Low-noise BC109?

[Synchrodyne]

My take from H.P. Walker’s article on low-noise discrete transistor amplifiers, Wireless World 1972 May p.23ff, is that for the magnetic cartridge RIAA-equalized case, the circuit configuration determines the theoretically best possible signal-to-noise (SN) ratio, 72 dB against a 2 mV input for the series feedback case. Thereafter circuit details, particularly first stage collector current, determine the closest available approach to the theoretical SN with actual transistors. HPW used a BC109C first stage in his design for an actual SN of 70 dB, so it does appear that there is very little more to be had by using lower noise transistors, and furthermore it would require transistor type-specific circuit optimization to obtain whatever increment might be available. For example the article showed 70.5 dB for a TIS97 at optimum collector current, with 70 dB available over a reasonably wide range of collector current.


Cheers,

[Argus25]

As I had mentioned in post #17, the collector current and the circuit configuration are important. However, regardless of the particular transistor's spec sheet and quoted noise figure, I found quite a lot of variation in the audio spectrum noise quality produced not only for transistors of different types, but those of the same type number from different manufacturers and batches.

So, to find the better transistor specimens it pays to actually build the exact circuit you intend to use, or as near to it as possible, have the transistor socketed, and select the better one/s with a listening test after a trial of a good number of devices. Some have very erratic LF noise, which is audible and not predicted by the theoretical noise spectrum they should have. Plus, in the experiment, you can vary the collector current too.

[AC/HL]

A handy cross reference of the BC108 family and derivatives. Bear in mind that the original design is over 50 years old now. Things move on.