De-pop circuit for speaker when powering on

[DOCa Cola]

I have started a little arduino project that includes an lm386 audio amplifier and a small pc-speaker to play some sounds. This all works fine. What i wondered about though is how i can prevent that pop noise that the speaker emits when applying power to the circuit for the first time.
This is also a typical issue i know from cheap radios when you turn them on or even various PCs where you hear their speaker pop when you press the power button.

I haven't really found anything in the internet that directly approaches this issue for this "small scale" application.
I have found a patent that seems to has a circuit using diodes. http://www.google.com/patents/US20060245602
I am still quite the beginner and the description for the circuits in this patent are not really easy to understand.

Can anyone give me advise on how to realize this? Again, this is not meant for a super surround high definition audio application, but just a small beeping pc speaker.

[rob77]

it's usually done by a relay which connects the speaker with some delay after power-on.

[jlmoon]

Single transistor driving relay coil (contacts breaking speaker connection) with transistor having RC time constant for a delayed bias or turn on threshold.  Don't forget to use a diode across the relay coil to suppress the reverse voltages back into the collector of the transistor when it goes into cutoff mode

an example

[jakeypoo]

You haven't posted a schematic, but I'll say it's caused by the charging of  the capacitors on the inputs of your OP AMP. It can be minimized by using the smallest possible caps there, but remember your input impedance and input capacitors form a high pass circuit so dont cut off anything above 60-100hz.

Similar to what Rob mentioned, the easiest solution is to not enable the output until after your input caps are charged.

Try replacing the LM358 with a equivalent audio amp that has a delay in the output turning on. Usually TI will call it something like "depop technology" basically it waits some time after power is applied before unmuting the outputs.

[DOCa Cola]

ah, right. i'll attach the schematic. I am using the basic LM386 20x gain configuration right now. I thought about using a relay, too, however with that i would get that clicking noise from the relay instead of the pop. Was the patent info i posted above useless? It at least looks to my unexperienced eyes not too complex to implement.

Usually TI will call it something like "depop technology" basically it waits some time after power is applied before unmuting the outputs.

Yes, thanks for the suggestion i could take a look at what TI has to offer in case we can't figure something else out here If someone got experience on what i could use for this project...

[jakeypoo]

Going to need more info than that. At least a full schematic of what's connected to the audio inputs. Are you connecting a cap to the bypass pin?

[planet12]

One source of "pop" will be the output 250uF (or whatever you've actually used) DC decoupling capacitor charging to 1/2 of Vs during switch-on. The LM386, being designed for single-supply operation, biases itself internally to 1/2 of the supply voltage.

Removing this the naive way by simply putting a relay between capacitor and speaker will just delay this charging until the relay switches on - you'll need to find another way to pre-charge the capacitor before connecting the speaker, such as a resistor a good deal higher than your speakers impedance between capacitor output and ground, then relay between that and the speaker, and just make the time constant before the relay switches on longer than the time it'll take to charge the cap.

A "better" way would be to use a dual-rail amplifier of a type that doesn't pop, but this may not be practical for your design.

[DOCa Cola]

some info regarding my setup
on the audio input side i've got a potentiometer (i think this is set to about 400 ohm currently) coming from the arduino output
i am using a 100µF cap to the speaker output
on the dc input pins i also got a decoupling capacitor 470µF.
on the gain pins i have currently nothing connected (20 gain). However i've got a switch to set the lm386 to 200 gain with a 10µF cap.
The bypass pin 7 has a 10µF cap to ground.

[mij59]

Hi,

You could try using a n-channel mosfet.

Disconnect the speaker lead from gnd.
Connect the speaker lead to the drain of the mosfet
Connect the source of the mosfet to gnd
Connect a capacitor (10 uF) between gate and gnd
Connect a resistor (470 k) between gate and pos. rail ( assuming it's 12V or so)

[DOCa Cola]

yes, the circuit is 5v based. i'll get a n-channel mosfet (don't have one at hand currently) and will try that approach.

[mij59]

yes, the circuit is 5v based. i'll get a n-channel mosfet (don't have one at hand currently) and will try that approach.

If the supply voltage is 5V you'll need a logic level mosfet. 

[DOCa Cola]

oh, ok. i wasn't aware of that. makes sense.
learned a lot from that thread already