What quad opamp can replace LM324 for this application?

[cm64]

I have the circuit below where I convert 0-3.3V 12-bit DAC output of a microcontroller to bipolar +/-10V to move a sensitive device.

I use LM324 since it is quad, rail-to-rail ect. But for a better output resolution hence for lower noise at the outputs what IC can be replaced with LM324? Is there any less noisy, DIP package IC which can replace LM324? I'm not familiar with the audio and analog electronics. I would appreciate any suggestions. I need something down to 5mV max 10mV noise floor.

 

[barry]

The antique LM324 is NOT rail-to-rail.

Why don't you search on the manufacturers website? It's easy. Without knowing all your requirements, here's one (of many) suggestions: TLV9354.

 

[danadakk]

First of all a LM324 is not RRIO.

A once chip solution with factory trimmed for CM. One even with follow on LPF.

Selection Table for Fully Differential Amplifiers | Parametric Search | Analog Devices

Analog Devices' Selection Table for Fully Differential Amplifiers lets you add, remove, and configure parameters to display; compare parts and choose the best part for your design.

www.analog.com

https://www.ti.com/amplifier-circuit/fully-differential/overview.html

Regards, Dana.

 

[FvM]

No doubt there are better OPs than LM324, but do you really see noise floor higher than 5 mV produced by this circuit?

 

[Easy peasy]

MC34074 ....

 

[KlausST]

Hi,

I agree with the others.
You need just an Opamp that is
* unity gain stable
* meets your output voltage requirements
* meets your noise requirements
* eventual other circuit requirements (like offset, frequency range...)
--> many, many Opamps will do.

Some ideas:
* the circuit can be reduced to just use 2 Opamps
* you need to add power supply decoupling capacitors
* I recommend to add low pass filtering capacitor (s). Especially when you work with DAC to "smoothen" the steps. And also to reduce high frequency noise at all.

Noise:
If you see noise in the millivolts region, then most probably this is not caused by the Opamp. It rather is caused by bad wiring, bad measurement, bad PCB layout, missing capacitors...

Klaus

 

[danadakk]

What is your desired spec for noise floor and its BW ?

The LM324 datasheet poorly characterized, eg. there is no PSRR vs Freq for
example to give an idea what noise performance occurs from power supply
contribution. Especially if using other logic, clocks in system.

This can be done with one fully diff opamp as posted in #3 above. G matching
and CMR very good as well. An example of trying to design diff amp discrete stages
using discrete designs (production IA stages are laser trimmed during manufacture) -

As far as bypass considerations not all caps, for same C, have same ESR
performance (use a combination of Polymer and MLCC to get best results) -

The OPA862 also seems to be quite capable.


Regards, Dana.

 

[floppy32]

Hi,

I agree with the others.
You need just an Opamp that is
* unity gain stable
* meets your output voltage requirements
* meets your noise requirements
* eventual other circuit requirements (like offset, frequency range...)
--> many, many Opamps will do.

Some ideas:
* the circuit can be reduced to just use 2 Opamps
* you need to add power supply decoupling capacitors
* I recommend to add low pass filtering capacitor (s). Especially when you work with DAC to "smoothen" the steps. And also to reduce high frequency noise at all.

Noise:
If you see noise in the millivolts region, then most probably this is not caused by the Opamp. It rather is caused by bad wiring, bad measurement, bad PCB layout, missing capacitors...

Klaus

Hi,
The LP filters should be at the output or before the input?

 

[KlausST]

Hi.

To most effectively reduce (low voltage) noise I'd put the filter at the very most end of the signal path.
With this you attenuate the noise generated by all previous stages.

But if you want to suppress noise caused by switching, ESD, ... where the noise is caused by sharp pulses, then I'd put the filter as close as possible to this switching source.
With this you reduce dV/dt and you reduce the amplitude, so that (input stages of) Opamps don't cause additional errors (plop, DC, ringing) caused by overload.

Klaus