use op amps with single supply

[aruna1]

hi
is there anyway to use dual supply opamps like TL082 or TA4558 with single supply?
thanks

[Soeren]

Hi,

Yes, any op-amp (single or dual supply) is used the same in that respect - you create an artificial ground, either by just connecting two resistors from + and - to a common point and a capacitor in parallel with one of them to help maintain the voltage (the cheapest way, but not always enough).
To lower the impedance of this virtual earth, an op-amp used as voltage follower will keep the output from sagging under load.

This virtual ground doesn't need to be at the midpoint of the supply and in some applications it may be better to keep the "ground" closer to one line, to get more overhead in the other direction.

If used for AC signals (like eg. audio) the virtual ground must be tied to the real ground through a capacitor calculated for the lowest frequency used and the output should be referred to the real ground as well (AC coupling through a capacitor).

If you Google "operational-amplifier artificial-ground", I'm sure you'll get plenty more to chew on.

[aruna1]

hi
i read about virtual ground but still dont have clear idea how things connect each other. can you please point me to  some kind of "idiot's guide to opamp virtual ground"  ?

[Soeren]

Hi,

i read about virtual ground but still dont have clear idea how things connect each other.

Two resistors of say 2k2 each (for close to 1mA at a 5V supply), one from positive supply line and one from 0V.
Join the free ends in what is your artificial ground.
Use a cap of eg. 10µF (or whatever's handy) between this node and 0V.
This will do for most circuits.

BTW. It is sometimes called a supply splitter, if that helps getting a picture of it.


If the artificial ground is loaded too much, it will start to sag in the direction of the load. In that case, a common voltage follower op amp is used as a buffer.
If the ground is very capacitive, a resistor of ~100 Ohm on the output, with the feedback connected to the "outer" end of it will help.

can you please point me to  some kind of "idiot's guide to opamp virtual ground"  ?

I cannot find any guide on this particular subject, but it's mentioned in passing in several application notes on single supply op-amp design.

[aruna1]

Two resistors of say 2k2 each (for close to 1mA at a 5V supply), one from positive supply line and one from 0V.
Join the free ends in what is your artificial ground.
Use a cap of eg. 10µF (or whatever's handy) between this node and 0V.
This will do for most circuits.

BTW. It is sometimes called a supply splitter, if that helps getting a picture of it.


If the artificial ground is loaded too much, it will start to sag in the direction of the load. In that case, a common voltage follower op amp is used as a buffer.
If the ground is very capacitive, a resistor of ~100 Ohm on the output, with the feedback connected to the "outer" end of it will help.

so where does this artficial ground point connect in the op amp ? Vee pin or non inverting in put?

[waltr]

The virtual ground connects to the non-inverting input.
What this does is put the mid-range of the op-amp's input/output between the + and - power supply. This way the op-amp can swing + & - from the virtual ground without running into the power supply rails.

Vee connects to the power supply ground or Common (0V).

Prototype an inexpensive op-amp with a virtual ground and play until you understand how it works.

[aruna1]

Hi Waltr
so do I have to do this even my input signal has no negative side? (so output doesnt go negative side)

and do I have to do this even with single supply?

[TrickyNekro]

Just to say if I were you I would go with a voltage divider - Darlington configuration... Ask Soeren to explain, I bet he would be happy... :-p

[aruna1]

Just to say if I were you I would go with a voltage divider - Darlington configuration... Ask Soeren to explain, I bet he would be happy... :-p

@soeren

Hi can you please explain this 

[TrickyNekro]

Generally, I mean something like the picture below, unfortunately their regulation isn't very good...

You know there is a standard voltage drop between base and emitter, for a single transistor that's 0.7V, for a Darlington it's 1.4, cause is a double transistor, or a transistor in series.

The bad thing is that their regulation heavily depends on the load...

Anyhow, say you have a 12V voltage supply, how come you don't use a 7806??? That's simple enough.
And there may be also 78L06 that's the low power equivalent that also comes in a "standard transistor" package... To - 92 if I remember right.

Anyhow, there is also the LM317 that you can use.... Many many choices really ;-)

Best Regards, Lefteris
Greece

[aruna1]

Generally, I mean something like the picture below, unfortunately their regulation isn't very good...

You know there is a standard voltage drop between base and emitter, for a single transistor that's 0.7V, for a Darlington it's 1.4, cause is a double transistor, or a transistor in series.

The bad thing is that their regulation heavily depends on the load...

Anyhow, say you have a 12V voltage supply, how come you don't use a 7806??? That's simple enough.
And there may be also 78L06 that's the low power equivalent that also comes in a "standard transistor" package... To - 92 if I remember right.

Anyhow, there is also the LM317 that you can use.... Many many choices really ;-)

Best Regards, Lefteris
Greece

oooooooook
I have no idea what you are talking about  maybe you are replying to wrong topic?

