# Is it possible to use a dual rail power system to power op-amp which is designed for a single supply?

I am designing an amplifier circuit using the LM386 amplifier. I am using the schematic below as a guide. I have access to 3 pins: +5V, virtualGND and -5V. Is it possible to power this amplifier using those pins? This amplifier only works with a single supply (not dual-rail apparently) so I'm wondering whether it would work with +5V as the supply voltage and replacing the ground in the circuit below with the virtual ground (which the audio input signal oscillates about). Thanks in advance, any help would be greatly appreciated.

• Your assignment for tonight is to write "The LM386 is not an op amp" 100 times. – Matt Young Aug 25 '15 at 14:23
• @MattYoung I think that's a borderline implied assertion. – Andy aka Aug 25 '15 at 14:26
• @matt young and andy aka, curious as to why you say the LM836 is not an op-amp? is it because it doesn't do negative feedback? – BenG Aug 25 '15 at 21:44
• @BenG Because it just isn't. Its input impedance is $\approx 50\text k \Omega$ and gain by default is 50. It is an audio power amplifier. – Matt Young Aug 25 '15 at 21:47
• Ahh. thanks @matt young in my addled brain I lumped them together. But of course the only real similarity is gain :) – BenG Aug 25 '15 at 21:50

It won't work as you wish - the ground in the circuit must be connected to the most negative supply rail i.e. the -5V rail. You might be able to couple your audio to the circuit thru a capacitor but it could be quite noisy. This could be overcome with an audio coupling transformer but, maybe just try running it from a single 5V rail (yes it will work at 5V or even 4V for some versions) then consider using a boost regulator to give you a clean (after extra filtering) 10V or 12V.

• You're correct that the ground in the circuit shown should be connected to the negative supply, not 'virtual ground', but I don't follow why it won't then work - the input and output are shown coupled by capacitors, and the input seen by the LM386 will then be referenced to its negative supply rail as the datasheet says it should be. If your power supply is noisy you may get noise, but that's just the same with dual rail supplies. ti.com/lit/ds/symlink/lm386.pdf – nekomatic Oct 1 '15 at 10:02
• Because it might be very noisy. – Andy aka Oct 1 '15 at 10:46
• @nekomatic my guess is that GND is usually pretty noiseless compared to Vcc/Vee, since they drop voltage under load and catch all the hi-freq noise easily - and GND doesn't (if it does, you can always make the ground plane bigger etc.). As such, I'd say that a device intended to work Vcc/GND will get about 2x of the PS noise when powered from Vcc/Vee rails (note that op-amps and other split-supply devices are, AFAIR, designed in a way that tries to cancel out the supply rail noise if possible). – user20088 Dec 1 '16 at 16:20
• @vaxquis the OP says in the question, that it is a virtual ground so I suspect it's formed by two resistors and a capacitor so I don't consider it to be as infallible as you might but who knows. Hopefully should get round to opening my mail tonight. If not then I'm considerably stupid LOL. – Andy aka Dec 1 '16 at 17:49
• @Andyaka I never considered it infallible - it's just that IMO even vGND can be more resilient to load noise vs PS lines due to its symmetric nature and the ability to easily increase its capacity. I guess it'd also be dependent if OP's 5V/vGND/-5V ain't really e.g. a +10V/GND setup with vGND made by splitting the supply line or other similar setup. Most supplies with real (non-virtual) -5V supply I've seen provide a real GND vs a virtual one; only some didn't (i.e. a old simple xformer PSUs without center tap). – user20088 Dec 1 '16 at 18:43

I've seen this amp operating with positive and negative rail supply

I've seen this amp operating with positive and negative rail supply https://www.youtube.com/watch?v=1zRamh9HMQc

which should be no surprise really. Then you don't need the output capacitor.

• He uses +- 4.5v ... which makes sense because the datasheet for this amp shows that after 0-9v it doesn't really give much more (up to 12v though sometimes I wonder if it handles input range better run at a higher voltage to reduce clipping even if its not going to give any more actual amplification power. – Michael Nolan Oct 21 at 14:34