The objective of the circuit below is to get the difference of the hall effect output and a 1 volt reference, then amplify it. When I first built the sensor and tested it, it worked very well. This is how it behaved: when no magnetism was present it would be outputting around 0.3 volts, and when south pole magnetism was present it would output 4.5v.

After leaving the sensor connected for a few days it "decided" to switch directions (sorry for the poor terminology), meaning: When no magnetism is present it outputs 4.8v and it would no longer react to south pole magnetism, but when a north pole magnet is present it will go to 0.5ish volts.

Why would it all of a sudden flip? @Spehro Pefhany pointed out below a mistake I made by adding 0.1uf capacitors. I removed these, and it did not change anything but did I potentially damage the chip?


I think I may have solved why it "flipped", but still need a solution. When I physically rotate the hall effect sensor from 0 to 180 degrees, the signal inverses from 4.9v to 0.3v. So I am assuming this is from external noise? The sensor itself is nowhere near a magnet or current carrying wires when I perform this test. Any ideas?

enter image description here

  • \$\begingroup\$ Please measure and report the voltages at all of the amplifier pins. \$\endgroup\$ – Elliot Alderson Dec 30 '19 at 17:01
  • \$\begingroup\$ I made some large edits to my question, it appears as though it may be the amplifier? With no magnetism at the hall effect, the voltage is at 1.000V, and the voltage at the shunt output is 0.999V. The other values are listed in my question above \$\endgroup\$ – MattG Dec 31 '19 at 0:45

You should not have a 0.1uF capacitor on the amplifier output. Remove that and it may magically start working again. From the datasheet:

AD626 is stable driving capacitive loads up to 50 pF (G10) or 200 pF (G100).

So it's probably oscillating, which can manifest itself in strange behavior such as apparent large DC drift with temperature or time.

  • 1
    \$\begingroup\$ good call....... there is a Filter pin with 100k that is intended for slew rate reduction. Driving low ESR caps can cause ground shift with positive feedback. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Dec 30 '19 at 17:24
  • \$\begingroup\$ Thanks @spehro Pefhany for the answer, I did not spot that in the datasheet. I removed those capacitors and it still did not solve the issue, however I did do some more troubleshooting and edited my question above. I'm not sure what the correct etiquette is for making large edits, so I apologize if that's not in the stack exchange spirit. I'm wondering if I damaged the chip by using those 0.1uf caps? It is a pretty pricey chip so I hope not! \$\endgroup\$ – MattG Dec 31 '19 at 0:44

One thing I noticed regarding the voltage reference is that you are running slightly higher current through it than it is rated for. The datasheet says its operational range is to 10mA (although the absolute max is twice that, but who knows how it handles overcurrent). The datasheet also says that the voltage reference is sensitive to electrostatic discharge. Try to measure it to see if it still provides the 1V. Also try to look for any loose wires (or iffy connections that don't make good contact). Aside from that I can only advise you to measure Hall output to see if that is still okay.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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