How to make this circuit unity gain?

[AFex54]

https://i.imgur.com/2g5Hoeg.png

I built this amp but the gain is unncessary I just need it to be a buffer between the headphones and source to supply the power the headphones need.
I think the way to achieve this is to remove the caps and resistors on the op amp's inverting inputs but will this have an effect on the amp other than the reducing the gain 1 (like alter the frequency response)? is there any other changes that should be made to avoid negatively affecting the signal i.e sound quality ?

 

[WHONOES]

Removing R2 and C2 (and the corresponding parts from the other half of the circuit) will yield a gain of unity. Have looked at the datasheet for the NE5534 and there is no indication that it may become unstable in this configuration. A simple way to check if this is so is to read the datasheet and if it mentions a minimum gain for stability (often 5 or 10 times) then you may have issues with a unity gain circuit. It will have an effect on the bandwidth of the circuit. It will be increase by the gain factor you have removed. This is determined by the GBW (Gain Bandwidth Product). Divide this by the gain set and it will give you the likely bandwidth you may achieve although this will most likely be affected by other factors such as layout.
It is possible to improve on the basic configuration you have shown with the addition of one more transistor and 2 diodes for each half of circuit.

 

[Audioguru]

The datasheet from Philips shows a 22pf capacitor between pin 5 and pin 8 when the gain is 1.
This class-A amplifier will have enough output power to blow up you headphones and ruin your hearing if the input level is too high.

 

[WHONOES]

Ok. The Texas Instruments datasheet that I saw didn't show the capacitor. An improved version could give a little better linearity. Having said that, the average person probably wouldn't discern the difference.

 

[Audioguru]

Philips bought Signetics and took over their NExxxx ICs.
The NE5532 is compensated for a gain down to 1. Philips and Texas Instruments say the NE5534 is compensated for a gain of 3 or more. They both show a 22pF external compensation capacitor for low gains.
The compensation is frequency compensation (to prevent oscillation) not for linearity.

 

[WHONOES]

The datasheet that I referred to is NE5532 specific and makes no mention of an external compensating capacitor requirement. It states that it is internally compensated for unity gain. Clearly other manufacturers products differ.
See link to datasheet : http://www.ti.com/lit/ds/symlink/ne5532.pdf

 

[AFex54]

Removing R2 and C2 (and the corresponding parts from the other half of the circuit) will yield a gain of unity. Have looked at the datasheet for the NE5534 and there is no indication that it may become unstable in this configuration. A simple way to check if this is so is to read the datasheet and if it mentions a minimum gain for stability (often 5 or 10 times) then you may have issues with a unity gain circuit. It will have an effect on the bandwidth of the circuit. It will be increase by the gain factor you have removed. This is determined by the GBW (Gain Bandwidth Product). Divide this by the gain set and it will give you the likely bandwidth you may achieve although this will most likely be affected by other factors such as layout.
It is possible to improve on the basic configuration you have shown with the addition of one more transistor and 2 diodes for each half of circuit.

thanks, it worked. I read that Non-inverting amps cant have a gain lower than 2 but this seems to contradict that.

after removing R3 feedback resistor aswell it seems to sound better.

how would you implement the extra transistors and diodes into the circuit?

 

[WHONOES]

I'll put together a schematic and post it in the near future. Your new configuration with the extraneous components removed is a voltage follower and has unity gain.

 

[WHONOES]

OK. See attachments for schematic as promised. The amplifier on my schematic is a TL072. I only used it as I don't have a model for the NE5532.

 

[Audioguru]

First you said "NE5534" and I corrected you. Then you said the correct NE5532 IC that does not have pins for an external capacitor for each channel so it has the capacitors inside it.

Your amplifier has nice low distortion.

 

[AFex54]

OK. See attachments for schematic as promised. The amplifier on my schematic is a TL072. I only used it as I don't have a model for the NE5532.

this helps a lot. how necessary are those resistors? I already have the bd140 and diodes on hand but would need to place an order for the resistors

 

[WHONOES]

Hi AG. You are right, I did inadvertently say NE5534 when I meant NE5532 as in the circuit under discussion. I am fully aware that the 22pF on the 5534 is for stability, the linearity mentioned was not referring to the opamp but the configuration following it. I did not actually say that the 5532 does not have pins for adding a compensating capacitor only that there was no requirement for it on the datasheet. I will leave that there I think.

The 2 emitter resistors in my circuit effectively control the quiescent current (QI) in the output stage. They are not absolute and could be in the range of 10 to 22R. As shown the QI is about 7mA but is dependent on variations in the transistor Vbe's and the diode forward voltages.
The circuit is shown with a 33R load. If you wish to use an 8R load you will need to reduce the values of R6 and R7 to 2K2. You may need to put a small heatsink on the BD139 and BD140. If you find that the amplifier is suffering from thermal runaway then glue the two diodes to the heatsink in between the 2 transistors. But I would suggest that you try it in the simplest format first and only resort the heatsink etc if things start getting a bit warm.
There is always the option of a FET output stage but I have not worked on that yet.

 

[Audioguru]

The drive to the joining of the diodes is a way that was taught in school that I do not agree with. The opamp does not turn on each output transistor, instead the opamp turns off each output transistor. The 4.7k resistors turn on the output transistors and when they need more base current then the 4.7k resistor gives them less current.

 

[AnalogKid]

thanks, it worked. I read that Non-inverting amps cant have a gain lower than 2

That is true for a conventional feedback network with both series and shunt elements. Without the shunt element the circuit becomes a voltage follower; non-inverting and unity gain, sorta like an emitter follower but without the DC shift.

ak

 

[WHONOES]

Yeah. I could have made it a lot more complicated but, the brief (sort of) was for a simple circuit with reasonably good performance which is what it is. It is only required to deliver 100mW or so, any more would be damaging.

 

[AFex54]

thanks for the help, looking forward to see how this affects the sound.

 

[AFex54]

This helped tremendously, the amp has more depth, weight and speed to its sound.
would there be any way to build even further on this design, maybe adding additional transistors?
this may seem excessive but its driving a particularly power hungry planar magnetic headphone, they require much more power (current?) than traditional dynamic headphones.
I also moved up to 15v rails and added additional capacitors to the power supply, surprisingly the capacitors made the biggest difference of all, the amp finally sounds like it can 'keep up' with the music, mainly in the bass.

from looking at pictures of their internals most powerful high end headphone amps use discreet designs , is the op amp going to limit any improvements past here?

 

[AnalogKid]

surprisingly the capacitors made the biggest difference of all,

Really decent audio needs high transient currents, but the transformer secondary's inductance limits how fast it can supply current to the filter capacitors. Big ass capacitors store a lot more energy, and so they discharge less on those transients, presenting a lower dynamic impedance voltage source to the circuits. Putting big caps right up against the output transistors is even better.

ak

 

[(*steve*)]

Putting big caps right up against the output transistors is even better.

A note for @AFex54, this means electrically, not physically

 

[Audioguru]

When a few hundred Watts amplifier is put in a car, some people add a huge capacitor since the battery can supply only (!) 600A.