Headphone amp, Turn on pop [solved]

[Michaelaudio]

Oh yes, right... that doesnt work then.
2 Batteries wont fit in my enxlosure sadly.

Bluetooth speaker always using a Boost converter, that doesnt make a lot of noise. I used this one: https://www.ebay.de/itm/SX1308-Step-Up-Spannungswandler-DC-Booster-Schaltregler-Arduino-Raspberry-Pi-/253381738734?mkcid=16&mkevt=1&_trksid=p2349624.m46890.l49286&mkrid=707-127634-2357-0

When No music is playing, the noise is fucking loud, it disappears when i Play music thou. I have to find a Regulator thats AS small as this, with a better quality. (Or somehow modify that one)

Will it be safe when i relie on ur circuit you posted before? Or will a relay be unavoidable? If so, i might get it to work that it will turn Off, before the Rest of the circuit turns on.

[magic]

Not cancel, but add to create a complete, attenuated, inverted image of load current. Only in a dual supply situation, only if both rails have equal impedance (including reactance).

Cancellation would occur in case of cross-conduction in the output stage.

In singe supply configuration, you first draw current from the battery, pulling the rail down, and you charge the load coupling capacitor as you drive the load. Then you pull the capacitor to ground, using the stored charge to drive the load negative. Battery current is zero in the second phase, so rail voltage is unaffected. Hence the rail carries halfwave rectified signal.

BTW, TL072 has much worse PSRR from the negative rail, so arguably that's where ground should be. Fun

What I mean by cancel is, each rail will have the half wave rectified signal superimposed on it. Par of the signal will be fed back to the input, but it will be inverted, as the output will pull the respective rail down slightly, so all it will do is reduce the total gain slightly, due to negative feedback.

Yes, what you mean by "cancel" is what I mean by "add and complement". I happen to believe that one of those descriptions is more true to fact, but at least we understand each other

Now, the rest of the post you quoted argues why this doesn't happen in single supply circuits and even with dual supplies it's hardly perfect.

[Michaelaudio]

So for the relay solution, this should do the trick.

[Zero999]

The lack of power supply voltage is more of a problem.

Why not use an op-amp with a high current rail-to-rail output such as the MC33202?
https://docs.rs-online.com/c975/0900766b816ded01.pdf

[Michaelaudio]

Even when i have the full +/- 4,5Volts, i only have around 3V Rms, but i need about 5V Rms. I guess i wont come around the boost converter.

[Zero999]

Even when i have the full +/- 4,5Volts, i only have around 3V Rms, but i need about 5V Rms. I guess i wont come around the boost converter.

Another option is to add some transformers to the output, but they'd probably be too big.

A boost converter will work, but it needs to be properly filtered, to eliminate the noise.

[Michaelaudio]

I ordered two more, to see which performs best. I tried a big cap with my current one, but that only changed the noise from high, to a low hissing. Maybe ill be more lucky with a Pi filter.

[magic]

I dont exactly know what u mean with "Filter bias to ground"

The supply rail is noisy. If you simply divide it in half to derive the bias voltage, the bias voltage is 50% (-6dB) as noisy as the supply.
To remove the noise, you filter the bias voltage with a cap to ground. Right at the resistors, before the opamp, because a cap after the opamp has little effect - the opamp will simply charge/discharge the cap in order to transfer the noise from the resistor divider to the output.

IOW, remove C17, move C18 to the input side of the opamp.

But this means that turn-on/turn-off transients will get worse.
Relay protection makes transients kinda irrelevant, if it works reliably.

edit
Generally, every capacitor between the signal path and the supply rail (not ground) is an invitation for noise. This includes load AC coupling capacitors too. OTOH, symmetric caps to both rails reduce turn-on transients. Pick your poison.

[Michaelaudio]

Ahh, now i got it. Since im gonna usw the relay method now, with the turn in delay for the Rest of the circuit, so i dont worry about more transient.

I'll See what i can do with my Regulator, i found some nice Sheets from Ti/AD for Output filtering, i'll play around with it.

[Zero999]

I dont exactly know what u mean with "Filter bias to ground"

The supply rail is noisy. If you simply divide it in half to derive the bias voltage, the bias voltage is 50% (-6dB) as noisy as the supply.
To remove the noise, you filter the bias voltage with a cap to ground. Right at the resistors, before the opamp, because a cap after the opamp has little effect - the opamp will simply charge/discharge the cap in order to transfer the noise from the resistor divider to the output.

