# Split supply from 24V power supply using uA741 opamp

I want to connect a 0-3V analog sensor output to the 0-10V analog input of a PLC.

In order to avoid losing resolution, I have amplified the sensor signal using a non-inverting amplifier circuit (based on the old uA741) so I get the full resolution. It works nicely using a +-15V dual power supply, but I would like to replace this power supply with a 24V one, as this is the one I have available in my test area. I have checked that I can design a circuit to split the 24V to +-12V using opamps, but I have two questions:

1. Can I also use a uA741 for this? I guess that I can't, because Vcc max is +-18V, so the 24V is outside of the range. Any alternative to this old opamp that is easy to find?

2. Will the resulting +-12V be enough to avoid clamping when going from 0-3V to 0-10V for the uA741 non-inverting amplifier circuit (for even uA741 or other ones?)

• Show the rail splitting circuit because it all rests on this. It also rests on not using the 741 ever again for any application! You need to state accuracy of conversion of signal especially if you choose an op-amp that is only slightly better than the vintage 741. Commented May 22, 2023 at 16:15
• @Andyaka I have already read your post about reasons to not use 741, nice one!. However, I tried with 741 as it was the only one available in my lab right now and I thought it was a good way to play and learn. Anyways, I can buy any other opamp. What do you mean exactly by rail splitting circuit?. Regarding accuracy, I don't need much as my sensor measures ph and I only need to detect changes of it around 20% of the values: for example, ph changes from 6 to 8 when total ph range is from 1 to 12. Commented May 22, 2023 at 16:22
• So, if the input supply is +24 volts and 0 volts, is the output of your rail splitter +12 volts and -12 volts relative to 0 volts? Commented May 22, 2023 at 16:27
• Mandatory read: electronics.stackexchange.com/questions/304521/…. Commented May 22, 2023 at 17:56
• No, that's incorrect; a rail splitter does not generate a voltage below the 0 volts of the incoming supply. Show your circuit is my advice. Commented May 22, 2023 at 20:17

So, if the input supply is +24 volts and 0 volts, is the output of your rail splitter +12 volts and -12 volts relative to 0 volts? – Andy aka

@Andyaka I thought that +-12V for the output reail splitter is derived from the fact that if I have 24V power supply, the splitted value shall be always half of this 24V. So yes, +-12V relative to 0 volts.

No, that's incorrect; a rail splitter does not generate a voltage below the 0 volts of the incoming supply.

This is your main problem; your input signal is referenced to 0 volts of the incoming 24 volt supply and, your rail splitter is doing literally what it says on the tin; making a split supply mid-point half way between 0 volts and +24 volts. It's not making a negative rail below 0 volts and, that's what you need for the rubbish old 741 and many other ageing op-amps.

However, you might be able to get away with a rail to rail op-amp powered from 24 volts (or a lower regulated voltage derived from the 24 volts) but, the output will never get down to true 0 volts but, if you are happy that it gets down to maybe +100 mV then this would work.

• It is clear now, many thanks. I think that not getting to 0V is not a critical for my application, so I will probably try with a rail to rail op-amp. Commented May 22, 2023 at 20:58
• @bardulia make sure the output is R2R and the input range encompasses 0 volts (quite commonly available but easy to choose the wrong part). Commented May 23, 2023 at 8:18
• I don't see why the input signal would need to be referenced to the 24V supply's 0V, that would be like referencing it to the -15V on the original supply. You would reference it to the rail splitter output. Commented May 23, 2023 at 13:39
• @bardulia that looks like it should do the job. You could probably find a cheaper one with less bandwidth if that's of interest. Commented May 23, 2023 at 21:49
• @GodJihyo maybe it's a powered sensor of some type that makes it more convenient to keep it referenced to 0 volts. Maybe it's beyond the control of the OP? Commented May 23, 2023 at 21:50

$$\\pm\$$18 V is 36 V total, so you're within the limits with 24 V. The data sheet recommended operating conditions section says $$\\pm\$$15 V nominal, $$\\pm\$$10 V minimum and $$\\pm\$$18 V maximum, so $$\\pm\$$12 V should be fine.

For something like a rail splitter to power an amplifier you shouldn't need to worry about accuracy too much, one problem you might run into though if you're also using a an LM741 (or other non-rail-to-rail op amp) for the amplifier is how close to the positive rail your output signal will be. You want 0 to 10 V, the positive rail is going to be 12 V, so that's only 2 V headroom and with an LM741 that might get you into clipping.

Since your output will only be going down to 0 V you can sacrifice some of the negative headroom for more positive headroom by making your voltage splitter output voltage be something like 9 V. Now you'll basically have a +15 V / -9 V supply and you'll have 5 V between your max output signal and the positive rail.

Here's an example of that, I don't have the LM741 model so I used LT1001. The rail splitter bottom resistor is stepped from 15k giving $$\\pm\$$12 V to 9k giving $$\\displaystyle +\$$12 V \ $$\\displaystyle -\$$9 V. The input to the $$\\times\$$3 amplifier is a 4 V triangle pulse (to exaggerate the clipping as an example), with an even rail split it clips at 11 V, with the uneven rail split it allows the full output of 13.3 V.

If you need the virtual ground to handle more current than the op-amp is capable of, for a 741 maybe 25 mA, you can add a simple transistor buffer. Something like this should get you at least a couple of hundred milliamps.

simulate this circuit – Schematic created using CircuitLab

The 741 is well within supply voltage limits and can be used to make a rail splitter.

simulate this circuit – Schematic created using CircuitLab

Above is a simulation of such a circuit with a load switched from -2.5mA to +2.5mA and it is reasonably stable. Obviously your 24V supply must be floating wrt where the pseudo ground is connected or you'll just short it out, if one side of it is grounded to the PLC input you'll have to do something else like regulate the 24V down to 12 or 15V if necessary and use a DC-DC converter to get negative voltage (or use a single supply op-amp circuit and trash the 741s entirely - hint hint).

The amplifier 741 should be okay too- it is guaranteed to swing to +10 with a 12V supply provided the PLC input does not present a load of less than 10kΩ.

Of course a better op-amp would be preferable.