Analog switch IC that can be made smooth switching?

[Tubis]

I tried switching the audio with an 4066. But it switched abruptly. And if I put a resistor/capacitor on the trigger do delay it, it just switches abruptly after the delay.

So my question is...Is there an analog switch out there that isnt buffered on the trigger so I can make it switch more smoothly?

The thing about using a circuit like this is that it has plenty of headroom at low cost. But the nanosecond switching wont work for me.

[Enzo]

ICs like that are after all analog switches. They are on and off, like a mechanical switch. If you want a smoother transition, like a crossfader on a DJ mixer, then you need a mixer circuit, not a switch.

[Tubis]

Well, a single jfet transistor switch can be made to go smooth by a simple RC network so an analog switch isnt always absolutely on or off. The thing is that for example the 4066 has a buffer that makes the trigger be either 0 or V+ no matter what you put on its trigger input. My fought was if there is a switch IC out there witout this buffer. The fading isnt ment to be made audiable its just to eliminate that slight pop you can get when switching to abruptly.

[eschertron]

Some designs will put a jfet on the signal line post switching to 'duck' the audio and mute out the pops &tc. May be overkill for a simpler circuit?

[Enzo]

Also, if popping is the problem, perhaps you have unterminated signal lines? For example, in a tube amp if we were to switch between two signals coming from a plate through a cap, we would get a pop every time we switch. So we add a high value resistor to ground to keep the cap charged. A 1 meg resistor won't materially affect the signal, but it will stop the popping.

So when using a 4066, are all the input and output lines terminated to ground?

[Mick Bailey]

I've had good success silently switching guitar signals with analog switches, without the need for soft-transition switching, but there are a few points to watch;

1. As per Enzo's reply - the switch behaves just like a regular mechanical switch. If were to sub in a regular switch you'd have the same problem - switching signals where there's a level difference through stored charge or cable capacitance. It's not the switch that's the issue, just the circuit.

2. Analog switches can put noise on the power or ground rails. You need to design the circuit with this in mind and decouple the IC and not share common grounding with the rest of the circuit.

3. Circuit paths and lead dress have a bearing on noise.

4. Not all analog switches behave the same. In a prototype you need to socket the IC and experiment with different chips of the same or similar type from different manufacturers. Some chips behave better in a given circuit.

5. Unused switches need to be properly terminated.

It's possible to get the design noise-free, but you need to work at it. I have one pedal that used 8 switches (2xICs) to route signals internally. It doesn't pop. Similarly, if you were to build a Nyquist Aliaser pedal using analog switches they switch at high audio frequencies - several Khz. If well laid out they can operate noise-free.

[Tubis]

The pop isnt something extra thats added. Its the sound that happens when you abruptly switches between signals. It doesnt pop when switching without signal. This pop isnt very loud but its enough that Im not satisfyed. I realy dont get how you can get around this fenomena. It isnt suposed to be possible when switching abruptly. This fenomena goes away when using a smooth switching system like delayed jfets or slow optocouplers.

Quote Merlin Blencowe "If an audio signal is switched on or off by switching it instantly then we will get a sharp corner on the waveform. This abrupt change in the waveform will contain high frequency energy that may be audible as a click. There is nothing we can do to prevent the click except throw the switch during silent passages.
If we do want to eliminate the click we must switch the audio on and off more softly"

[Mick Bailey]

A quad switch gives you the ability to momentarily mute the signal using a spare switch (by shunting the output) during switching events in a mute-switch-unmute cycle. Fast enough not to give any problems. This is a pretty common technique.

[Tubis]

A quad switch gives you the ability to momentarily mute the signal using a spare switch (by shunting the output) during switching events in a mute-switch-unmute cycle. Fast enough not to give any problems. This is a pretty common technique.

Okej. But i cant get that to work in my brain. It does the same thing, abruptly switch the signal to off. That should produce the same fenomena.
But that would be a god idea if one have a lot of relays in an amp and then put just one optocoupler at the end to mute of any popping.

[Mick Bailey]

Tom Morello and many others use kill switches - just the same thing, a short to ground. No problem with pops and they're intentionally switching the signal on/off under full guitar output. It all comes down to what you're switching and where.

Lots of amps use relays without any soft transition, and most don't pop or need any form of suppression, but you'll often find pull-up or pull-down resistors to equalize levels.

[Tubis]

Try muting this tone on and off https://www.youtube.com/watch?v=G9oYClOrS-k This is the fenomena Im talking about. It doesnt get as severe with guitar tone because its not a single sine wave, but its still there.

A qoute again from this thread: http://music-electronics-forum.com/t6152/

"One of the problems with most switching IC's is the buffer circuitry on the control lines. They ensure a fast transaction, even with slowly changing control signals - so no matter how much de-glitching you do to your switching control line, the switch will still change state rapidly once the threshold is reached. This causes minimum power dissipation through the FET switching devices. It also causes "popping" with audio signals."

[g1]

A quad switch gives you the ability to momentarily mute the signal using a spare switch (by shunting the output) during switching events in a mute-switch-unmute cycle. Fast enough not to give any problems. This is a pretty common technique.

Okej. But i cant get that to work in my brain. It does the same thing, abruptly switch the signal to off.

It does not do the same thing, like Mick mentioned, it temporarily kills the signal by shorting it to ground. This happens at the same time your other switch is doing it's "click" noise.
It's often called a "ducking" circuit, and an example is shown here:
http://music-electronics-forum.com/t15742/

[Enzo]

Peavey has used a ducking circuit in their big tube amps for years. Whenever it detects a channel switching, it momentarily shunts the signal path with a JFET. That prevents big pops coming through. The downside is there is a brief moment of silence.

[Tubis]

It does not do the same thing, like Mick mentioned, it temporarily kills the signal by shorting it to ground. This happens at the same time your other switch is doing it's "click" noise.
It's often called a "ducking" circuit, and an example is shown here:
http://music-electronics-forum.com/t15742/

As i said that would work if the mute is soft. If you mute abruptly with the exakt same fast switch you havent got rid of the problem. You just killed of the signal abruptly when muting and that will produce the click that happens when you turn on and of that fast. TSL muting i guess is made soft with a jfet after all relays that has two problems, fast switching and bouncing.

[Mick Bailey]

The only switch IC I can think of that has inherent 'soft' properties is the H11F1.

Edit: By which I mean it's not threshold dependent - the LED side can be transitioned - sometimes just a cap across the LED is sufficient.