That's only true if the signal is referenced to one of the power supply rails. Apply the signal to the mid point and the two capacitor solution is just as good.

Here's an example showing an RC rail splitter and the effect of the loading on the power supply rails. All it does is reduce the gain a little. If the power supply had a bit of noise, it would be no worse, than referencing the signal to either rail and bypassing the bias point to the rail, the signal is referenced to.

[magic]

Yeah, that's essentially the CMoy amplifier. You replaced all the biasing drama with a virtual ground. Note a complete absence of capacitors in the signal path.

Which is perhaps an acceptable option, but I was mentally stuck in the realm of single supply AC coupled circuits, as suggested by the OP.

This circuit has DC coupled signal path, so potential disadvantages:
- DC offset of the source is passed to the load
- DC offset of the opamp too, at 9x gain
- in presence of DC offset, supply rails may get asymmetric due to high resistance of R3,R4
- if the two channels fail short to opposite rails, DC fault current flows through the two transducers in series

OTOH, the AC coupling of supply rails to ground saves transducers when only one channel faults or both fault to the same rail, which is surely a more likely scenario.

DC offset problems could be reduced by AC coupling between stages, like O2. Only a pair of small caps is required. This means the gain stage no longer gets to support the output stage with additional load current.

edit
Another (unrelated) disadvantage is less efficient use of the opamp's PSRR, compared to grounding the more sensitive supply.
There is also an advantage: any distortion of load coupling capacitor is removed from the signal path.

OP will have to decide which topology to choose. Last time I remember the topic of simple headphone amps, the OP was hell bent on AC coupling the load for the sake of safety/reliability and rejected all DC coupled split rail / dual rail suggestions.

[Michaelaudio]

@Zero999 isnt Ur Last schematic more or less the same Voltage biasing like i did it? I did the same thing, but with an OP amp, for a low impedance source.

I Just looked at this vid again, and i think i should be fine now: https://youtu.be/Z2GUoi63pJs

The relay should now turn on, before the rest of the circuit, and only Turn on for ~ 3 Seconds, so battery only supplies amps to the OP amps -> improving battery life a lot.

Also, i did some filtering on my converter. With a simple added RC Filter, i went from a 200mVpp noise, to 25mVpp. Thats good so far, but still audible. Ill mess around with it more, but im getting there.

[Zero999]

@Zero999 isnt Ur Last schematic more or less the same Voltage biasing like i did it? I did the same thing, but with an OP amp, for a low impedance source.

My schematic is completely different. The 0V point is at half the supply voltage, providing a +/-4.5V power supply. There won't be a turn on pop because the both the capacitors charge up at the same time and the output of the op-amp follows them.

I Just looked at this vid again, and i think i should be fine now: https://youtu.be/Z2GUoi63pJs

I haven't watched it in its entirety, just flicked through the schematics. It's much more complicated than necessary.

The relay should now turn on, before the rest of the circuit, and only Turn on for ~ 3 Seconds, so battery only supplies amps to the OP amps -> improving battery life a lot.

Also, i did some filtering on my converter. With a simple added RC Filter, i went from a 200mVpp noise, to 25mVpp. Thats good so far, but still audible. Ill mess around with it more, but im getting there.

That will work, but why is it so complicated?

[Michaelaudio]

My schematic is completely different. The 0V point is at half the supply voltage, providing a +/-4.5V power supply. There won't be a turn on pop because the both the capacitors charge up at the same time and the output of the op-amp follows them.

U have a 9V DC on the Input, Splitting it in half with the resistors. So in the middle u have 4,5 Volts, thats what u define as ur Ground. Ur "V-" is actuall ground, 4,5 Volts is ur virtual ground, setting an Offset to ur AC Signal. Or is it a real +/- 4,5 Volts supply, with the wrong Symbol at its Input?

How do i make it simplier? As u said, the relay may be too slow, so it has to turn on first.

[Zero999]

My schematic is completely different. The 0V point is at half the supply voltage, providing a +/-4.5V power supply. There won't be a turn on pop because the both the capacitors charge up at the same time and the output of the op-amp follows them.

U have a 9V DC on the Input, Splitting it in half with the resistors. So in the middle u have 4,5 Volts, thats what u define as ur Ground. Ur "V-" is actuall ground, 4,5 Volts is ur virtual ground, setting an Offset to ur AC Signal. Or is it a real +/- 4,5 Volts supply, with the wrong Symbol at its Input?

