[sdiy] Operational Transconductance Amplifiers

[Magnus Danielson]

Antti Huovilainen wrote:
> On Tue, 9 Mar 2010, Antti Huovilainen wrote:
> 
>> The response is due to the differential pair and is fundamental for 
>> BJT transistors. It can be somewhat linearized by predistorting the 
>> input and perhaps some other tricks.
> 
> The linearization diodes in LM13700 seem to use similar trick:
> If the input is fed with differential current, the differential voltage 
> over the diodes will be such that
> Id_diff = Id*tanh(Vd/2Vt)
> ->
> tanh(Vd/2Vt) = Id_diff/Id
> Vd = 2Vt*invtanh(Id_diff/Id)

This is by the way common to linearizing Gilbert cells for instance. It 
should not come as a big supprise as the differential pair as such is 
what have this property in both cases.

Cheers,
Magnus