Op-Amp selection table.

[Rixen]

I've been working on a table that compares op-amps across most of the parameters of interest to pedal builders, biased somewhat toward battery power considerations.

Here's a link to a google docs spreadsheet, which I am sporadically updating (and looking for a few suggestions for additions).

https://docs.google.com/spreadsheets/d/1VjHWtAaqLJBZi7CKDw5SlzluXgqv__cc42lV0K1jMvQ/edit?usp=sharing

some notes: TL07x has been used as a 'benchmark' and where a parameter has "worse" performance that cell has been shaded red.

Vs min and Vs max are single supply figures

Supply current at 9V has often been taken of the chart, as the datasheet table figures are usually measured at a different value

Single, Dual and Quad variations- sometimes a single, dual and quad are released as a different parts so it may not be obvious that they are the same family

Voltage noise- self explanatory

Current noise- this matters when a high resistance is present at the input, for example, a TL071 in non-inverting configuration with input terminated into a 1M resistor will have 0.01pA x 1000000 = 10nV√Hz @1kHz equivalent noise in addition to the voltage noise. For some devices this can be significant and can make them a poor choice for an input stage.

Slew rate- ability of the output to change rapidly. Low levels of this parameter can cause slew rate limiting distortion.

BW- unity gain bandwidth.

R-R output- if the output can be considered rail to rail, can give a little more headroom for the signal

Input offset current and bias current- these are a factor when inputs are terminated into high resistances, as they will generate a dc voltage that can saturate an amplifier at even moderate gain.

CMOS- cmos devices have lower electrostatic damage threshold than others, so in my opinion they are not a great choice as an input stage. My Red Llama is still working fine though...

Crossover distortion -device should be avoided in the signal path

DIP- package friendly for breadboard and stripboard.

[samhay]

Nice, but you should consider adding the NE5532/4.
The rail-to-rail column could be expanded to include single supply chips like the LM358.
Would be good to note the input type as well - BJT, JFET, CMOS.

[Kipper4]

Thanks Rixen

[Passaloutre]

Good stuffs! I just used the MC33174 in a circuit so I was curious to see how it stacked up. Here are the figures in order if you want to add to the table:

MC3317x,3,44,0.18,SDQ,32,0.2,2.1,1.8,almost,40000,200000,N,N,Y

Feel free to double check my numbers (https://www.onsemi.com/pub/Collateral/MC33171-D.PDF)

[Rixen]

Added NE5532/4 and MC3317
changed 'CMOS' column to reflect all input types
considering adding single supply operation column - this indicates that the input voltage range includes ground (negative rail)

[R.G.]

You're missing a really important column or two. You need to include the input common and differential mode voltage ranges, and whether the output inverts when one of these is exceeded.

The TL07x series is a good "bad example". It's input ranges only go to about 1.5-2V from both supplies, and its output DOES invert when the range is exceeded. About all it doesn't do bad is latching up its output when the input ranges are exceeded. These issues have bewildered people on this forum and others.

[Rixen]

Thanks R.G, I'll add those columns.

[GGBB]

Where are you getting the "Supply current mA typ, per amplifier @9V" values? Are you measuring/calculating these yourself?

[Rixen]

Where are you getting the "Supply current mA typ, per amplifier @9V" values? Are you measuring/calculating these yourself?

they are taken from the supply current/supply voltage graphs in the data sheets. Typical value.

[GGBB]

Where are you getting the "Supply current mA typ, per amplifier @9V" values? Are you measuring/calculating these yourself?

they are taken from the supply current/supply voltage graphs in the data sheets. Typical value.

The graphs - of course - thanks.

[Rixen]

You're missing a really important column or two. You need to include the input common and differential mode voltage ranges, and whether the output inverts when one of these is exceeded.

The TL07x series is a good "bad example". It's input ranges only go to about 1.5-2V from both supplies, and its output DOES invert when the range is exceeded. About all it doesn't do bad is latching up its output when the input ranges are exceeded. These issues have bewildered people on this forum and others.

Interestingly for the TL07x, the datasheet blurb proclaims the input common mode goes up to the positive rail, but in the characteristics table we see this is only for a 'typical' device and can be as bad as 4 V from the supply rails for a 'minimum' device @ +/-15V supply

[Rixen]

added:

OPAx227 : A very low voltage noise device, but watch out for the current noise into any high impedance

TSX711: Rail to rail input and output, input actually extends slightly past both rails and operation down to 2.7 V could make it useful for a lithium coin cell powered device

OPAx132: not as 'good' as the OPA164x, and more expensive from my vendor. Included for completeness.