[TrickyNekro]

I had the idea you were in need on a negative voltage with a single supply... right??? 

[aruna1]

I had the idea you were in need on a negative voltage with a single supply... right??? 

ya sort of but i dont see how you get negative supply from darlington couple or 7806

any way my original requirment is

I have 9V positive regulated dc supply
and i have a TL082 op amp which made to work with dual supply

since i dont have both negative and positive supplies I want a method to make TL082 work only with single supply

[Soeren]

Hi,

Here's an example with a TL082, with one of the amps used as buffer. The other one is free to use (but should be terminated correctly if not)

Remember a cap on the supply as well.
The value of C2 depends of how capacitive the load may be, if it isn't capacitive at all use eg. 100µ. R3 can be 47R to 100R
C1 can be eg. 1..10µF depending on the value of R1 and R2.

The artificial ground does not necessarily connects to one of the op-amps pin directly, it all depends on the circuit you're making.

It would be better if you post the circuit you're making, rather than asking for something that may not be needed at all.

[aruna1]

Hi soren

What I want is to amplify a ultrasonic receiver transducer signal. currently I'm using TLC272 single supply opamp for this.
here is the diagram.
now i want to replace this op amp with TL081 which has null adjustment.so I thought to begin with TL082.


and in your diagram lets say i want to make non inverting amplifire with gain of 10 how can i modify the circuit?

[Soeren]

Hi,

now i want to replace this op amp with TL081 which has null adjustment.so I thought to begin with TL082.

TL082 doesn't have offset adjust pins. Not that they're needed, any op-amp can be offset adjusted by biasing with a few resistors.

But, more important... Why do you want to adjust the offset in this circuit?

and in your diagram lets say i want to make non inverting amplifire with gain of 10 how can i modify the circuit?

In your circuit, lift the 0V end of X1 and R2 and connect them to two 2k2 resistors and a 1µF to 10µF cap connected as R1, R2 and C1 in the schematic I posted. I don't think you'll need to use an op-amp to buffer the art. ground.
You should do this for any op-amp, whether it's single or dual supply.

The gain in your schematic is 11 (R1/R2+1).
Did you measure the capacity of X1?
If it's more than 10nF, increase the value of C1 to at least 10 times (up to say 100 times) the capacity of X1.
Add a resistor of 910R (R1//R2) between C1 and pin 3 to match the impedance of each input. Even better, use a slightly larger resistor and make a (high impedance) splitter to keep the non-inverting input in check.

I can't imagined what you're trying to do with an LED on the output, but perhaps it's just for visual confirmation or something along that line.
When you make the artificial ground, then, when there's no input from X1, the inputs of the op-amp will come to rest at Vcc/2 and so will the output. Your amplified signal will thus be symmetrical around Vcc/2.

Lifting the ground side of the LED to Vcc/2 as well would keep it off with no input and (with luck) flicker weakly at 40kHz if the input is very strong. But then you'd need the buffer on the splitter.


Two CR2032 coin cells could be used to make a +/-3V supply.

[TrickyNekro]

I had the idea you were in need on a negative voltage with a single supply... right??? 

ya sort of but i dont see how you get negative supply from darlington couple or 7806

It's fairly simple and something like the circuit Soeren had drawn for you...

Söeren had drawn a buffer with an op-amp, I did it with a darlington... It's that simple...
You see transistors can be used as buffers... Funny thing is that that's their primary use.
To amplify current...

Most hobbyists see the transistor as a switch, or maybe some as a voltage amplifier, but it's made to amplify current...
The input is a current, the output is that current amplified with Hfe + 1 at the emitter and Hfe at the collector...

About the virtual ground... why still not use a LM317... It's possibly the best of all solutions...

You have a 9V regulated, nice... The LM317 outputs then 4.5V let's say....
If you perceive the LM317 output as the ground rail for the op-amp, then the real ground is at -4.5V for the op-amp and the real 9V positive rail is the 4.5V rail for your op - amp...

I'll post a schematic too, but I feel Söeren has covered most of the field needed to be explained...

Best Regards, Lefteris
Greece

[TrickyNekro]

And here is the circuit, the ground is the Ground for your microprocessor etc etc etc...

If 9V regulated is a must, then you should immigrate to 3V - 3.3V logic...

Just a note though... Should you use this circuit, then you must find out about precise rectification methods,
(NOTE: That's basically some Schottky diodes with some op-amps, one or two, depending whether you need half or full wave rectification)
as this circuit will give you negative voltages, that a microprocessor can NOT handle at its inputs...

Peace out,
Lefteris, Greece

[Soeren]

Hi,

And here is the circuit, [...]

Problem is... Your circuit is a voltage regulator with a static output voltage. If the supply changes, you no longer have the Art. Ground at the midpoint.

It could be done by driving the LM317 with an op-amp, a power op-amp could be used, or the op-amp could be used with a couple of power transistors, but now we're into a complexity level far beyond what's needed.

