Problem with volume control noise in a headphone amplifier

[StanleyB]

ne5534 compensation

I designed a little headphone amp that has a a 50K volume control, which is connected in between two electrolytic capacitors. When I adjust the volume control I get a crackling noise. It is not the pot. I tried several types. I am using NE5534, and I adjusted the offset bias, and made the gain larger than 3. I rerouted the ground as well.
Anyone got any ideas I have overlooked? I am desperate to get this thing working properly after about 3 weeks of fighting with it hours on end. HELP!!

 

[IanP]

Re: Volume control noise

Try to add compensation cap Cc between pins 5 and 8 ..
Something from 22pF to 47pF should do ..

Regards,
IanP

 

[banjo]

Volume control noise

Hi,

My opinion is that C200 is not needed. Your audio source should be able to withstand a DC load of 10K to 50K.

Second, I think C204 is much larger than it needs to be. My quick calculation is that your high pass cut-off is C204 & R204+R203. This gives a corner frequency of 2.6HZ. Since it is a headphone amp and since lower range of hearing is 20HZ, I would suggest reducing this capacitor to 2.2uF.

I think your static on adjusting the volume control is due to the charging currents for C204. This current is injected right into the inverting input of the amp which then get amplified and ends up as noise in the headphones.

Another thing you could try is to remove C204, leave C200 and terminate the lower side of the volume control to bias network on C205. Charging currents for the capacitor would end up in the bias network instead of in the input terminal of the amp.
With the volume control not directly tied to ground, you do not have to worry about DC isolating the two inputs of the amp.

 

[StanleyB]

Re: Volume control noise

Thank you guys. I tried the cap between pin 5 and 8 before, but forgot to add it to my circuit for clarity. It didn't work by the way.
The idea of connecting the center contact of the volume control to C205 is very interesting and I am going to try it out today. I can't remove C200 in this application. The signal comes from the output of an IC that specifies a 10uF cap at its output due to the offset voltage from that IC.

I shall report back on the outcome.

 

[banjo]

Volume control noise

Hi Stanley,

I was probably not clear enough. I was suggesting connecting the lower or ground terminal of the volume control to C205. The idea is to feed the bias voltage back up the volume control to the inverting pin and thus remove the need for the C204 isolation capacitor.
This amplifier can be thought of in terms of an op-amp. Op-amps with negative feedback have a property called a "virtual ground" between the two input terminals. What this means is that the op-amp will adjust itself at all times until the voltage at the inverting input exactly matches in voltage at the non-inverting input. With a DC bias on the non-inverting input, the op-amp will require this same voltage on the inverting input. Hopefully, by feeding this voltage back through the volume control, the coupling capacitor can be removed.

C205 provides an AC ground for the audio frequencies, so to the incoming audio, then volume control still appears to be connected to ground.

 

[nguyennam]

Re: Volume control noise

Your way of using NE5534 is a bit strange for me.
You can consult Application note for it from ON semi (https://www.onsemi.com/pub/Collateral/AND8177-D.PDF), you will see that it is connected in a very common way.

In general, if the crackling noise appears only when you adjust the volume so the reason must be from the volume potentiometer. Dust from worn surface of potentiometer is what you need to wash out.

You also may use a 0.1uF capacitor between volume pointer and ground to eliminate the noise in case of very little of dust.

nguyennam

 

[StanleyB]

I found out that I needed to connect a 120K resistor to the right side of the cap. This discharged the DC potential on the cap, and solved the noise issue. The circuit has been used in a commercial product of mine of which more than 3000 pieces have been sold since I first asked the question.