1
\$\begingroup\$

I would like to ask for help regarding a certain phenomenon related to amplification by the Analog Devices AD620 instrumentation amplifier (device home page, device datasheet).

schematic diagram

Vcc is 5V.

The AD620 receives a differential voltage from the KMZ10B magnetoresistive sensor, which has a Wheatstone bridge construction.

MY GOAL:

I wanted to amplify voltage from KMZ10B (which is about several tens of millivolts). In my circuit this voltage depends on direction of the magnetic induction of the neodymium magnet. A direct voltage results gives such a characteristic:

enter image description here

On the above graph:

  • X-axis = Neodymium magnet angle (degrees)
  • Y-axis = voltage (V)

Then I decided to implement the voltage reading from AD620.

HOW I DID THAT:

I used 4.7 kOhm resistor between pin 1 and 8. Datasheet of AD620 says that this resistance (called Rg) must be calculated like this:

Rg = 49.4 kOhm/(G-1)

G - gain

I used 4.7 kOhm resistor, so gain is 11.51. Reference voltage is about 2.5 V (from second op-amp: MCP6022) and voltage divider (half of 5V from Vcc)

I created a chart of this circuit (voltage measured between output of AD620 and GND) and it looks like below:

enter image description here

Now I decided to compare signals before and after amplifying.

I took the results in Excel made before using AD620, I multiplied them with Gain (11.51) and added a 2.5 volt reference (below).

y=((x*11.51)+2.5)

It was supposed to show if the amplifier really strengthens as the calculations show. And look, I compared both charts:

enter image description here

LEGEND FOR ABOVE GRAPH:

  • Blue chart is a real, amplified signal
  • Orange chart is a calculated signal.

PROBLEM:

It looks like something's going wrong and in my opinion it depends on in-amp AD620. The signal rise up quickly and falls slowly. What can it be caused by?

\$\endgroup\$

1 Answer 1

Reset to default
1
\$\begingroup\$

In your diagram, you have placed your 47k resistor across the outputs of the Wheatstone bridge (+Vo and -Vo). This has loaded up the bridge, resulting in measurement error.

The whole reason we use instrumentation amplifiers like the AD620 is to provide a high input impedance and avoid loading effects.

Remove the 47k resistor from the +/- inputs of the AD620, and place it across the RG pins (pins 1 and 8 as per the datasheet).

Let us know if you're still having trouble afterwards.

\$\endgroup\$
2
  • \$\begingroup\$ Oh, I'm really sorry, my fault, it wasn't 47k - it's only 4,7k (missed comma). Let me correct this on schematics. Calculating (49400/4700 + 1) gave me 11,51 of gain. But about removing this Rg at all - are you sure? :( I need gain, because in next step this signal is measured in ADC in my ATmega8. Signal without amplifying could be too hard to measure on Atmega and that's the reason. \$\endgroup\$
    – Proboszcz
    Sep 26, 2018 at 23:20
  • \$\begingroup\$ You aren't totally removing Rg, you are placing it across the dedicated Rg pins (pins 1 and 8). Right now it's across the input, which is incorrect. Placing it across the Rg pins will give you your gain and your necessary high impedance input. \$\endgroup\$
    – Synchrondyne
    Sep 26, 2018 at 23:27

Your Answer

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

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