Relay switching causes *CRACK!*POP!* in audio chain - help wanted please.

Hey guys,

I'm modifying a 4-ch headphone amp, adding in a microphone & preamp for talkback.
I'm using a relay to switch the talkback mic, but I've found that when I energise/de-energise the relay, there is a loud *CRACK!* in the headphones' output.

The headphone amp PSU is 7815 & 7915. The mic preamp is powered from 7815 thru resistor for 12v. The relay is 12v, but running from 15v directly from the 7815 output.


Can someone suggest a solution to my problem, please?

THANKS!

UPDATE: It's easy to Google things... if you know the name of what you're looking for.

And it turns out the name is "Flyback". Here's an interesting article: http://en.wikipedia.org/wiki/Flyback_diode

I shall go and try it out and report back my findings.


(P.S. I found the solution through this forum, so, thankyou )

A resistor in series with the "reversed" diode seems to make it a little better.

It's still not perfect, but that could be due to the fact that's it's all running from a shared power rail.

Is the relay switching mic in/out from micpre or micpre from headphone amp (do you have a blog diagram or schematics)?  I just read about contact protection (from seemingly great book Noice reduction techniques in electronic systems by Henry W. Ott) and perhaps could try to help later today if no one has already answered to your problem.

Here is some info
http://www.leachintl2.com/english/english2/vol6/properties/00049.html

and more with google searches like "contact protection networks for inductive /capacitive loads".

edit: this might not solve your problem, but contact protection is interesting read anyways so you should check it out  8)

DaxLiniere said:

A resistor in series with the "reversed" diode seems to make it a little better.

It's still not perfect, but that could be due to the fact that's it's all running from a shared power rail.

Click to expand...

I believe, in addition to the reverse diode, you need a "snuffer" across the switch; typically an RC series-circuit with ca. 100R and 0.1uF

abbey road d enfer said:

a "snuffer" across the switch

Click to expand...

ITYM snubber.

JDB.
[depending on your layout and the stiffness of your supply rails, the diode can go across the relay or across the switch]

A snubber should not usually be necessary for the small relays we use for audio, but if it helps of course why not. See if the rail drops when switching the relay; if it does a couple of 100 uF of decoupling capacitance close to the relays might help. And make sure you have no DC on the signal you're switching.

Samuel

Just re-read your first post: where in the signal path are you switching? You're not opening the mic input, are you? I guess a FET switch would be the better solution anyway for this application. But perhaps you show us a schematic first...

Samuel

+1

If the mic is a simple electret with fet built in and biased up at DC, you need to make sure there is no DC being switched, which will cause a step jump when switched.

JR

+2 for schematics of what you are doing.

since I used to work in power control, I used to have to switch heavy loads with contactors and know a lot about snubbers and flyback(recirculator) diodes, TVS devices, etc.

Let's see what you are doing and we can suggest a quick solution.

Hey everyone! Thanks for your interest!

I only supplied a little bit of info in my first post because I figured if I bored everyone with mindless details, my questions was unlikely to be answered.


So, the story is:
*It's a small 12v relay reclaimed from something else (who knows what), coil is 273r I have a slightly larger 12v relay lying around that has a 139r coil.
*I haven't 'inserted' it into the audio signal path yet. Only power is connected as I was just testing it when I noticed the crackles.
*Switch (not currently installed) will be a microswitch (inside a footpedal)
*The mic is electret, running into a "pre-champ" basic 2 transistor preamplifier (Jaycar), but I was planning on creating the mic-mute by shorting the preamp's output (aka grounding the next module's input).
*I currently have 1N4004 & 3k3r in parallel with the coil. I know this is not the 'correct' placement for the resistor, but I couldn't see a good reason for putting it in series with the coil, which would only slow it's energise/deenergise time, which I guessed would make the snap, crackle, pop situation worse.

As it is currently, I could live with it, but any improvements would be great. (Short of rebuilding the PSU which would be more effort than it's worth to remove the remaining amount of click.)
I am intrigued by the idea of adding an RC network to the switch. I guess that would be in parellel with the switch, but then I suspect power would just circumvent the switch and flow through the 100R, energising the relay, right?

Thanks again for your interest, everyone!

Another thought, perhaps it is a good idea to put a resistor in series with the relay coil, considering it's a 12v relay running from a 15v regulated supply.

3v voltage drop, 273r coil resistance... how does the rest of that song go?

100R in series with coil, plus 'reversed' diode in parallel with coil (not coil+resistor) has done the trick beautifully!

Thanks again everyone

ah yes I was about to suggest that but you beat me to it.

The 'flyback' diode will take most of the energy out of the transient but you will still get enough of a transient back through the Vreg to cause audio problems.  The series resistor limits the small current in the transient and the small capacitance of the wiring/PCB trace will likely take care of the rest with voltages this small.

Whenever you use coils/inductors with DC switching you want a little series resistance and then possibly a small decoupling cap to suppress any inductive kick(coil side snubber).  In a situation like this you want to make a decoupling cap's voltage rating many times higher than normally used as the voltage spike will be higher than your source voltage.  Sometimes you might also use a reversed biased zener/TVS(rated for slightly above the operating voltage of the coil) to ground to *clip* those voltage spikes too if nothing else helps tame them. 

Since nobody else has mentioned it yet... There is a way to just about completely elimnate the pops when forced to use relays on a commn rail, is this is sort of overkill, i dont know, you be the judge. 

Feed the relay coil with a current source And put a switch in parallel with the coil. (can be any variety ccs, simple 2 transistor type is fine). When you want the relay off, close the switch.  Put the reverse diode in parallel too, but the ccs is going to do the real work ensuring that the current draw on the rail is unchanging. The real advantage of this approach is if you have multiple relays, use low voltage types and put the coils in series. A 15v rail could support 4 relays with 3v coils. All driven by one ccs.