OPA277 noise 0.1-10Hz?

Mattias,

You are correct that there is an inconsistency in the OPA277 data sheet.  I also think the problem is that the frequency axis of input voltage noise graph is mislabeled.  Art Kay looked into this issue some time ago and that was his conclusion (please see the attached presentation).

Marek,

Ok, many thanks for swift response!

Best Regards

/Mattias

M Haglund said:

Hi!

I'm having a 10Hz system using the the OPA277 amp.

I the ds the voltage noise is specifyed as:

Input Voltage Noise, f = 0.1 to 10Hz 0.22uV PP and 0.035u Vrms

If i look at the "INPUT NOISE AND CURRENT NOISE SPECTRAL DENSITY vs FREQUENCY" the noise up to 10Hz is dominated by the 1/f and the graph shows the noise density @1Hz ~ 50nV

I'm for 1/f noise learned to  use the equation en_1/f = (e_1/f@1Hz)(√[ln(f_H/f_L)])   use f_H = 10Hz; f L= 0.1Hz

en_1/f  = (50nV)(√[ln(10Hz/0.1Hz)]) = 107nVrms

....which is much larger than the 0.035nVrms given in the ds? 

Where do i go wrong?

Many thanks

/Mattias

 

Phill

Hi Phil,

Thanks for asking, but sorry, i asked this over a year ago and cannot remember/find the customer info anymore? :-)

Phill,

The peak-to-peak noise measured in a 0.1 to 10Hz range is commonly used as a measure of the low frequency noise encountered in slow measurement systems. It is a specification seen in many data sheets for precision op amps and instrumentation amplifiers and provides a way to compare the low frequency noise of various amplifiers.

Regards, Bruce.

Mattias,

There was an error in the original PDS noise graph where the horizontal axis were mislabeled by a decade (1Hz to 10kHz instead of 0.1Hz to 1kHz) - see below. There is an errata sheet next to PDS (see attached) pointing the error.  The final result is that the actual voltage noise spectral density should have been 50nV/rt-Hz at 0.1Hz instead of 1Hz as originally shown in PDS.  Taking this into account, you may calculate the noise for 1Hz using following equation: