measuring mic pre noise

At this extreme, low base resistance is key.

The 8 Amp part has small base resistance so that it can be hammered to 8A Ic 1.6A Ib (overdriven switch) without huge base voltage.

The BF459 is a video amp. Bandwidth may be limited by Ccb against Rb. Low Rb is a virtue.

> most low noise parts I am familiar with are much higher current gain

The lower 3dB NF resistance is set by roughly Re+Rb. The upper 3dB NF is about hFE higher resistance. The OSR is in the middle. High hFE parts have a wide range of favorable source resistance, reach a very low NF, and may be marketed widely. Low hFE parts have narrower range of optimum source and won't reach really-low NF.

BF459 hFE at 4mA is like 30 to 50. It will give a 3dB NF at 10 ohms and at around 400 ohms. OSR is maybe 60 ohms, NF here is about 1.4dB.

> most low noise parts

2N5088 specsheet for 100uA shows 3dB NF 1K5-40K, 1dB at 10K; for 30uA shows 2K-1Meg(?) and 0.75dB at 15K. The high hFE part is wider between 3dB NF points and gets closer to zero NF. However I suspect the optimization for high hFE conflicts with low Rb.

Another aspect: both the cited parts have high breakdown voltage. Doping for high breakdown conflicts with doping for high hFE (roughly as square-root of voltage- see A.S. Grove). In practical circuits with small external resistance, low Rb is better than high Rb. holds the leaky junction together longer.

Interesting.. That's why I didn't post anything when I first saw it. An 8A part with a beta of only 25 must handle hundreds of mA of base current so clearly merit in low Rbb.  I suspect there is some benefit from modern process technology being cleaner than back in the day. It was an old trick to search out medium power devices for low base resistance, but they had to be screened for process noise.

I still have a small stash of 2sb737s in my back lab in case of emergency (.35 nV/rt Hz and circa 500 hfe).. but the BD459 looks like an interesting exercise in MacGuyver design, if you've got some lying around. 

JR

Neither of the parts cited is VERY readily available.

But it confirms that such devices are worth trying for sources under a few hundred ohms.

I've always had the feeling that big junctions *could* work at low noise impedance. I think reliability/yield issues have eliminated dirty-part excess noise. I had "no" hiss from TIP120 devices after electret capsules in an unsilent rehearsal hall (meaning a 1989 TIP120 at 28mA is as quiet as the FET in a 99-cent capsule).