Preventing ringing in audio mute circuit

[Psi]

I've been trying to solve this for a few days and not getting anywhere so i thought id ask here.

I have a MCU controlling two opto switches, one passes or blocks the audio the other mutes the output to ground. The mcu is clocked by internal RC at 1mhz.
The issue i'm having is that the mute opto always produces a pulse of super fast output ringing when turned on (~60mhz then ~10mhz).

The signal is instrumentation level audio and this burst of ringing produces an audible 'click/pop' when transiting into the mute state.

Does anyone have any good ideas how to stop or suppress this ringing?
I've tried disabling the MCU and driving the optos manually and i still get the problem, so i don't think its a digital noise issue.

Any help appreciated.



Yellow is AC audio and blue is DC opto drive from mcu. (Logic is inverted, so mcu low = opto on = mute)
This is a silent signal being muted.

[hamster_nz]

Do the inputs have a dc offset?

Sharply switching off a signal with a DC offset will give a transient.

Maybe a coupling cap on the input to block DC?

[DrGeoff]

What kind of optos are these?

[Psi]

Opto is this one
https://nz.mouser.com/datasheet/2/408/TLP222A_datasheet_en_20170710-1134823.pdf

I dont see any DC offset, if there was one i would expect to see that on the scope view above.
It would show as a different level before and after the transition into mute state.
hm.. then again, i do have scope in AC mode so it could easily be masking this.
Thanks i will check this more thoroughly tomorrow and post results.

[kg4arn]

I don’t have a RIGOL scope but it looks to me like CH1 is AC coupled based on the yellow symbol at the bottom left. This will block the DC level.


***. Oops. I Didn’t read carefully *****

[Psi]

yes, i realized this as i was writing the post and edited it.
It probably didn't say that when you first read it.

[ciccio]

I see from the datasheet that the opto's outputs are mosfets, not resistors as shown in your schematic.
Did  you try a really resistive opto, such as a Vactrol? They are defunct, but Silonex has the NSL-32 that is still available from Farnell (it was sold also by LUNA optoelectronics), and there are even clones using the old Vactrol's name: VTL5C3 or similar available on ALIexpress.
The problem is maybe that all products using LDRs are becoming obsolete for some environmental reason, so maybe they must be avoided for new designs.
Why not use a standard opto driving a "mute" transistor (2SC2878 , obsolete but still available, or similar)?

Best regards

[aldi]

I found this, a similar problem, do you have a DC- blocking capacitor somewhere?
https://electronics.stackexchange.com/questions/325777/audio-soft-mute-using-optically-isolated-mosfet

[Psi]

Thanks, some good info there.

[Kalvin]

You may also try adding 10 nF - 100 nF capacitors across the LED1 and LED2, which will effectively slow things down a bit.

[dmills]

Charge injection?

Place a honking great cap across each of the the LEDs so it fades up and down rather then snapping on or off, and switch LED 2 on before switching led 1 off and vice versa.

Regards, Dan.

[David Hess]

It is not clear what is happening so the first thing I would try is AC coupling the optocoupler output and placing a resistor across the optocoupler output so that the capacitor remains charged to the average DC voltage.

[Audioguru]

I made background music circuits for speakerphones. To prevent switching the music off at the peak voltage of the signal (causing a loud POP) I slowly ramped on the Jfet used for muting.
I don't know if an optocoupler can slowly ramp on. An old fashioned LDR turns on fairly slowly and might work better.

[cvanc]

Just trigger one switch at a time to see if it's only one opto that's causing the issue?

And yeah, maybe stagger the trigger so the shunt element turns on first.

[Psi]

I had time today to look into the issue a bit more.
The DC offset idea is defiantly part of the problem, just need to figure out why there's DC where they should not be.

Just trigger one switch at a time to see if it's only one opto that's causing the issue?

And yeah, maybe stagger the trigger so the shunt element turns on first.

It's the Mute opto that's causing the pop

I made background music circuits for speakerphones. To prevent switching the music off at the peak voltage of the signal (causing a loud POP) I slowly ramped on the Jfet used for muting.
I don't know if an optocoupler can slowly ramp on. An old fashioned LDR turns on fairly slowly and might work better.

Yeah, i may have to do that.
Even with the DC removed i still get a small pop if changing between mute and a peak in the audio.

[BrianHG]

DC offset in your circuit, if it's not from your source signal, may be from one mosfet in the opto turning on or off prior to the second, or one of each protection diode is switching with an initial delay.

You are using the wrong type of opto for audio.  Use this guy: H11F3M
http://www.onsemi.com/pub/Collateral/H11F3M-D.pdf

looking at page 7 on the datasheet, you see that this photo-fet is designed for audio attenuation applications unlike opto-mosfets like which you are currently using.

Also, low pass filter the driven LEDs as this optocoupler has a smooth linear resistive output transfer VS the LED drive current unlike the optos you are using which harshly avalanche from megaohms right down to a few ohms without any linear transition.  This should act like a fast moving volume control if you tune the right 2 resistors and cap at the LEDs so that a pop will never be generated, though, muting the audio will take around 1/20th of a second.

Also, with the over 300M-Ohm resistance in the H11F3M, you should only need 1 series optocoupler with an output load as your mute switch.

[ciccio]

http://www.onsemi.com/pub/Collateral/H11F3M-D.pdf[/url]

Thanks for this info: I did not know that device, and I'll order some samples to-day for a project I'm working on.
Best regards

[duak]

I wonder if the reverse or body diode across the FET is acting as a photodiode and is injecting some voltage into the circuit.  The pulse is about the right amplitude.  I think the FET is diverting some of the photccurrent, but i there will be a time delay before the FET turns fully on. The remainder of the damped sinusoid could be due to the cabling ringing.  This could be confirmed by observing the voltage across the FET out of the circuit.

[cvanc]

http://www.onsemi.com/pub/Collateral/H11F3M-D.pdf[/url]

Thanks for this info: I did not know that device, and I'll order some samples to-day for a project I'm working on.
Best regards

Yeah, I also did not know this device existed.  Looks very useful.  Thanks BrianHG!