I am looking into using an Allegro A1324LUA-T for sensing very small magnetic field changes.

The data sheet shows that the sensor gives a 5mV/G change in voltage so I can sense a 0.1 Gauss change if I can capture voltage change of 0.5 mV.

Does anyone have any experience measuring such a small change in voltage on a rather large output signal? The sensor outputs 2.5 V when no field is present. I am wondering if there is any equipment that can pick up 0.1 mV change with good accuracy.

  • 1
    \$\begingroup\$ 10 microtesla - that’s less than the geomagnetic field in my area (70uT). Probably better off with a GMR sensor \$\endgroup\$
    – Bryan
    Jun 17, 2023 at 5:57
  • \$\begingroup\$ Post a link to a datasheet. Is the output absolute or ratiometric, and what is the tempco?If ratiometric, a Wheatstone bridge will help. Are you looking for long term stability, or can you switch your target field on and off? It might be a case of buying one, hooking it up to your ADC for a few months and seeing what it does, while you change its temperature, power supply voltage, and wave test magnets at it. You will learn something surprising doing that anyway. I don't know what you will learn, but trust me, engaging with reality always teaches something unexpected. \$\endgroup\$
    – Neil_UK
    Jun 17, 2023 at 6:03
  • \$\begingroup\$ allegromicro.com/en/products/sense/linear-and-angular-position/… \$\endgroup\$ Jun 17, 2023 at 6:35
  • \$\begingroup\$ Sorry I meant the total magnitude of the field is not 0.1 Gauss the change in magnitude of the field is 0.1 Gauss. The constant signal would be in 100s of Gauss as the signal is biased with a permanent magnet. \$\endgroup\$ Jun 17, 2023 at 6:37

2 Answers 2


The data sheet says the output is ratiometric, so that's a good start. It means you can use a Wheatstone bridge arrangement, and a weighscale ADC like an HX711 as an example, to reject the steady mid-scale output, and just read small changes.

However, drift and tempco of the device itself could be a problem.

On page 5, despite the device claiming to be temperature compensated, it specifies a drift of less than +/- 10 G across the temperature range. To be confident about 100 mG, you will need to hold the device at a constant temperature. An alternative would be to either find a way of switching your target field on and off, or to know when it was off, to take a new zero reference level.

There doesn't appear to be any long term offset drift specification. If you need long term stability, this could be the one that bites you.

On page 4, the input referred noise is given as 1.3 mG/Hz. That should mean you won't need much filtering to keep the noise well below 100 mG. If you aim for (say) 10 mG rms noise, you would need a filter of less than (10/1.3)2 = 60 Hz bandwidth. That's easy enough to implement in software.

Resolution is not an issue for sigma-delta weigh-scale ADCs, they have it to burn, the question is noise and drift, specifications for which are harder to come by and interpret. Noise of even the cheap HX711 seems to be more than adequate on a quick scan of its data sheet, though its spec of 90 nVrms is hampered by not giving the bandwidth. They do spec 0.4 mV offset drift, though the conditions for that are not clear. Unlike drift in your magnetic device however, drift in an ADC is easily removed by using a zero-ing switch at its input.

My recommendation is you get a cheap weighscale sigma delta ADC, there are plenty of options about, hook it up to your Hall device, and wave some small magnets or drive a Helmholtz coil near the Hall, and see what you get. Then you'll discover which device is going to limit you, and how careful you need to be with temperature.


I designed an instrument to read current from about 1A to 10,000A (AC RMS), using a 1000A/100mV shunt. So reading 1 A involved an input of 100 uV. I used an instrumentation amplifier AD620 connected for unity gain, followed by an 8 position programmable gain circuit for 50, 100, 250, 500, 1000, 2500, 5000, and 10000 A full scale. Accuracy was specified as 1% FS, but actual performance was mostly better than 0.2%, or 0.1 A on the most sensitive range, which corresponds to 10 uV resolution. The AD620B has a gain error of 0.01% and 10 ppm linearity. It used a 10 bit ADC, which has a granularity of about 0.1%. This was measuring 60 Hz power line frequency current, using a true-RMS computation algorithm.

So, if your transducer has a sensitivity of 5mV/G, a change of 0.1G would correspond to 0.5 mV or 500 uV. That is well within the capabilities of an instrumentation amplifier like the AD620. I designed my circuit around 1988, and there many more choices offering even better specs and lower cost. According to Wikipedia, the magnitude of Earth's magnetic field at its surface ranges from 25 to 65 μT (0.25 to 0.65 G). The 2.5 V offset is no problem for a good IA with adequate CMRR.

To be concise, then, there should be no problem doing what you need.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.