Relay "pop" noise problem

[gtudoran]

A bottom pit for me ... tried everything ... and nothing helps. So this is the schematic (there are 6 loops like this but for the sake of space used i've draw only one). The problem: even with nothing inserted in J1 and J2 and even with the audio input grounded ... i get a pop whenever the relay is switching (on and off) ... i don't have any idea why (as you can see there is no connection between the audio path and switching system). The logical part has it's own 5w supply and also the relay module has it's own 5v power supply (both of them regulated 78L05 for the logic supply and 7805 - TO220 for the relay supply). Even-more... the supply for the relays has 220uF at the regulator input. Also tried to add a capacitor (100n) between the B and C of the switching transistor... well that's it ... don't know what else to try. Any idea would be appreciated.

This is the schematic of one loop:


The relay is driven by a 5v logic form a PCF8574 I2C expander....

Best regards,
Gabriel Tudoran
Analog Sound

PS: do you think that useing a dedicated driver like ULN 2003/2803 would solve the problem?
PPS: the relay model is AXICOM IM03 - SMD type

Thank you in advance for any help / idea that you can provide.

[GibsonGM]

I'm by no means an expert with relays + audio...but what I see that MAY help (not sure tho) is to add SOME resistance to the collector at your 5V.   For example, at the top of the 1N4148, you could add a resistor that feeds only the relay coil.     This would, to me, limit the current the coil is drawing and make it 'less spiky', if that is a proper term

Say the relay needs 100mA at 5V....so, you use 5v/.1A = 50 ohm resistor - which I would decrease to 33 ohm just to give it  'headroom'.   I believe your 'pop' is a current spike coupling into the audio circuitry right next door....and perhaps a selectively placed cap would also be required to "slow down" the spike?    Anyway, this resistor is of course separate from the stuff associated with your LED.

Just a thought, hope it leads somewhere for you!

[samhay]

Only time I have seen a resistor in series with the coil is to drop voltage - you could try this instead of using a separate 5V regulator for the coil.
R2 seems quite small for a relatively high gain transistor. Not sure how much it will help, but you could upping this a far bit.
Any chance it is the LED causing the popping? If so, you could directly drive this from your uC.

[gtudoran]

@samhay thank you. I use a separate regulator for the relay supply voltage, with 220uF before the regulator and 10uF after the regulator. I could use a 10-20k there (i will give it a try) ... maybe that could be the answer... i burnt out all the logical explanation for this... before that, i used 2n3094 and the relay didn't energize ... so i thought i use a higher gain tranny. So yes i guess i could upper the base resistor a bit.

Best regards,
Gabriel Tudoran
Analog Sound

[samhay]

No worries.
If you are having trouble with the transistors turning on far enough, you could also trying using a MOSFET instead. A BS170/2N7000 works nicely here.

[therecordingart]

Is it possible that the current change through the coil at the time of switching induces a current  in the neighboring connection(s)? Maybe put the short on the other side of the switching action? Guess a complete stab in the air since it sounds like you've tried everything else.  

http://imgur.com/klEXirY

[gtudoran]

Hmmm i didn't quite get it.

Is it possible that the current change through the coil at the time of switching induces a current  in the neighboring connection(s)? Maybe put the short on the other side of the switching action? Guess a complete stab in the air since it sounds like you've tried everything else.  

http://imgur.com/klEXirY

@samhay hmm i will also try a fet switching ... might also help as the gate is voltage controlled not current controlled.

[merlinb]

FET switching won't be any better than the BJT. I suspect you need pull down resistors on the the insert loop,bypass loop, input and output. (The device you're using to listen to the pop might also be sending DC back to the relay...)

[therecordingart]

Hmmm i didn't quite get it.

Is it possible that the current change through the coil at the time of switching induces a current  in the neighboring connection(s)? Maybe put the short on the other side of the switching action? Guess a complete stab in the air since it sounds like you've tried everything else.  

http://imgur.com/klEXirY

@samhay hmm i will also try a fet switching ... might also help as the gate is voltage controlled not current controlled.

