I'm still a bit fuzzy about point 1 and 2. When do you bypass the virtual ground to both rails and when do you only bypass to the most negative rail? Does it depend on whether you reference the signal to virtual ground or to true ground?

I think that the example one is quit bad one in terms of an bypass. In the first case we have a AC voltage divider, which is not good. All noise present at Vcc will appear at VGND only 0.5 smaller (capacitor divider).
The second case is better one, because this time C1 and R1||R2 for a low pass filter, so if we properly select the component we can attenuate the noise.

 

The OP never told us why he needs a negative supply.
I have made hundreds of circuits with opamps that need a positive and negative supply but without a negative supply and without a virtual ground circuit by simply biasing the opamps at half the supply voltage like in the virtual ground circuit, and using coupling capacitors.

 

The important part that needs to be discussed is what you choose as a COMMON reference point at which all inputs and output voltages are referenced. It does not have to be ground.

Here are three ways to create a reference beside GND.

1. Use a pair of resistors to create a voltage divider. Follow this with capacitor filtering.

2. Use the voltage divider and an op-amp to create a low impedance voltage source.

3. Use an IC virtual ground such as TLE2426.

Reference: https://tangentsoft.net/elec/vgrounds.html

Hi Mr FishAndChips, thanks for sharing your insight. Could you please tell me the why of the precise why of the caps in the virtual ground method ? You said it's for filtering, but what does it filter precisely? High frequency noise or just about any noise? Why only 250uF ? Wouldn't using higher valued caps add more "inertia" to the virtual ground ? Say for example one of the rails generates a ripple, then the ripple shorts through to the other rail through the virtual ground or not really? Or does it work by adding inertia to voltage changes so that if they werent there the ripples would show up more rapidly at the ground point? Also, don't passive filters with a capacitor also require a resistor or maybe a chokie ? My digital design teacher once told me to make a filter using just a capacitor and nothing else, and I thought he messed up because just a cap can't be a filter (I guess). But I am sure you didn't mess up, maybe the voltage divider resistors take part in the AC filter? Please MrFish&Chips give me the light

 

All circuits have resistance, inductance, and capacitance, even though you did not employ one intentionally.

When you put a capacitor in a circuit it still acts as a filter taking into account all the inherent and hidden resistances in the circuit. Even the capacitor alone has internal resistance which we call Equivalent Series Resistance. This, in itself, is a very important parameter to be considered when one is choosing a capacitor for a given application. Every application is different and calls for different analysis.

Capacitors used in power supplies fall into two major functions, (1) storage of charge, and (2) suppression of noise.
Storage capacitors tend to have larger values, 100μF - 10mF depending on the power needed to be supplied. It is still a filter, one with a very long time constant. Because of the large capacitances, these are usually polarized electrolytics made from aluminum or tantalum.

Capacitors used for noise suppression are chosen depending on the frequency of the noise. The capacitances can range from 1nF to 10μF and are typically non-polarized ceramic capacitors. It is not unusual to see two or three different values placed in parallel in order to cover a wider range of noise frequencies. Again, they still act as filters, in this case with very short time-constants.

You do not have to include an actual resistor in order to make a capacitor function as a filter.

 

MrFish&Chips

Who said anything about fish?

 

Who said anything about fish?

I don't fish and I aint fishing.

 

Who said anything about fish?

Here in the UK they love Fish&Chips. Everybody eats it.

 

All circuits have resistance, inductance, and capacitance, even though you did not employ one intentionally.

When you put a capacitor in a circuit it still acts as a filter taking into account all the inherent and hidden resistances in the circuit. Even the capacitor alone has internal resistance which we call Equivalent Series Resistance. This, in itself, is a very important parameter to be considered when one is choosing a capacitor for a given application. Every application is different and calls for different analysis.

Capacitors used in power supplies fall into two major functions, (1) storage of charge, and (2) suppression of noise.
Storage capacitors tend to have larger values, 100μF - 10mF depending on the power needed to be supplied. It is still a filter, one with a very long time constant. Because of the large capacitances, these are usually polarized electrolytics made from aluminum or tantalum.

Capacitors used for noise suppression are chosen depending on the frequency of the noise. The capacitances can range from 1nF to 10μF and are typically non-polarized ceramic capacitors. It is not unusual to see two or three different values placed in parallel in order to cover a wider range of noise frequencies. Again, they still act as filters, in this case with very short time-constants.

You do not have to include an actual resistor in order to make a capacitor function as a filter.

Ok I understand. Thank you a lot for your time. But how do you find the "topology" of the filter if the resistances are all spread out everywhere? In that example, you have the capacitors directly from V+ to V-, so would you say the filter is composed mainly of each capacitor + their internal series resistances? Or would you include the resistances in the voltage divider as forming other filters as well ? For example the capacitor close to V+ and the resistance close to V-, and vice versa, do those form filters as well ? In fact, would you find the actual precise voltage at that virtual ground, at all times and all frequencies, if you combine the parallel impedances of R1 and C1, and R2 and C2, and then find the final voltage divider of the series result ? Then you have a formula dependent on frequency and time, and hence you know what frequencies the filter filters best ?

For instance, combining R1 and C1 you get (R1 * jwC_1) / (R1 + jwC_1), and combining R2 and C2 you get (R2* jwC_2)/(R2 + jwC_2). So now you have the new voltage divider, and can find the final voltage at that point, which gives you the filter as a function of frequency and time? For example taking the voltage divider formed by the two resulting series impedances formed by R1 and C1 and R2 and C2, and then multiplying by e^jwt, and finally taking the real part of that, you get the voltage at the virtual ground at all times and frequencies ?

Thanks a lot.

 

It is derogatory to bastardize someone's username.

 

When a capacitor is employed as a charge storage device, the significant resistance to consider is the load itself.

When a capacitor is employed for high frequency suppression, the mathematical analysis is much more complex. Inherent inductances in component leads and circuit traces must be included. This is not a simple and straight forward analysis.

 

It is derogatory to bastardize someone's username.

That is where I was going with my deleted pizza comment. My comment gets deleted and your's is "liked" by the same moderator.

 

That is where I was going with my deleted pizza comment. My comment gets deleted and your's is "liked" by the same moderator.

I suspect that the moderator believes that your deleted post adds no value to the thread.
My post is in direct response to the TS that his action is not appropriate and he needs to make amends, i.e. I am not beating about the bush.

 

It is derogatory to bastardize someone's username.

Lol, no, I was just kidding you

 

Lol, no, I was just kidding you

You need to take my statement seriously.

It is derogatory to bastardize someone's username.

This is not a joking matter.

 

You need to take my statement seriously.


This is not a joking matter.

Ok MrFish&Chips.

 

Ok MrFish&Chips.

I am going to give you some time to seriously think about your behaviour on a public forum.

 

I am going to give you some time to seriously think about your behaviour on a public forum.

I have over 25 accounts here. You can't ban me. If you banned me simply for saying "OK MrFishAndChips", I think you have some serious mental problems. Very childish. I have no idea what you're on about abusing your power as a moderator here. I will report you to the owners!

 

I have over 25 accounts here. You can't ban me. If you banned me simply for saying "OK MrFishAndChips", I think you have some serious mental problems. Very childish. I have no idea what you're on about abusing your power as a moderator here. I will report you to the owners!

Wow.

Just show a little respect for other people's boundaries... keeping in mind that these are the same people who you're asking help from, and who voluntarily give that help.

You wouldn't need to have 25 accounts if you used just one of them with a little more respect and decency.