[TrickyNekro]

More than true... But when he said he had a regulated steady supply I assumed he was working at a bench...

Anyhow, Soeren's schematics, is a true half high rail voltage regulator... If you don't have a that steady power source, go with it...


BTW, btw..... Why not use an ATX... flexible as it is it can give you a variety of voltages... Plus it's steady enough, and has some crazy current delivering capabilities.

[aruna1]

The gain in your schematic is 11 (R1/R2+1).
Did you measure the capacity of X1?
If it's more than 10nF, increase the value of C1 to at least 10 times (up to say 100 times) the capacity of X1.
Add a resistor of 910R (R1//R2) between C1 and pin 3 to match the impedance of each input. Even better, use a slightly larger resistor and make a (high impedance) splitter to keep the non-inverting input in check.

I can't imagined what you're trying to do with an LED on the output, but perhaps it's just for visual confirmation or something along that line.
When you make the artificial ground, then, when there's no input from X1, the inputs of the op-amp will come to rest at Vcc/2 and so will the output. Your amplified signal will thus be symmetrical around Vcc/2.

Lifting the ground side of the LED to Vcc/2 as well would keep it off with no input and (with luck) flicker weakly at 40kHz if the input is very strong. But then you'd need the buffer on the splitter.


Two CR2032 coin cells could be used to make a +/-3V supply.

Hi

sorry for the late reply.I was stuck with some assignments and now I'm back.
I measured the capacitance and its around 2nF.
LED is for visual confirmation.I'm planning to feed the output of the op amp to LM567 tone decoder.

and another noob question;
lets say I use a single supply op amp for this circuit. then how should i connect ultrasonic receiver unit to the op amp input? is it correct to connect one pin of the transducer to op amp non inverting pin and other to the ground (0V) supply rail (in single supply) as in my circuit I posted above?

or should i connect one pin of transducer to non inverting input and other to a voltage divider as you mentioned above?

I have a feeling that if I connect one pin to a 0V supply rail it will cut off negative part of incoming signal and will reduce the sensitivity.

[aruna1]

And here is the circuit, the ground is the Ground for your microprocessor etc etc etc...

If 9V regulated is a must, then you should immigrate to 3V - 3.3V logic...

Just a note though... Should you use this circuit, then you must find out about precise rectification methods,
(NOTE: That's basically some Schottky diodes with some op-amps, one or two, depending whether you need half or full wave rectification)
as this circuit will give you negative voltages, that a microprocessor can NOT handle at its inputs...

Peace out,
Lefteris, Greece

LM317 trick is very cool. should give it a try.
thank you very much

[Soeren]

Hi,

I measured the capacitance and its around 2nF.

Then the 100nF coupling cap is fine.

LED is for visual confirmation.I'm planning to feed the output of the op amp to LM567 tone decoder.

OK, but what's the purpose of the '567?
If you want to measure distance, you want as little delay in getting the return pulse (or pulse train) back into the controller. If you use a '567, you'll get a delay with a magnitude depending on the quality of the signal (the lock time).

lets say I use a single supply op amp for this circuit. then how should i connect ultrasonic receiver unit to the op amp input?

Basically the same way as with a dual supply.

is it correct to connect one pin of the transducer to op amp non inverting pin and other to the ground (0V) supply rail (in single supply) as in my circuit I posted above?

or should i connect one pin of transducer to non inverting input and other to a voltage divider as you mentioned above?

With the 100nF capacitor, you could use a voltage divider (with high value resistors) on the non inverting pin instead, as the cap will "remove the ground reference".
I'd use the voltage divider (with a cap) for the "ground pin" of te transducer however, as you then won't need to lower the input impedance of the op-amp.

I have a feeling that if I connect one pin to a 0V supply rail it will cut off negative part of incoming signal and will reduce the sensitivity.

Due to the cap, the swing will be shifted, but since the input isn't at a known potential, it's not as predictable.

Go with two resistors and a cap - it's cheap and easy.

[aruna1]

OK, but what's the purpose of the '567?
If you want to measure distance, you want as little delay in getting the return pulse (or pulse train) back into the controller. If you use a '567, you'll get a delay with a magnitude depending on the quality of the signal (the lock time).

I tried with 567. before using 567 , circuit was responding to clap sounds also. (LED blinks when I clap) this not acceptable. but after adding 567 it only responds to the 40kHz transmission signal.
anyway can you explain this lock time and quality thing? I couldn't figure out such thing from 567 datasheet

With the 100nF capacitor, you could use a voltage divider (with high value resistors) on the non inverting pin instead, as the cap will "remove the ground reference".
I'd use the voltage divider (with a cap) for the "ground pin" of te transducer however, as you then won't need to lower the input impedance of the op-amp.

this I didn't understand. perhaps you can attach a diagram