Replacing RC4558 opamp with NJM2068DD question

[bamavette]

I have replaced quite a few 4558 opamps with NE5532 and always put a 100nf cap across the rails when doing so. Never had an issue.

When replacing a 4558 with a 2068 is the 100nf cap still needed? I think so but wanted to ask the experts. Yeah I know it is easy to just put it in and it won't hurt anything. I am just curious more than anything.

Thanks.

 

[rayma]

Some kind of local decoupling is always needed for op amps due to the wide bandwidth. The across the rail style
may work best for you if the board layout is less than optimum (no ground plane, etc.).

 

[indianajo]

I've modified several products with 741s and 4558s that did not need local decoupling. 1 v/us slew rate. I have found that replacing with anything 5 v/us or faster slew rate does.

 

[rayma]

TI says for the LM741:
For proper operation, the power supplies must be properly decoupled. For decoupling the supply lines,
a 0.1-µF capacitor is recommended and should be placed as close as possible to the LM741 power supply pins.

And for the RC4558:
Place 0.1μF bypass capacitors close to the power-supply pins to reduce errors coupling in from noisy or high impedance power supplies.

 

[WhiteDragon]

Sounds about right. Decoupling 100n caps pretty much mandatory for any opamp.
When you start getting over 100 MHz many will use 10n and the datasheet will usually specify.
At greater distances allowed usually combined with 10uF as well for each rail for any opamp

This case good old 100n close to the pins, the original design should just have them already

 

[indianajo]

Sounds about right. Decoupling 100n caps pretty much mandatory for any opamp.

I have repaired several products with 4558 and 741 that did not have them. Problems were not the op amps oscillating.