Hi,

If you have a single-supply inverting opamp with a voltage reference on the non-inverting input like figure 3 from this link, would there be any benefit of adding another capacitor C3 of equal capacitance as C2 from the reference to the positive supply rail like in the attached image? Or would the result be equivalent if we ommit C3 and double the capacitance of C2?

Edit: sorry, that should be bypassing, not biasing. Can't edit the title.

 

If your input and output are referenced to the negative rail as they appear to be in this circuit then adding the extra capacitor would just add any supply noise into the op-amp input.

 

Ah, i see. So you should only put bypass caps between the voltage reference and the rail to which the signal is referenced.

 

I'm sure there are exceptions but that's a good general rule.

 

Hi,

If you have a single-supply inverting opamp with a voltage reference on the non-inverting input like figure 3 from this link, would there be any benefit of adding another capacitor C3 of equal capacitance as C2 from the reference to the positive supply rail like in the attached image? Or would the result be equivalent if we ommit C3 and double the capacitance of C2?

Edit: sorry, that should be bypassing, not biasing. Can't edit the title.

Hi HRS,

Just to add some detail to what the other members have said:

You definitely should not connect a capacitor from the reference point to the positive supply line.

The whole objective of the reference point is to provide a voltage, which is typically half of the positive supply voltage, but is still referenced to ground (0V) (or to any other point in a circuit).

Not only does C2 couple the reference point to ground, but C2 also shorts any ripple voltage and noise on the reference point (from the positive supply line) to earth so that the ripple and noise do not get into the non-inverting input of the opamp.

The value and type of capacitor have a big influence on the effectiveness of C2. For audio work, polypropylene capacitors are best, but where low distortion is not important a ceramic capacitor will suffice.

Both of these capacitor types have a low effective series resistance (ESR), which means they couple well and attenuate the ripple and hash well. For higher values of capacitance, tantalum capacitors are best, because tantalum capacitors also have a low ESR. Note that you need to take care when using tantalum capacitors but, in this application, they are ideal and safe because the current is limited by the top resistor.

You mention about the value of C2, so just to explain:

As you no doubt know, the reactance (XC) (like resistance) of a capacitor decreases as the frequency increases according to the formula, 1/(6.3 * F * C),
where:
XC is the reactance of the capacitor in Ohms
F is the frequency in Hz (cycles per second)
C is the capacitance in Farads

So, C2, in conjunction with the top resistor, forms a potential divider which attenuates the ripple and hash originating from the supply line and, because the capacitors reactance drops with frequency, the attenuation is greater at higher frequencies than at lower frequencies.

From this, it follows that C2 would need have a high capacitance value to attenuate low frequency ripple. Note that C2 has no effect on the gain of the amplifier, so you can make C2 whatever value is necessary to attenuate the ripple and hash from the positive supply line.

But there is another aspect to consider: leakage current. Electrolytic capacitors have a relatively high leakage current, aluminum types being the worst. Tantalum capacitors have a much lower leakage current but, the leakage current is still significant where high value resistances are concerned.

Leakage current has a quadruple whammy:
(1) unpredictably lowers the reference point
(2) generates noise
(3) generates distortion
(4) is highly temperature dependent

To all intents and purposes, non electrolytic capacitors, ceramic, polyester. polypropylene... have no leakage current.

Of course, all these problems can be circumnavigated by using positive and negative supply lines and an actual earth (0V).

phonic