Ground just means the point from where all voltages are referenced to. The schematic I posted uses a rail splitter, or virtual ground.

How do i make it simplier? As u said, the relay may be too slow, so it has to turn on first.

If you a rail splitter, as I suggested, then you don't need a relay.

[Michaelaudio]

Yes, but im doing the same, it doesnt matter If there is an OP amp Afterwards. Also ud still have the Cap charging Via 9V to 4,5V, which would cause a Pop on the output.

[Zero999]

Yes, but im doing the same, it doesnt matter If there is an OP amp Afterwards. Also ud still have the Cap charging Via 9V to 4,5V, which would cause a Pop on the output.

You're using a single supply non-inverting amplifier topology, which uses the negative side of the power supply as the ground. The circuit I've shown, uses half the power supply voltage as the ground.

Here's a couple of schematics showing the difference between a single and split supply.



In this case, the TLE2426 can be used, that way the large capacitors are no longer needed.
https://www.ti.com/lit/ds/symlink/tle2426.pdf

[Michaelaudio]

I cant see how the "Split supply" circuit could work. If ud apply a signal to it, it would cut the sinewafe in half. The -6v still isnt -6V, its ground, therefore If i feed a Signal to the Input, without biasing it, it cant work.

[Zero999]

I cant see how the "Split supply" circuit could work. If ud apply a signal to it, it would cut the sinewafe in half.

The split power supply circuit works fine. In this case, the signal is AC, so the biasing resistors only carry a tiny circuit. The capacitors need to be large enough to pass the AC signal, but similar sized capacitors would be needed for the signal supply circuit.

The -6v still isnt -6V, its ground, therefore If i feed a Signal to the Input, without biasing it, it cant work.

Ground is just a reference point, from where all voltages are common to. It doesn't have to be the negative side of the supply. It can be the positive. Here's an example showing two 9V batteries in series, with the ground reference, in the middle, on the negative and on the positive.

[Michaelaudio]

I know, its just the definition, how u define ground. But if u now connect lets say ur phone to the input, ur input ground of the circuit is at 4,5Volts, while ur Phone ground is at 0V. So u biased the ground, Not the signal path so to speak. Isnt that a Problem?

[magic]

TLE2426 can be used, that way the large capacitors are no longer needed.

Another advantage: higher peak output at low frequencies.
A disadvantage: load driving efficiency decreases by 50%.

I know, its just the definition, how u define ground. But if u now connect lets say ur phone to the input, ur input ground of the circuit is at 4,5Volts, while ur Phone ground is at 0V. So u biased the ground, Not the signal path so to speak. Isnt that a Problem?

Only a problem if there is some external connection between the phone's ground and the negative of the PSU, for example through mains-powered chargers with three prong plugs.

[Michaelaudio]

I'll try that out in my current Board, If the pop is gone then, i can just use that board how it is, and focus on the psu noise.

[Zero999]

TLE2426 can be used, that way the large capacitors are no longer needed.

Another advantage: higher peak output at low frequencies.
A disadvantage: load driving efficiency decreases by 50%.

Good point. The capacitors are charged in series and discharged individually, like a charge pump, which halves the voltage, rather than doubling it.

I know, its just the definition, how u define ground. But if u now connect lets say ur phone to the input, ur input ground of the circuit is at 4,5Volts, while ur Phone ground is at 0V. So u biased the ground, Not the signal path so to speak. Isnt that a Problem?

A circuit can have different grounds, which isn't a problem if they're not connected. The mains has protective earth/ground and neutral serves as the 0V reference for the circuit. Another example is a device connected to the mains via an isolated switched mode power supply: the 0V ground on the secondary, is completely different to the mains neutral.

[Michaelaudio]

So ive tried it Out and it works, i still have a slightly pop tho. I dont know If its bc of my quick and dirty test, or if there will be Always a slightly pop. If so, the question is, If i will have this with the relay solution too.

[Zero999]

So ive tried it Out and it works, i still have a slightly pop tho. I dont know If its bc of my quick and dirty test, or if there will be Always a slightly pop. If so, the question is, If i will have this with the relay solution too.

Did you try matching the capacitor values?

It's very difficult to eliminate the pop. I've done this before, using the relay method and there was still a tiny pop.

The question is: is the pop acceptable? I'd say if it's not uncomfortably loud, it's good enough.