If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

# DRV425: Noise analysis

Part Number: DRV425

Hi, I am wondering how to fully evaluate the DRV425 for noise. There are some noise densities listed, e.g., 1.5nT/root Hz for the sensor and 170nV/root Hz for the output voltage. There are also charts. My goal is to understand what the noise output of the DRV425 will be with respect to the sensed input magnetic field so that I can predict what the SNR is. I plan to feed a precision A/D with the DRV425 and I want to make sure I don't over specify the A/D. If it matters any, this is for a low sample rate project (<60Hz) and I am looking to measure small (less than the earths) magnetic fields to as high a resolution as possible. Thanks.

• Hello Mark,

Please take a look at this post.  Here I get the noise density plot and calculate the RMS expected output.  If you filter to the 60Hz by either filters or averaging on the ADC I calculated the noise will be about 16.6nT RMS or 109nT peak to peak.  I have attached the noise plot so you can view for the bandwidth your measurement will be.

Worksheet in DRV425 Noise Analysis.xlsx

• Hi Javier, this mostly answered my question, but I wish to ask some further clarifying questions. From what you show, the 16.6nT RMS comes from assuming a noise bandwidth of 1000Hz. I think I get that. But what is the purpose of the graph (Figure 47) in the data sheet that shows the shunt sense amplifier output voltage noise density vs noise frequency? Where or when would this come into play in an analysis?

Also, in your previous response in 2017 you mention you have a TINA simulation of the DRV425. Is this available to the public? Thanks again.

• Hello Mark,

I think I looked into this before and saw the flux-gate noise was much larger.  You can do the sum of squares to get an estimate.  Please calculate for you bandwidth.  Below is an example and you can see why I assumed I could ignore the noise from the amplifier.  I also ignored the 1pole filter multiplier.  Depending on your bandwidth filter you will need to adjust but for a rough estimate I do not believe you need this.

For output noise of the differential amplifier and assuming a 1 pole filter the calculation is RMS = 170nV *√Hz*1.57.   You will need to multiply by 6.6 to get peak to peak. To see how I got the calculation see our Analog Engineer Pocket Reference page 60, 61.  Or in more detail is this training from our Op-Amp team.

The Difference amplifier noise is output based and the Flux-gate noise will be the input.  To combine these you need to multiply the flux-gate noise by 12.2mA/mT*Rshunt*4V/V to calculate the noise at the output.

The noise I assumed of 16.6nT RMS is from the bandwidth of 60Hz.  This is captured from the spreadsheet I attached and it can be observed in the Tina simulation.  The excel sheet is a capture of the Tina Simulation.  Please excuse the typo.  Column B should be noise density.

At 60Hz assuming perfect filter and a 100Ω shunt the calculation would go as follows:

Flux-gate:

Noise @ 60Hz =16.6e-9T

16.6e-9*12.2nA/nT*100Ω*4V/V = 8.10e-5 V RMS

Diff Amp:

170nV√(60) = 1.32e-6 V RMS

Total

√{(8.1e-5)²+(1.32e-6)²} = 8.101e-5  RMS output noise

I have attached the noise simulation for the flux gate.  It is only to simulate the noise level.  I translated V-->T.

DRV425 Noise.TSC