Balanced output example?

[Jopn]

Does anyone have a simple balanced output schematic?

With the popularity of preamp and speaker sim projects, I'm having a surprising time trying to find a basic schematic that I can apply to add a balanced output to an effect.

[Kipper4]

Did you mean a line level output?

[Fender3D]

Connect an audio trasformer like this http://www.mouser.com/ds/2/177/5c0047-48-75398.pdf model 560Q @ pedal output

[artifus]

http://www.jensen-transformers.com/apps_sc.html#musical

[Jopn]

I'm thinking of this for adding an XLR output to a pedal or preamp.

Is there not a simple op-amp way to split the signal, phase invert one side, then feed the two opposite signals to the XLR output? 

[Gurner]

Is there not a simple op-amp way to split the signal, phase invert one side, then feed the two opposite signals to the XLR output?  

http://www.seekic.com/circuit_diagram/Amplifier_Circuit/Amplifier_Circuits-Audio/BALANCED_OUTPUT_WITH_OPAMPS.html

(though you could use a better opamp, than the 1975 "everybody was kung fu fightin'" example in that schem). It's fbetter to use close tolerance resistors....else you can buy ICs that convert unbalanced to balanced for a low component count solution.

[Johan]

Look at the passive balanced outputs used by mackie mixers.just a resistor....gets the cmmr working but is very simple
J

[merlinb]

I'm having a surprising time trying to find a basic schematic that I can apply to add a balanced output to an effect.

Seriously? This is the first image if you Google "balanced output":

[Gurner]

Just bear in mind that example posted by merlin will be low input impedance (which oughtn't to be a problem here, as you probably wish to balance your signal at the end of your circuit)

[merlinb]

Just bear in mind that example posted by merlin will be low input impedance (which oughtn't to be a problem here, as you probably wish to balance your signal at the end of your circuit)

To make a high impedance version of the same thing, just move the input of the inverting opamp so that it is taken from the output of the noninverting opamp (same as Gurner's post, if a little less busy):

(Fifth image when you Google "balanced opamp output"...)

[Gurner]

To make a high impedance version of the same thing, just move the input of the inverting opamp so that it is taken from the output of the noninverting opamp (same as Gurner's post, if a little less busy):

I'm just a bit too lazy to tidy up a grabbed google image!

Jopn - bear in mind the above diagram is a bi-polar supply solution (whereas most folks will be on single supplies)...to make it single supply, put the bottom of that 470k resistor to  1/2 VCC & also the +ve input of the lower opamp to 1/2 VCC  too.

[Jopn]

Sorry for the frustration if it appears I didn't try to research this on my own.  I certainly came looked into conceptual diagrams like the first one that merlinb posted, but in attempting to learn I was hoping to find a "and this is how it's often implemented in the DIY stompbox world...".  Then I could take the theory of the conceptual diagram, compare it a practical application example, and hopefully start to learn better how to make such transitions myself.

Anyhow, I'm not looking for an argument, I just felt the implication was that I'm lazy and I wanted to take a moment to explain myself.  Enough drama, let's talk circuits....

So merlinb's second diagram posted looks nice and logical to me.  But just to clarify, it's expecting that I'm feeding -9v to one op amp, and +9 to the other, correct?

"put the bottom of that 470k resistor to  1/2 VCC & also the +ve input of the lower opamp to 1/2 VCC  too"
- I'm not yet familiar with what the 1/2 VCC refers to, but I can figure that out (obviously has to do with the VCC+ and VCC- pins, I just have some reading to do...)

Thanks guys!

John

[Gurner]

"put the bottom of that 470k resistor to  1/2 VCC & also the +ve input of the lower opamp to 1/2 VCC  too"
- I'm not yet familiar with what the 1/2 VCC refers to, but I can figure that out (obviously has to do with the VCC+ and VCC- pins, I just have some reading to do...)

1/2 VCC just means half the supply voltage (therefore if its a 9V supply, you need a '4.5V voltage reference' to connect both those points I mentioned to...sometimes you'll see this refereed to as bias, Vb, Vref, 1/2 VCC etc)...there are several ways to achieve this ranging from the most basic (two resistors of equal value connected between 9V &  0V)....to slightly more involved (voltage followers, regulators etc)

[Jopn]

1/2 VCC just means half the supply voltage (therefore if its a 9V supply, you need a '4.5V voltage reference' to connect both those points I mentioned to...sometimes you'll see this refereed to as bias, Vb, Vref, 1/2 VCC etc)...there are several ways to achieve this ranging from the most basic (two resistors of equal value connected between 9V &  0V)....to slightly more involved (voltage followers, regulators etc)

Ah, perfect!  Thanks!

[rankot]

I know it's an old thread, but info here is still valid, and I wanted to add some more. I've been recently fooling around with balanced outputs, and came across this one, as shown on ESP: https://sound-au.com/project51.htm



It is quite simple, and simulation shows a slight disbalance, which can be cured if R5 is 110 Ohm. I also thought to use 10k instead of 33k, to lower the noise.

But I have another idea, maybe simpler (but not necessarily better) - to use JFET phase inverter for this. Is it acceptable, or not?

[ElectricDruid]

That's a nice clear one, Ranko. Again, it's a bipolar power supply example, so we have to add Vref supply and biasing.

There's a load of stuff about balanced ins/outs on the That Corp site, but unsurprisingly, it all uses their extremely high quality line driver / receiver chips. Still, worth a read.

[PRR]

> Is it acceptable, or not?

I say "not".

First, the DC bias is wrong. By JFET follower action the top of R23 will be say +1V, so 16V across R23. R21 drops the same, from +15V to -1V. The voltage across the JFET is reversed. Yes, you can change the gate bias and "fix" this.

Looking at the big picture: if you "need balanced" it is to overcome hum/buzz/crosstalk in the cable. You want BIG signal on the line. The cathodyne can only swing half voltage on each side. And the 100k value you chose is not much current; values far under 1k are needed to suck-up induced line noise.

> simulation shows a slight disbalance, which can be cured if R5 is 110 Ohm. I also thought to use 10k instead of 33k, to lower the noise.

The result should be exact as shown in the ESP plan. You should beat your simulator until you find the error. Hiss in a line driver should always be far below the hiss of your source (mike, pup, tape) because of all the gain between.

[rankot]

> Is it acceptable, or not?

I say "not".

First, the DC bias is wrong. By JFET follower action the top of R23 will be say +1V, so 16V across R23. R21 drops the same, from +15V to -1V. The voltage across the JFET is reversed. Yes, you can change the gate bias and "fix" this.

Looking at the big picture: if you "need balanced" it is to overcome hum/buzz/crosstalk in the cable. You want BIG signal on the line. The cathodyne can only swing half voltage on each side. And the 100k value you chose is not much current; values far under 1k are needed to suck-up induced line noise.

> simulation shows a slight disbalance, which can be cured if R5 is 110 Ohm. I also thought to use 10k instead of 33k, to lower the noise.

The result should be exact as shown in the ESP plan. You should beat your simulator until you find the error. Hiss in a line driver should always be far below the hiss of your source (mike, pup, tape) because of all the gain between.

OK, I will stick to ESP, then.