Advantages of using bi-polar power supply?

[goldstache]

I've built about 50 or so circuits and am tackling a compressor circuit which has a bi-polar power supply.  Wanting more info on why some pedals use just a V+, Vref, and GND as the configuration where as others have V+, V-, Vref, and GND. 
I am self taught so......  I was hoping someone could fill me in on this a bit, from a novice standpoint.  Thanks in advance!!!!

[R.G.]

Using a bipolar supply lets you design circuits where the signal is at 0Vdc, not some offset bias voltage. This removes the issues of bias voltage noise, generating the bias voltage, and in some instances needing to capacitor couple between stages.

[merlinb]

Wanting more info on why some pedals use just a V+, Vref, and GND as the configuration where as others have V+, V-, Vref, and GND. 

Mostly it is a simple case of headroom- with a bipolar supply you have more of it! Eliminating a few coupling capacitors is also useful too, of course.

[goldstache]

So let me get this straight. 
It eliminates 1/2 of V+(Vref) being applied to amp inputs for biasing??

[R.G.]

So let me get this straight. 
It eliminates 1/2 of V+(Vref) being applied to amp inputs for biasing??

Well, kind of. If you have a single supply, V+ and ground, you must bias the circuits running from that power supply to someplace between ground and V+. Generally, dead in the middle is a good choice, especially for opamps. But different circuits may need different biases, depending on the details of the circuit. Discrete transistor circuits may have a bias voltage unique to each stage. For opamp circuits, it is often most convenient to make only one Vbias at half of the total power supply voltage.

But if you have a bipolar supply, where the total voltage between the most-positive and most-negative voltage supplies is the same as the V+-and-ground scenario above, then the power supply itself is making you a bias supply centered on the total power supply voltage. If your circuit can use that, then yes, it eliminates half of V+ being supplied to opamps for biasing. Your circuit may or may not be able to do that, but many opamp circuits will be able to.

This can save (1) the voltage divider/capacitor chain often used for a bias voltage generator and (2) in certain circuits, the capacitors between stages that we usually use to block DC voltages.

Mostly it is a simple case of headroom- with a bipolar supply you have more of it!

I believe this may have been intended to mean that you can get twice the voltage from a single transformer winding. This is true if one uses the half-wave doubling circuit. However, to the circuits using the power supply, there is no difference in headroom using V+ and ground and +/- half of that same V+ centered on ground, other than the elimination of a bias generator and some blocking caps. Their available power supply voltage is still the difference between the most-positive voltage and the most-negative voltage.

[goldstache]

Ahhhhhh! 
Thanks!  Starting to sink in!  Does anyone have a favorite voltage doubling circuit to get 18volts from a 9volt supply.  Maybe using charge pump or the like!!??
Really great of you all to take the time!

[Cliff Schecht]

R.G. briefly mentioned noise but in some circuits, this turns out to be a killer. Having a bipolar supply tends to be lower noise because you don't have to add a bias circuit that elevates the signal path to some reference. This bias supply adds noise into the signal chain (both passive and active reference circuits add noise!) and unless you are running a fully differential setup, that noise never goes away. Compare this to a bipolar supply that has some ripple associated with both rails (even after filtering). This ripple is typically small in amplitude and for a 60Hz power source, is low in frequency content. Any op amp will reject power supply noise as a function of the devices specified power supply rejection ratio (PSRR). It turns out that this is a frequency dependent term though and the PSRR goes down for higher frequencies. But for audio you can run a switching power supply at multiples of the highest frequency that you'll deal with for audio (say 1MHz) so that even if there is excessive ripple, proper limiting of the amplifier bandwidth will naturally squash any of this high frequency crud before it becomes a problem.

[Gurner]

The biggest advantage for me is you can be pretty certain that your 0V doesn't waiver (unless you've designed an awful grounding arrangment) whereas in high current single supply circuits, the rail can sag, which in turn sags the 1/2 VCC point which can have all sort of unanticipated/nasty outcomes (think oscillations, motor boating ...but not in a good,scenic 'life on the ocean waves' way!)

[Mark Hammer]

While +/-4.5VDC officially counts as as a bi-polar power supply, I have to say that I have never really seen one, largely because power supplies tend to correspond to the more common regulator voltages: 5, 6, 9, 12, 15, etc.

So, while the question focussed on bipolar vs single-ended supplies, it bears noting that most of the time when you see a power supply that either requires two batteries, or provides +/- DC voltages in some other manner, it is likely going to be +/-9v.  At that point, the "advantages of bipolar supplies" also includes advantages stemming from greater overall supply voltage, and the headroom that provides.

[puretube]

yes, +/- 1.5V doesn`t offer more "headroom" than +9V...

[amptramp]

Part of the attraction to positive and negative supplies is that the output of an op amp is derived from the ± supplies.  If the negative supply is at ground, the current through the ground that the amplifier sources is carried through the same ground impedance that the signal ground goes through.  With positive and negative supplies, the ground is never the source of the current although it may sink the current the amplifier puts out.