0

Attached is a stripped down version of the circuit. The relay is controlled by a push button going to a microprocessor then powered by a BJT with a flyback diode in parallel.

The output of the DSP/etc. needs to stay engaged when the input is disconnected, and the single push button controls other aspects of the DSP which is why the relay is required.

Image of the circuit all unlabeled resistors are on Meg ohm order of magnitude, caps are microfarad

I feel like I'm going crazy trying to diagnose where the popping sound is coming from. I've pulled down every point of contact for switching. I tried an opto coupler to ground on the output but that just messes with the output decoupling capacitor. I have tried running the relay on a separate power supply. I have tried shorting out various buffers/the summing amplifier. I know there will be some amount of contact bounce but the sound is much louder then the signal, and when comparing to devices using relay bypass; it is much louder.

Edit: A little more clarification is the non inverting terminal of the summing amp is going to half power supply. Which is why I used a different ground symbol. Could the fact that the input cap to the summing amplifier is at half power supply but the switch terminals are connected to a lower potential, cause the pop?

I have just tried swapping out the relay for a manual switch and the pop still happened.

OnlyServeSwill
  • 1
  • 2
  • Hm could you show us a picture of your hardware? (Your physical board)? – Christianidis Vasilis Sep 08 '23 at 19:09
  • It is a PCB I designed, but the "DSP/ETC" is the majority of the circuit so I don't think a picture of the crowded PCB will be too helpful since it's mostly irrelevant components. I have a scope and verified that the pop is not generated by the dsp area of the circuit. – OnlyServeSwill Sep 08 '23 at 19:15
  • This seems to be a duplicate of https://electronics.stackexchange.com/q/459924/10810. – Kuba hasn't forgotten Monica Sep 08 '23 at 20:20
  • A pop will occur if you switch during sounds on one side and silence on the other side. I worked with a sound system that slowly muted the signals before switching then slowly ramping up the signals after switching. Slowly to avoid a pop but fast enough not to notice the delay. – Audioguru Sep 09 '23 at 00:55

2 Answers2

0

I'd use a pop-free audio signal switch, rather than a relay. For more, see this question.

But first, ensure that it's really a switching problem, not a control-coupling-to-audio problem. Replace the relay with a quality switch. If the pop is still there, you'll know that a simple switch in that position won't work, relay or not.

Kuba hasn't forgotten Monica
  • 40,960
  • 1
  • 39
  • 120
  • I just swapped it out with a manual switch to test, and the pop was still there it sounded like the same kind of pop too.. This means that it's most likely not control coupling correct? It needs to be a relay, or something similar, as it is triggered by a control signal from a microprocessor. – OnlyServeSwill Sep 08 '23 at 20:51
0

When the relay switches there will be a short delay between cutting off one signal path and establishing the other, so a few milliseconds of silence while switching is inevitable.

When signal is applied to a filter, there is always a settling transient. If your DSP implements a filter (or even just a delay), since you switch its input on and off, there will be a transient (could sound like a pop) at the output every time the input is switched on.

In addition the DSP adds latency, so when its input is turned on the output will remain silent for the duration of this latency. So after switching the signal will be cut off for about the sum of the relay switching time and the DSP latency.

A solution could be to not switch the DSP input, but the output instead.

bobflux
  • 77,207
  • 3
  • 91
  • 222
  • It still happens when I short out the DSP/etc or leave it omit. I tried clamping the output to ground using an optocoupler timed with the relay switching but that discharges the output coupling cap. My next move will be doing that but instead of output to ground; it will be between the final buffer and the output coupling cap and connect to half power supply. – OnlyServeSwill Sep 08 '23 at 23:55