[sdiy] twilight zone question

[Mike I]

Jason,
On the TL072 datasheet  look at the "input common mode range" spec... on the
OP275 datasheet look at the "input voltage  range" spec....  these  both
cover a narrower range of voltages than the full +15 to -15 range between
the supply rails. Running on  reduced supply voltages like +15, 0V (to the
op amp it will look like +/-7.5V depending on how you set it up) will make
performance even worse.  The op amp can go into phase inversion if the
common mode range is exceeded  on the inputs... if it is a really poorly
behaved op amp it can also latch up in this state, and the only way to
restore normal operation is to turn off the power. From your latest post it
looks like phase inversion is happening. The OP275 is actually worse than
the TL072, assuming the spec sheets contain reliable info.

This may or may not be the cause of the problem in your case, but I'd try
replacing the TL072 with LM358...the LM358 is cheap, but good for jellybean
single-supply stuff like this.
M.


----- Original Message -----
From: Jason Proctor <jason at redfish.net>
To: Mike I <mirwin at marketbridge.ca>
Cc: <synth-diy at dropmix.xs4all.nl>
Sent: Monday, October 16, 2006 10:15 PM
Subject: Re: [sdiy] twilight zone question


> i'm starting to think that mike is right with his mad tl072 theory.
> consider the following -
>
> + input = 5v
> - input = 7.5v
> output = 0v (correct)
>
> + input = 0v
> - input = 7.5v
> output = 15v (????)
>
> seems like the 072 spazzes if the input is < 1v. but i'm not
> understanding why i'm not seeing this everywhere. it's only when the
> 0v comes off the protoboard, and i know the two grounds are fine.
> weeeird.
>
>
>
> At 11:16 PM -0700 10/16/06, Mike I wrote:
> >Hi Jason,
> >About an hour ago I looked at Ken's site and the schemo for his clock
> >divider module and some other divider module - both circuits use LM358 in
> >the same way at the inputs..... I forget which CGS circuit you have, but
> >either one should be able to be connected directly to the 0 to 5V outputs
of
> >the PIC processor (caveat: I don't know anything about this Blacet
module,
> >and I assume it puts out TTL-compatible logic levels). There should be no
> >need for an intervening 0 to +15V level translation, the LM358 in the
> >existing CGS board will do a 0/+5 to 0/+15 translation directly. I would
> >advise against replacing the LM358 in the CGS board with TL072 though. If
> >you look at the TL072 specsheets, the inputs should not be brought any
> >closer than 3 to 4 volts of the most negative supply rail  (in your case,
> >ground). The LM358 inputs can be brought right down to ground and it will
> >still behave properly.
> >Good luck,
> >Mike