- Ask a related questionWhat is a related question?A related question is a question created from another question. When the related question is created, it will be automatically linked to the original question.
I´m using an INA163 (instrumentation amplifier) with a dual supply, +-5V. But, to get a better performance in my project, I don't want to use a negative source anymore. So, my idea is to connect V- to the ground. But, I tried it and the amplifier responds in a very strange output. I can't figure out why.
When I use the dual supply, my output is something between 3.5 and 4.8 V. Using a single supply, I expect a gap smaller than this. To be exact, the INA163's data-sheet says that there's a lost of 1.8 V in each side of source. So, my gap should be (5 – 1.8) – (0 + 1.8) = 1.4 V. My inputs are very similar, with a small variation of milivolts. The interval measured with dual supply shows me that this small one should be enough. But my output with single supply is always constant (4V) to a very different values of input.
The only change that I did in my circuit is the V- to the ground. I researched in INA163's data sheet looking for something about single supply, but there's anything. Also read a Texas' application report, A Single-Supply Op-Amp Circuit Collection, SLOA058, but I still can't find out what's is happening. Anyone have any idea what's happening?
+12V and ground should allow the IC to properly function as long as the input common-mode range is within the limits specified in the datasheet and the output voltage range is not being violated. Here is an example of the INA163 being used single supply:
Notice that the input has a common-mode voltage is 6V, which is halfway between the two power supplies. Also the reference pin has been tied to a voltage between the supplies which prevents the output stage of saturating at either of the supplies. I've attached this simulation file which can be used in TI's free simulator: Tina-TI.
Just to clarify, you have an input differential voltage of 2V and a gain of 280? This will saturate the output of the amplifier as it will be attempting to produce 2V * 280 = 560V at its output, but will be limited by its power supply. The easiest method for determining the minimum input common mode voltage for an instrumentation amplifier is probably to sweep the input common mode voltage from 0 to mid-supply and determine when the output of the INA is at the proper voltage. For example, in the schematic shown below I have the INA163 configured in a gain of 280, with the reference pin connected to ground. I'm applying a 10mV input differential voltage, so the expected output should be 10mV * 280 = 2.8V
Now I can perform a DC transfer characteristic in Tina-TI to sweep the value of Vcm and determine when the output of the INA163 is at the proper value:
Notice the the output of the INA163 is not at the proper value until the input common mode voltage is at approximately 3V. This test should be performed for a multitude of input differential voltages to ensure proper functionality. Also, it may be helpful if you could post a schematic of your circuit here such that I can better understand your system.
Actually, I have an input centred in 0 volts with 250mV of variation.
To simulate this, I used the circuit bellow. In simulation, my output is correct, but I can't get this in real tests. Could you imagine where's my problem? Is there a common error in single supplies like this?
It's strange cause my output is correct when I've used a dual supply. I've changed only the supply and I can't get the right results anymore.
That circuit should not simulate properly. Duplicating your schematic with the output filter omitted, I get the following output for a 250mVpp input signal centered around 0V:
The INA163 will not be able to amplify signals below the negative supply voltage as you have shown in your schematic. With dual supplies the input signal no longer falls below the negative supply. Also, having such a large capacitor on the output, even with the resistor may cause instability.
Bruce beat me to it, but he's absolutely right. The input resistors are only trimmed for gain error and have no bearing on the CMRR of the system. Here's a quick screenshot from TINA showing the differential output voltage for an input common mode voltage of 10V (two ideal op amps, with one resistor mismatched on purpose):
Looking at your nominal impedance and damping requirements, I have to ask if your sensor is a geophone?
Ahhh my e2e telepathy is still calibrated! One thing to consider when selecting amplifiers for geophones is that the peak in the geophone impedance near resonance can often co-inside with the 1/f rise in current noise that bipolar input amplifiers exhibit. Because of this, current noise can be much more significant than it would initially appear. Here's an impedance measurement I took of a 10Hz geophone with a nominal 400 Ohm impedance :
One amplifier to consider is the OPA209 family. It's input voltage noise is 2.2nV/rtHz, which is slightly worse that the INA163 or ultra low noise op amps like the OPA211 or LME49990, but its input current noise is 500fA/rtHz, which is significantly less than other low noise bipolar amplifiers and may provide a noise benefit here.
As for filter design, our free program FilterPro will do component calculations for fully differential MFB filters (including component tolerance), it might be worth checking it out as it can be a huge time saver for filter designs.
