[sdiy] Prophet 10 op-amp swaps?

Dave Manley dlmanley at sonic.net
Wed Jul 31 23:10:42 CEST 2013

Hmmm, 'unsolicited agression' with a simple question.  Who knew.  Call the
UN.  Let me help clarify with an example: "The opamp may not be the only
thing having stability issues".  There.  That is an example of aggression.
 See the difference?  I accept you apology.

The reason why I asked about the 'GHz' bandwidth is because I didn't see
anything in the mentioned opamps that were anywhere near that and it
seemed like hyperbole.  I like to review datasheets before posting, and
did so.  I also missed Eric's mention of the LME49710.  Which does have a
much higher GBW than any of the other mentioned chips.  Not surprisingly,
Eric reports the problem only occurs with the LME49710 - the LF356 and
TL07x parts don't.

1. Damian's suggestion he found googling to add decoupling isn't bad, it
is worth a try.

2. The suggestion about measuring the DC output voltage of the opamp
should not be an issue.  The author was talking about opamps with high
input bias currents, leading to input offset voltage, thereby causing a DC
output, is not an issue with the LME49710 - it is measured in nA and it
also has low input offset voltage.  A review of the datasheet shows this.

3. Some of the other suggestions are nonsense if he had reviewed the
Prophet-10 schematics.  Do not ground your scope probe on the 'negative
supply of the opamp' - that will create a dead short to the -15V rail. He
adds a warning, but a little research and less >>>typing<<< on his part
would have eliminated this suggestion.

4. He suggests putting a cap in 'series with the feedback resistor'.  If
the opamp in question is the top opamp (U824) in the schematic, that opamp
doesn't have a feed back resistor.  The opamp is connected as a voltage
follower with the output directly connected to the negative input.  The
usual method of reducing the bandwidth is to put a small cap (10's of
picofarads) in parallel with the feedback resistor.

A question for Eric - are the opamps socketed?  Putting an opamp like the
49710 with a GBW of 55MHz (high but still ~1/20 of a GHz!!!) in a socket
is asking for issues with parasitics.  If socketed: as Damian, will warn
you - if you decide to remove the sockets take care not to damage the pcb

If you don't remove the sockets, then Brian's suggestion is probably best.
 Look for a newer 'audio' grade opamp with a lower GBW than the 49710's


More information about the Synth-diy mailing list