When current passes through a coil a magentic field is created around that coil. A separate conductor within that magentic field will have a current induced within it. So at the time of switching I imagine that the rapid current changed in the relay coil could potentially induce a current within a neighboring connection on the relay. Think of the circled area like a transformer. Putting pulldown resitors on the audio in/out might help?

http://imgur.com/klEXirY

I could be completely false. I'm probably bro-sciencing right now.

[MrStab]

i'm no expert either, but thought i'd offer a few ideas as someone on the outside.

i spent 2 weeks trying to minimise pop on a pair of small (latching) relays a few months ago, and in the end, i had to give in to the possibility that sometimes there'll always be a tiny click in the audio path from the mechanical aspect of the relay. there's just one post deep on these forums where somebody said a tiny bit of pop is inevitable, past a point.

i have, however, had better results with non-latching relays in the same circuit, but their placement involved only low-impedance signals (unlike when i used the latching ones). the pin spacing was further apart, though, and as my circuit was quite high-gain, i have a feeling this may be a factor.

• have you tried isolating the ground(s) from the relay switching? ie. a separate wire to the final grounding point from the relay, not connected to the signal ground or other power grounds, except at that point.
  
• i've also read that connecting the relay ground via. a 100R (i think) resistor in series can help.
• does the same circuit pop if you use an analogue switch?
• is there any difference if you use high or low-impedance sources only?

sorry if any of the above is obvious or stupid. or obviously stupid.

hope you get it working!

[gtudoran]

Thank you for your time. The problem is this: the pop is there with nothing inserted and  the input grounded (main input). The PCB doesn't have any ground plane (i usually do copper poring but ... not now).

So i guess is not a problem about impedance and dc from the input....

Best regards,
Gabriel Tudoran
Analog Sound 

i'm no expert either, but thought i'd offer a few ideas as someone on the outside.

i spent 2 weeks trying to minimise pop on a pair of small (latching) relays a few months ago, and in the end, i had to give in to the possibility that sometimes there'll always be a tiny click in the audio path from the mechanical aspect of the relay. there's just one post deep on these forums where somebody said a tiny bit of pop is inevitable, past a point.

i have, however, had better results with non-latching relays in the same circuit, but their placement involved only low-impedance signals (unlike when i used the latching ones). the pin spacing was further apart, though, and as my circuit was quite high-gain, i have a feeling this may be a factor.

• have you tried isolating the ground(s) from the relay switching? ie. a separate wire to the final grounding point from the relay, not connected to the signal ground or other power grounds, except at that point.
  
• i've also read that connecting the relay ground via. a 100R (i think) resistor in series can help.
• does the same circuit pop if you use an analogue switch?
• is there any difference if you use high or low-impedance sources only?

sorry if any of the above is obvious or stupid. or obviously stupid.

hope you get it working!

[MrStab]

ahh, okay - sorry, i read the "nothing connected" bit, then re-read it, and somehow thought i'd got it wrong the first time!

taking that out of the equation, my main suspicion is current surge/grounds. if you haven't already, i think you should try isolating as much as possible and maybe the 100R resistor i mentioned. both are advice from RG, iirc.

maybe some kind of decoupling could also help.

[gtudoran]

Hmmm could my wiring be a problem, the relays are wired as in the first picture... the 2nd picture seems... better.

[MrStab]

the only difference in the 2nd one is that the board input seems to be grounded when in bypass mode - otherwise they work pretty much the same.

[R.G.]

Bypass switching pop seems to get rediscovered by each new generation, like a lot of pedal technology basics.

