From: Winfield on
D from BC wrote:
> acannell wrote:
>
>> Is there a clever way to make a current amplifier which acts like a
>> NPN transistor except that it has 0 Vbe? That is, with a "base"
>> resistor of 10k and an input voltage of 100mV, 10uA gets amplified
>> into 1mA at the "collector"? Preferably a method that does not use
>> an IC's (opamps, etc..)
>
> The OPA861 comes to mind but it's an 8 pin IC..
> Clipped from datasheet...
> "The OTA or voltage-controlled current source can be viewed
> as an ideal transistor. Like a transistor, it has three terminals
> - a high impedance input (base), a low-impedance input/output
> (emitter), and the current output (collector). The OPA861,
> however, is self-biased and bipolar."

I like the opa861idbv with its cute SOT23-6 package.

TI also offers the Burr-Brown SO-8 package opa860, which
has the same OTA stage, but adds an output buffer.

These parts replace the opa660, which was discontinued.
Read the opa660 datasheet to see the schematic.

While TI calls these operational transconductance
amplifiers (OTAs), and they characterize them as ideal
active transistors, they're actually basically current-mode
opamps with the transconductance node exposed and the
Gv=1 output stage either omitted or made available with a
separate connection pin, as in the opa860 (and opa660).
This type of OTA functionality can be had in several Analog
Devices opamps, like the ad846 and ad844, which bring
out the transconductance pin, labeling it "compensation".

TI's ability to set the quiescent current of the input transistors
is a nice touch, because you can reduce the +input current
(what they call the base) and reduce the output-current offset,
at the expense of speed and a reduced the maximum output
current capability (note, the maximum output current can
substantially exceed the quiescent current).