Audio Switching

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

JRJR

Active member
Joined
Oct 8, 2014
Messages
34
Hi guys!

hope you're all doing well.

I have a problem without the right solution:

I need to design a switching system that can route one mono audio source to three different outputs via a rotary switch.
It needs to be remote controlled (4pin mini TRS cable would be cool), and without the remote connected I need a default routing to one of the outputs.

I just tried a simple relay switching with a 2 transistor opener that had the perfect behaviour but gave me unacceptable switching pops & clicks. I could also imagine a digital switch but would need one that does not require programming.
Classic analog multiplexers/CMOS switches like the CD4066 have too high distortion.

Do you guys have any idea on how to prevent switching pops/clicks or a device (Maxim has some stuff, but all I2S controlled) that would work? VCAs, some transistor circuitry, ICs?

Thanks in advance!
J
 
Simple enough. Just do two relays in series. You don't need transistors. The popping is because there is DC across the relay switches. It could be that there is a cap on an output and the leakage is causing DC in which case you just need a "drain resistor" to drain away that voltage. Meaning try adding large resistors to ground (like 22K on each input and 1M on each of the other switch contacts).

If you want a more detailed answer, post a drawing. Just draw something on a piece of paper and take a pic with your phone.
 
I've been playing with various solutions for guitar amps to monitor controllers.
Arduino can be used here. Or Teensy or Pi.
You can have them do other functions as well.
Also vactrols. Check Elliot Sounds as he has an excellent silent switching system using vactrols and it's cheap.
Or like mentioned using a relay that triggers a relay works well.
Finally you can use an attiny to be programmed which triggers the latching switch.
 
Real switches click too, but you can mitigate the clicks with HF pre/de-emphasis, albeit forgoing some headroom. I found about 12 dB of pre/de-emphasis effective.

JR
 
If you want absolutely clickless switching then you have to have about a 10ms ramp up and ramp down.  That’s good to somewhere around 40Hz for zero crossings IIRC. That’s usually close enough.

The SSM2402 used to be the go to part but it’s obsolete.
 
JRJR said:
I just tried a simple relay switching with a 2 transistor opener that had the perfect behaviour but gave me unacceptable switching pops & clicks. I could also imagine a digital switch but would need one that does not require programming.
Classic analog multiplexers/CMOS switches like the CD4066 have too high distortion.

Do you guys have any idea on how to prevent switching pops/clicks or a device (Maxim has some stuff, but all I2S controlled) that would work? VCAs, some transistor circuitry, ICs?
You need to identify the cause of the clicks. there are three things:
If there is DC offset, you need to provide DC decoupling
Relays can induce significant RF energy if there is no snubber; this energy propagates in the audio circuitry. You also need to make sure the switching currents does not interfere with the audio ground.
Interrupting an audio signal always produces noise, because it goes form 0 to something, except when switching at zero-crosssing, which rules out relays.
Overall, the best solution is using VCA's, as suggested by Gold.
 
Easiest would probably be something pre-made, like these: https://www.aliexpress.com/item/315Mhz-Universal-Wireless-Remote-Control-Switch-DC-12V-4CH-relay-Receiver-Module-and-RF-Transmitter-315/32817572941.html

HTB1nMyQdRgXBuNjt_hNq6yEiFXal.jpg



And if you get clicks (and you absolutely don't want clicks), remove DC offset or low frequencies in signal. Including from the input of the following stages (in case they DC bias without capacitor coupling)

Jakob E.
 
boji said:
What about IR or bluetooth? You could build up your 3-way switch with suggested vca's, fets or relays around an arduino.
Relays could be used so upon power off it would go to your default input.
Unless the source of clicks is identified as being control voltage or current feedthrough, that would not solve the problem.
 
Interestingly, I ran across a recent discussion along these lines at EEVblog,  One contributor mentioned this part:

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

I had never seen that part before.  The  "Typical Applications" section even show it being used to "tune" a SV filter.  Looks like a useful candidate for switching as well.

Bri
 
Brian Roth said:
http://www.onsemi.com/pub/Collateral/H11F3M-D.pdf

I had never seen that part before.
This is commonly found in guitar FX pedals, for On/Off switching, but also as a variable resistor in phasers and compressors or gates. The FET inside is optimized for THD reduction, but not up to pro audio standards.
As switchers, provided the signal range is in accordance with the limits, performance is very good in both On and Off states. During transition, where the signal is attenuated, THD is about 1-2%. It's generally not audible if the timing is correct. All FET-based switchers do that. The nice thing is that, due to the particular arrangement of the FET, THD is minimized without additional components. Being optocouplers, there is no control voltage feedthrough.
 
abbey road d enfer said:
This is commonly found in guitar FX pedals, for On/Off switching, but also as a variable resistor in phasers and compressors or gates. The FET inside is optimized for THD reduction, but not up to pro audio standards.
As switchers, provided the signal range is in accordance with the limits, performance is very good in both On and Off states. During transition, where the signal is attenuated, THD is about 1-2%. It's generally not audible if the timing is correct. All FET-based switchers do that. The nice thing is that, due to the particular arrangement of the FET, THD is minimized without additional components. Being optocouplers, there is no control voltage feedthrough.
I've used simple JFET shunts in value (cheap) noise gate designs and could hear the distortion during transitions with slow release times. Admittedly not very audible on complex signals but clearly audible on pure tones.

