# Dual power supply for Op Amp

We have dual power supply units (+- 12V DC with common GND) at our college lab using which I am able to implement my project (digital thermometer) which uses op amps as shown in the figure.

I'm using an LM35, three op amps, an Arduino Uno, and an LCD display. I need to make this a product which can be run using only one external power supply (+12V DC adapter for the Arduino).

I tried using a voltage divider circuit to supply +6V and -6V DC to the op amps. However, when I connect the LM35, feedback resistors and other components, the voltages measured at various stages are not as expected. In fact, the input voltage itself changes it's value from +6V to +2V and the other voltage changes to -10V from -6V. One thing I have noticed is that there is a constant voltage difference of 12V between the input pins.

Why am I getting weird values of voltages? Is this called the "loading effect"? What are the other possible ways of supplying power to the op amps using the 12V adapter?

EDIT: This is the schematic that works with the dual power supply unit (+-12V) at my college. I want to make this device portable. Hence, I can't use the dual power supply all the time. Need a solution for this...

• Your diagram doesn't match your description. According to the diagram, your op-amps are powered by +12V and -12V, and you have a voltage divider providing 1.3V. If you've replaced the 12V with some voltage divider providing 6V then yeah, you are going to have problems. A voltage divider can't provide power to the op-amps. An unloaded divider can provide a reference voltage, but can't deliver enough current to power an amplifier - the amplifier will act like another resistor in parallel to part of the divider and change output voltage. – JRE Sep 17 '15 at 14:31
• @JRE I totally agree with you. I just wanted someone to confirm that. The diagram which I've shown is the one which I use in college. It works perfectly fine. The voltage divider output is given to the summing amplifier because the transfer function of my application requires a constant voltage of 1.3V to be added to the output. Anyways, you cleared my doubt! Thanks. – Navin Sridhar Sep 17 '15 at 14:37
• In general, when asking a question, you should show the circuit that doesn't work, not a different one that works. – Pete Becker Sep 17 '15 at 16:29

1. The µA741 can be used in single voltage applications.
2. The 741 is pretty lousy as far as op-amps go.
3. There are better op-amps that work better on single voltage supplies.
4. You cannot power an op-amp from a voltage divider (insane things like two 1 Ohm resistors excluded.)

What you need to do is to look up single voltage operation for op-amps, and see what you need to do. Hint: It does involve a voltage divider, but not powering the amp.

Here is an example of a 741 running on a single voltage:

• I've got an LM324 and an LM358 too. Since two of my op-amps will be used in the inverting configuration, I don't think I can go ahead with using single supply as it will clip the negative output. – Navin Sridhar Sep 17 '15 at 14:45
• No, it won't. You work around a virtual ground that is at 1/2 of VCC. So, with 12Volts, you have a virtual ground at 6Volts. – JRE Sep 17 '15 at 14:48

What are the other possible ways of supplying power to the op amps using the 12V adapter?

You can use an inverting switching regulator like this: -

The one above takes +5V in and produces -5V out but equally you can use the same device to convert +12V into -12V. Read the data sheet of course.

• This seems to serve my application well, but this IC uses way too many extra components. The 7660 is another IC I'm planning to use. – Navin Sridhar Sep 17 '15 at 14:43
• @NavinSridhar Well, your design uses two op-amps that are basically redundant along with 4x 10k resistors and no-doubt op-amp decoupling capacitors so how do you justify your "way to many extra components"? – Andy aka Sep 17 '15 at 14:48
• I have an LM324 which is a quad op-amp IC. I also had the option of using a single op-amp in the difference amplifier config. But it doesn't have very good repeatability and doesn't offer accurate gain all the time. I think the 7660 can't provide high current, so I might as well go with this DC-DC converter. I'm just waiting for better solutions. – Navin Sridhar Sep 17 '15 at 14:55
• The LM324 sounds a better idea than dinosaur crappy 741s. I'm 57 and I have been doing analogue stuff for near enough 40 years and I've never used a 741 in all that time and there's a reason... did I say they were crappy LOL. – Andy aka Sep 17 '15 at 15:00