Originally posted by Dinosaur in EngineerZone feedback community on May 10th, 2010 (before Amplifier community was active)
I have a ad620 amplifier board that has the circuit as per attached with a gain of 495
At an input of 0.00 mV I need to pull the output of the ad620 down closer to -5 v
At the moment it sits around -0.3
I have tried what I normally do with other adc's, and that is put a bias resistor between
-5 and one of the +in or -in. (Shown as 510K on the circuit) To do this with the ad620, I measure the output of the
load cell with the In+ and In- disconnected and select a resistor that will get me say -3.00 vdc.
Then when I connect to the ad620, this rises to about -0.7 and no amount of resistors will bring it lower.
The output of the ad620 shows what is expected with that input.
The resistor initially was about 29k. This also introduces a large dead area on the loadcell, where I need
to apply a significant load before it will register.
I have been doing this kind of offset adjustment for more then 20 years with the Ad7701 and others, and have never
struck this kind of problem before.
The technique that you are using makes sense. You are unbalancing the wheatstone bridge with the extra resistor, which should create an offset. One way to think of this is that you are putting the resistor in parallel with one of the resistors in the bridge. In addition to creating the offset, this will also change the gain of the bridge somewhat.
In order to move the output voltage lower, a resistor should be tied from the +5V to the -IN terminal. Alternatively a resistor could be tied from the -5V to the +IN terminal.
The schematic shows a capacitor connected to the output of the AD620. We normally discourage this practice, since adding capacitors to the outputs of amplifiers often causes them to go unstable. The AD620 can drive a few hundred pf, but for capacitance more than that, a resistor (say 100 ohms) should be placed between the capacitor and the output of the AD620.
In the question it sounds like the AD620 has abnormally high input bias current and is pulling the bridge input voltage to -0.7V. This is typically a sign that the input voltage is out of the in amp's input range. I would recommend checking the power supplies of the AD620 and verifying that -Vs is indeed connected to -5V. My guess is that -Vs may be connected to ground. If the power supplies check out, it is possible the input stage of the AD620 is damaged (say by an overvoltage event). I would recommend checking with a fresh AD620.
As an alternative to unbalancing the bridge, the reference voltage could be driven with a negative voltage. This method is easier to analyze and does not change the gain of the bridge. However it will likely result in a more expensive BOM, since you will probably need an additional op amp stage. The AD620's reference pin needs to be driven with a low output impedance source, such as an op amp, rather than a high output impedance source, such as a voltage divider.