I used the popular technique of injecting 50% (-6dB) of AC drain signal into gate control voltage to effectively reduce the distortion below easy audibility.

JR 
 
JohnRoberts said:
I used the popular technique of injecting 50% (-6dB) of AC drain signal into gate control voltage to effectively reduce the distortion below easy audibility.
i don't have details (maybe patented), that's apparently what they do in H11F3, but not in discrete form. Since it's an opto-FET, the gate is not accessible.
 
With an opto FET, you have essentially a floating gate, so the 'inject some compensating signal' into the part will be difficult. However, another trick available to you is that you can use two FETs back to back (connect the sources, put the drains facing out) and the FET resistance variation will cancel itself out, as long as the FETs are reasonably matched. Some of the opto FETs do that internally for you, since it allows bidirectional current flow within the FET switch. So, if you specify an opto FET that is rated for AC switching, it will already have that trick installed.

In that configuration, a FET switch will have very low distortion, and because the "gate" is floating, you can use it as a high side switch without worrying about getting your gate drive circuit to float along with the signal.
 
There's also high voltage cmos gates. If done correctly performance could rival a relay. Some do not need digital control. Something like ADG1411 doesn't need digital control, it can work with single supply (or dual), the control inputs can take Vdd voltage and the Ron is so low that you don't even need buffering. Also, if I'm not reading the datasheet incorrectly, it looks like it runs on 0.38mA. So it could run on a 9V battery continuously for literally probably a month.
 
Monte McGuire said:
With an opto FET, you have essentially a floating gate, so the 'inject some compensating signal' into the part will be difficult. However, another trick available to you is that you can use two FETs back to back (connect the sources, put the drains facing out) and the FET resistance variation will cancel itself out, as long as the FETs are reasonably matched. Some of the opto FETs do that internally for you, since it allows bidirectional current flow within the FET switch. So, if you specify an opto FET that is rated for AC switching, it will already have that trick installed.

In that configuration, a FET switch will have very low distortion, and because the "gate" is floating, you can use it as a high side switch without worrying about getting your gate drive circuit to float along with the signal.
I guess that's the case in the HIF11. The datasheet just shows one FET without distinction between drain and source, but I guess it's a back-to-back arrangement.
 
Hi,

I agree withsquarewavehere, ADG 1411 would work well in this scenario.

You could also consider MAX338/MAX339, CMOS analog multiplexers which are designed exactly for your case. (1 out of 3)

The advantages here is the easy implementation of those chips. You don't really need anything else but the chip and a bit of decoupling.

Cheers,

Thomas
 
squarewave said:
There's also high voltage cmos gates. If done correctly performance could rival a relay. Some do not need digital control. Something like ADG1411 doesn't need digital control, it can work with single supply (or dual), the control inputs can take Vdd voltage and the Ron is so low that you don't even need buffering. Also, if I'm not reading the datasheet incorrectly, it looks like it runs on 0.38mA. So it could run on a 9V battery continuously for literally probably a month.
I've used a lot of these CMOS switches in the past; I agree with the notion their performance is adequate as long as they're used within limits. However, they won't solve the issue that hard switching signals does produce clicks, by interrupting a signal that is non-sero and replacing it with another. As JR mentioned, switching at zero-crossing is one solution (that's how it's done in the digital domain), or ramping the levels. Discrete JFET's allow this, at the cost of significant distortion in the transition period.
One could use zero-crossing detection to sync the command signal to a CMOS switch.
The use of an HIF11 is probably one of the most clever solution, but still requires attention to timings.
 
totoxraymond said:
You could also consider MAX338/MAX339, CMOS analog multiplexers which are designed exactly for your case. (1 out of 3)
MAX338/MAX339 Ron is 200 ohms and not particularly flat so it's not ideal for high performance audio - there will be measurable distortion without buffering. ADG1411 Ron is not particularly flat either but it's only 1.5 ohms so the distortion will be much less.

Note that when talking about these low Ron gates (the low Ron is what makes then special), the junction capacitance goes way up. An ADG1411 gate that is "on" has a capacitance of 116pF.
 
Back
Top