SET sound question

[chip_mk]

But the 2SK1815 isn't a bipolar transistor

Sorry, my bad. I had 2SC1815 on mind, in fact this particular chart.

 

[chip_mk]

The gm is Ic / Vt.

Sure, in ideal world where Rb = 0

 

[DF96]

But there is a really big difference between using a gain-consuming NFB loop and simple degeneration in terms of the impact on the signal.

Degeneration does not consume gain?

we all know what "no NFB" means if we suppress our pedantry.

It is not pedantry to ask whether "no NFB" actually means 'some NFB, albeit no GNFB''.

I find this thread interesting because some responses highlight the pervasive problem that if all you can "see" is THD then naturally you think every sonic characteristic root cause is discoverable if only you stare at the THD graph long enough.

I am not going to search through this thread to see who first mentioned THD, but it could be you. My experience is that it is most often mentioned by those who wish to ignore it, while implying that others (who have not mentioned it) are bound by it.

It's not quite so simple, people are all too focussed on talking about voltage gain and of course there is no active element that creates voltage gain - they are all transconductance devices.

A device is what you consider it to be. Use the model which best fits the application. A triode operated into a very high impedance is best seen as a voltage gain device, while a BJT operated from a high impedance source into a low impedance load is best seen as a current amplifier. Either can also be regarded as a transconductance device, given different conditions.

Whether you decide to control the charge by voltage or current is up to what suits your circuit.

Yes.

 

[DF96]

As far as I understand BJTs are different from mosfets in that the low impedance input was purposeful to create electron holes in between the doped areas which allow electrons to jump the gap.

No, the low input impedance of BJTs is a nuisance. It can be regarded as non-ideal behaviour, which is why the current gain is rather random from sample to sample and varies all over the place with collector current.

Intrinsically BJTs need to be current driven but I think from an effective standpoint they are just low input impedance, low vbe turn-on mosfets.

No, different effect so different behavour.

If there was a true ultra low input impedance current driven transconduction device it would revolutionize the industry, but unfortunately as far as I know voltage is required to practically make an electronic switch.

Common-base BJT driving a resistive load?

We must be living in different universes. So “gm is non-linear but at least it is consistent”; good luck with that.

But this is exactly how precision circuits are designed: gm is predictable. 'BJT is a current amplifier' is what we teach newbies who don't understand the exponential function.

 

[Bigun]

A triode operated into a very high impedance is best seen as a voltage gain device, while a BJT operated from a high impedance source into a low impedance load is best seen as a current amplifier. Either can also be regarded as a transconductance device, given different conditions.

Exactly my point, a device by itself is not a voltage amplifier - it needs a load, an impedance is needed to convert the devices intrinsic output current signal to a voltage.

 

[hellokitty123]

Common-base BJT driving a resistive load?

Only useful in limited circumstances. I meant an electronic switch that functions entirely in the current realm.

 

[Bigun]

Not at all, the common base is actually quite common and useful. It is used as the top device in a cascode and folded cascode; it can be used as a VAS device too TGM4 amplifier

 

[hellokitty123]

Yes, those are by definition, limited circumstances.
My point was if a true current input transconduction device existed with an ultra low impedance on both the gate and the source then circuits could function almost entirely in the current realm. Parasitics would no longer matter. It would be the ideal electronic switch. Except trying to troubleshoot such a circuit would be tricky. I'm no transistor scientist but I'd guess a true current amplifier can't be made.