Sub oscillator for CEM3394

[Martin Czech]

> Duane R Balvage wrote:
> > Then I thought, "save a step (and parts) and make a comparator out of
> > the op amp." This worked... a nice square wave
> 
> I'm a bit sceptical about using TL07* or TL08* as comparators. It often
> works, but sometimes they latch up. Seems to depend on manufacturer,
> lot, maybe supply voltage or whatever. Another case where one should be
> careful when replacing 741s, BTW.
> Some comparator ICs have also the advantage that they can produce TTL
> output levels although being operated from +/-15 volts.
> 

one observation:
I have a little parallel port expansion for Atari STE, a 8 bit
adc/dac.  It's called "Galactic Sound Sampler". It uses TL084. If the
microphone input recieves to much voltage (switching or line level),
the adc locks, ie. allways 255 comming out regardless what you do.
This won't even go away if the power is switched off.  It helps to swap
the TL084.

You may think, well + and - are high impedance inputs. No current can
flow (allmost no).

But: These are junction FETs. Ie. there is a breakdown voltage, where the
input acts like a breakdown diode, and in revers polarity there will be
normal forward diode current, if you exceed the potential of the jfet
by 0.6V. However, doing this via transients (like in my example above)
or even continuous will not even disturb the input
functionality/destroy the input junction, but it will also cause
substrat currents to flow in the opamp chip, and if you go too far the
whole chip will trigger as a scr: latchup, burnout. I think there is a National
app. note where series resistors and clamping diodes where proposed in
order to avoid such damage/malfunction.

This could well be the reason why a TL081 does not work as replacement for a
741. The input breakdown is different, and the whole chip layout and process
is different, so no wonder that some critical situations come out different.

This is just my $0.02.

m.c.