1. Any discontinuity in your audio signal will sound like a "pop" or a "tick" or a "thump". The difference between these is only in how long the transient lasts, and how abrupt the transition is. Loudness of the artifact depends entirely on how big the transient peak is.
2. Anything that causes that discontinuity can be a cause. Like hum, there is no one answer. To eliminate ticks, pops and thumps, you simply have to do many things right. Miss one, you get transients. There are many sources.
3. Pulldown resistors fix only open-capacitor pops, nothing else.
4. Leaky capacitors that let DC through to the output can cause pops that pulldowns are not strong enough to dispose of, and you get pops.
5. Hard-contact switching - metal contacts on metal contacts - causes pops from the instantaneous switching of audio if you happen to switch when the signal is non zero. This cannot be designed out of metal contacts, and the only way to avoid it is to make the contacts touch only when the signal happens to be zero.
6. Slow switching generally removes all audible transients by making the transitions so slow that they fall outside the audio band. This works only if you do the slow switching well, of course. Boss and Ibanez do this.
7. It is possible to get pops from the discontinuity of the power and/or grounds to the circuits. If bypass switching involves relays or LEDs, the current these use is large enough and sudden enough to cause the power line or ground wiring to move a fraction of a volt when it happens. The circuit then obligingly amplifies this as a pop. The cure is the same as for signal switching - make the transition slower, or smaller, or both. It is possible to fix this in most cases by more-educated wiring of the power and ground distribution. Other band-aids are to add resistors, caps, diodes, transistors and such to mess with how the switching happens.
8. Relays are a special case. In addition to the normal problems of switching just the signal relays can add pops as noted in 7 by heavy, sudden changes in current on the power and ground wires AND by direct coupling of the coil switching voltage into the audio lines. They also have coil wires carrying sudden full-power-supply voltage changes being very close to signal wires carrying millivolt audio signals. Since there is capacitance between everything in the universe, some of this sudden voltage change gets across the coil-to-signal capacitance, and you hear ticks and pops. This can be fixed by (1) using shielded relays; they exist, but are hard to find and expensive; (2) slowing down the coil voltage change a lot, so the frequency content of the coil voltage is poorest at high frequencies and doesn't couple across well; (3) making the source and load impedances on the signal wires low so that the finite charge coupled across is held to a lower level; you still get a pop, but it's dramatically lower in level.

And you have to do it all correctly to be pop free.

[gtudoran]

Thank you R.G for your explanation and thank you all for your time. I've solved the problem by replacing the relays with some bigger ones (5V). But i run into another problem now and i can't understand why.

I drive 6 relays with an ULN2003 (datasheet: http://www.ti.com/lit/ds/symlink/uln2003a.pdf) wich is driven by logical 5v signals form a port expander PCF8574 -  and something strange happens: the first 3 relays are working as they should... but the last 3 are not sometime they engage and sometimes not ...

I've measured the command voltage (on the ULN input pins) and it's 1.8V. On the relays that work, the voltage drop across the coil is 0.8 - 0.87v but on the other relays (the last 3 of them) the voltage drop is 1.9 - 2.3v

The relays are 5V type and the supply voltage at the ULN common pin is also 5V.... does the voltage at the common pin of the ULN should be higher? I've tried with 3 different IC and is the same thing.

I have to mention that the behavior of the last 3 relays are dependent by the state of the first 3 - when 1 one or more of the first ones are engaged ... the last 3 are working, but when none of the first 3 are engaged the last 3 are starting to behave strange.

Do i need ... pull up resistors on the input of ULN?

Best regards,
Gabriel Tudoran
Analog Sound  

[R.G.]

Check your relays to see if they have internal diodes. Some relays have the de rigeur catch diode inside the relay. That might be too high a current load for the 2003 to pull down. If there is a + and - pin notation on the relay coil, it usually means an internal catch diode on non-latching relays.

[gtudoran]

Thank you R.G. ... i don't think is that. They are just ordinary 5v relays. Is like the last 3 outputs 4/5/6 can't pull down the voltage and i can't figure out why... is like a $#@ curse ...

[gtudoran]

Problem solved with 15k pull-up resistors on the PCF output lines. All good now. Thank you all for your help and time provided.

Best regards,
Gabriel Tudoran
Analog Sound