26:38 bob_ledoux Re: Low-noise audio amps
Just in case you haven't seen it, look at the article here:

Download

 

10092 2014-06-02 10:48:24 Glen Leinweber Re: Low-noise audio amps
Dave,
    2N4401 was the first one measured. Those PNP's stated
measured lower voltage noise with similar collector current (2mA):
MPSW51A, NTE129P, 2N4033.
A MPS2907A and 2N4403 measured lower voltage noise than 2N4401.

I wonder why PNP's seem to have lower voltage noise than NPNs?
Noise of bulk base resistance (rbb) seems critical for low voltage noise.
And I wonder how much difference the manufacturer makes, especially
when so many cloned devices are around?

10093 2014-06-02 12:46:54 arfghans Re: Low-noise audio amps

Here's a novel low-noise circuit that is quite educational, by the late great W. Marshall Leach. Uses 2N4401/2N4403, always a great choice for low-noise audio.
Moving Coil Cartridge Head Amps

 

I used a derivative of this with a step-up transformer to achieve an equivalent input noise figure equal to a 2.7 ohm resistor. Scroll to the bottom of this page under Ultra Low Noise Preamp.
Electronics

 

-Gary, WB9JPS