[sdiy] Really obscure OTA question

[jhaible]

If you look at the BE junctions of the input transistor pair,
asuming there is no tail current (Iabc = 0)
these are two diodes connected back to back in series,
so no current will flow across them. But at some voltage,
one of the BE junction will conduct (zener diode ...),
and this is what you want to avoid. I guess the "5V" is
just a lower limit for the zener voltage - it may happen
at 5.5V, or it may happen at 7V.
*If* it happens, the amount of damage will depend on
the current (external resistors). I guess even a rather
small current will degrade the transistor (noise, balance
between the pair ...).

JH.

-----Ursprüngliche Nachricht-----
Von: harrybissell <harrybissell at prodigy.net>
An: <synth-diy at dropmix.xs4all.nl>
Gesendet: Donnerstag, 30. Januar 2003 06:56
Betreff: [sdiy] Really obscure OTA question


> While studying a circuit (the PV-1 based on the design by Bob
> Moog...used
> with permission  ;^)... I noticed that his OTA current source had one
> input
> at ground, and the other at +6V (divider of two 22K resistors)...
>
> now we all know the differential voltage limit is 5V for this part.
>
> Question 1:  What would be the possible benefit to driving the input
> this hard ?
>
> Question 2:  If you see the CA3094 OTA with buffer output... the spec
> sheet says
> differential input voltage 5V... but no device damage will occur if
> input current is limited
> to 1mA or less...
>
> no such disclaimer on the CA3080...  otoh I have used these in higher
> differential voltage
> apps without damage.
>
> Observation:  I tried a PV-1 running on +/- 15V supplies (now that
> voltage is 7.5V)...
> with a NATIONAL LM13600, or 13700... circuit operation was normal. With
> the
> NJR 13600, or 13700... the OTA would not shut off when the bias current
> was reduced
> to zero... output current continued to flow.
>
> I found it was very sensitive to that higher input voltage. Making the
> input 7V or less made it
> resume operation...
>
> Now what's with THAT do you think ???
>
> Theories, please...
>
> H^) harry
>
>