[sdiy] Some OTA questions

[harrybissell]

Good evening young Skywalker   :^P

Jeff Farr wrote:

> I've been working my way through the EN app. notes on the 3080 and
> there are some nagging questions left, first, there is an example of a
> S&H circuit where a 3080 in switch mode is feeding a capacitor with a
> FET buffer.  If I understand this correctly Vdiff simply switches on
> the 3080 and since Iout = Ic it's Ic that is doing the "sampling" so I
> would connect my source (noise) to that.

Think of it this way... if Gm bias (current) is zero, the output is
effectively
open circuit (hold mode).  The inputs are the signal input, the Gm bias
is the sample clock.

> Do S&H's usually work with a
> current in?

might be in some cases...  If its a sample hold you want try the LF398
chip...

> Second, I haven't yet figured out how to make use of Id
> in a 3280, can someone point me to a good app note about that?

You should see my twin brother Jim Patchell for that... Basically Id will
allow you to use a higher level signal in... reducing distortion and
increasing
S/N ratio... at the expense of lower input impedance.

> What
> about the differences between a 3*80

3080 = plain vanilla OTA
3280 = DUAL OTA with linearizing diodes, probably the best OTA available


> and one of the LM1*00 OTA's?

LM13600 = Dual OTA with linearizing diodes and darlington buffers... the
bias
current is fed to the buffers, so that they are optimised for best
operation. Downside...
fast bias changes might THUMP the output buffer

LM13700 = same as above BUT the darlington buffers are generic and NOT
connected
to the bias input.  They work well over 'most' of the range but are not so
good at really
low bias currents.

NE5517 = same as LM13600 but there is a buffer between the bias input and
the
darlington buffer, that might make it work better in some apps than the
LM13600...
otoh I have not seen any marked improvement in my sample evaluations...
and found
some distortion I liked less than the 13600...

H^) harry   (this could go on and on of course...)