Remote switching design help

[Chuck]

Ok,
So I put together my second attempt at R.G.'s remote footswitch bypassing system.  The one shown on www.geofex.com.

This time using the NEC EA2-12NJ minature signal relays.
I still get an audible click in the amp when switching.

I tried increasing and decreasing the capacitor which should slow down the switching transistor.
I also tried moving all the DC current using portions (switching diode, LED, etc) as far from he audio circuits as possible.  (Even removing them at times.) 

I've tried three different relays now.  I had my hopes on the EA2.

I'm seeking suggestions on reducing the click noise.  Anyone have a sugguestion?


I'm also seeking any other remote switching ideas which would not produce tone sucking, hum, buzz or switching clicks.  If someone knows of a design which does tha please let me know.


I appreciate any response.

Chuck

[RaceDriver205]

Are you familiar with the 4066 and its use in switching?
If not I can help you there, its better IMO than a relay.

[Gilles C]

From what I already tried, I found out I could have as much popping from relays or 4053s or 4066s than from a regular footswitch.

Some precautions must be taken in how the ground and power is laid out, and some resistors and caps must be used in some places.

I can't give you the perfect formula because I still have to find it, but I just wanted to tell you that these things happen, even with relays or analog switches.

Gilles

[R.G.]

Do your switching ground and signal ground touch?

[Chuck]

R.G,
I currently have a 1 meg ohm resistor between the grounds.  I tried direct connect and the 100 ohm as you suggested with no difference in the click, only in the buzz noise.  I'm not getting any real noise with the 100 meg ohm, other then an audible click when the relay switches.  In both directions.

I'm open to more suggestions R.G.  I've tried two relays now, and I etched a circuit board and used a bread board.  The circuit board is ly quieter as for any transient noise.  Ther's just this click, nearly a pop, but more of a click when switching that can be heard from the amp speaker.

Thanks Gilles, I wonder if all the relay and 4066 designs click and I'm just too picky.

RaceDriver.  I'll probably try the 4066 on the breadboard design next weekend.  Gotta take my daughter to college tomorrow, so I'll be gone a few days.

[Gilles C]

Nah... I didn't mean they had to pop. Just that some precautions must be taken to make sure they don't click or pop.

I'm quite sure that the problems I had were caused by a bad layout of mine. Mainly the ground that I didn't separate enough, if at all. I don't remember, but I should work on it again some day.

Gilles

[R.G.]

Clicks and pops can only happen in certain limited ways.
1. Direct coupled. If you suddenly connect a different DC level, however small, you'll get a "pop" whose magnitude depends on the size of the DC difference. The higher the resistance of the amp you feed this to (e.g. a 1M guitar amp input) the more prominent the pop will be.
A variant of DC transients is common mode.  It is also sufficient to make a pop if you suddenly move the ground wire and signal wire both. This is called common mode noise, where the signal and ground both move together.  Guitar amp inputs have zero common mode noise rejection.
2. Capacitively coupled. If you connect you signal wire to a noise source through a capacitor, and then place spikes on the noise source, the noise will be capacitively coupled to the signal line.The magnitude of the transients is dependent on the size of the capacitance coupling the signal line and the impedance of the amp receiving the signal. Again, a high impedance guitar amp with zero common mode rejection is almost a classical worst case. A "capacitor" is any two conductors separated by an insulator - like for instance, two wires separated by a half-inch of air. So other wires just passing near a signal wire can induce a transient. The lowest frequency of the transient that's noticeable depends on the speed of the transient (that is, how much high frequency content it has) and the relative sizes of the parasitic capacitances and the amp impedance.
Notice that the signal ground shield is vulnerable to this as well, because of the zero common mode rejection.
3. Magnetically coupled. If you have a signal carrying loop of wire and a noise carrying loop of wire, the magnetic field from the noise loop can magnetically couple into the signal loop and generate a tick. To minimize this, you make the loop areas as small as possible and move the wires apart. M-field interference works best on low impedance signals, though.

Common problems with all switches, both electronic and mechanical, is that the control signal couples into the signal line. A purely mechanical switch (that is, a stomp switch) works pretty well because there is no electronic signal to couple. A relay has more problems because the coil current is going through a coil of wire very close to the signal carrying conductors. With relays, you can't change the distance between control signal and signal carrying conductor, so you slow down the transition. Relays also carry a lot of current to activate their coils. So if your coil current ground connects to your signal ground, you can cause ground transients. In both cases, slowing down the speed of the transient edge makes it harder and harder to transmit the click through the capacitances by lowering the frequency content.

I am surprised you're having so much trouble with the EA2's. I've used them with no problems.

Here's an experiment to try, though. Rig up a way to turn the relay on and of by turning a potentiometer.  For instance, grab an unused and unloved 10K pot. Chances are that it will carry 20ma of coil current with no damage, but there aren't any guarantees. Hook up the pot as a two terminal variable resistor between your coil power and the relay coil. Now turn on your relay by twisting the pot shaft slowly. Do you still hear a click? If so, then the click is being generated by something in the signal path because the pot method ramps the coil current up and down so slowly that it practically can't cause a transient. If not, the problem is related to the turn on/off time of the coil.

[Chuck]

Thanks RG,
I'll try that when I get back from this trip.  Great idea and tips.  I'll digest it.

But just so you know my motivation here is the situation.

I built this cool rack enclosure which looks like it was made by Marshall.



It has a pedal drawer behind the speaker grill cloth covered front piece.  (Note:  The pedals shown are not the final product.)



I want to built the switcher for the pedals behind the right side panel.  The power supplies for the pedals and 4 ru rack are behind the left side panels.  (Note:  The jack on the right is the patch to the amp head.)



I didn't want to waste one of the 4 RUs on the switcher.  And I can't buy a switcher which fits in my space unless I rip it apart and try to jam it in.

I'm going to try and pot a relay on the breadboard as soon as I can get some breadboard time.  It's a wonderful plan and excellent click/pop/noise primer.

Thanks,  I truely appreciate the help.

me

[RaceDriver205]

Which of the units is yours! I think its the middle one, but it looks so 'real' ! 
How on earth did you make it?

[Seljer]

Nice rackmount Crybaby 

[Chuck]

RaceDriver,

Yes, its the center one.  A Marshall Rack Mount/Pedal Drawer Case. 

I bought about $100 in Marshall Parts and some elephant hide/tolex.  The cabinet is made from 1x12s, routed and glued and covered with the Tolex.  The switch, corners, skid plates, feet, white piping, Marshall emblem and speaker tweed are all authentic.  The hardest part to find where the gold alumn sides.  There are 4 ru rack rails front and rear which I happened to have laying around.  Took two weekends to build.
It matches the TSL100 amp head and 1960 speaker cabinet exactly.

Now as soon as I get back I'll try the pot idea.

C ya

[Chuck]

Ok, I tryed it on a breadboard.

Just the Ea2 relay, wired to 4 jacks.  In, loop send, loop receive and out.

Then I put a 10K Audio pot on the 12 v supply to the relay.  Plugged in an amp and slowly turned the pot.
I have a slight click click.  Even at higher volumes, the click in the amp speaker is about the same loudness as the click from the relay itself.
I tryed a second relay with the same effect.

So based on my previous attemps, I know this should be lower if I etch a board.  So next I plan to etch a board, with PC mounted jacks and the relay and try the pot again.

Maybe I'm still being too picky.

Chuck

[R.G.]

If the click in the speakers at higher volumes is comparable to the mechanical click of the relay itself, you may be at about as good as you can get.

The pot method is dead slow, so the only click you get is the click from the relay actually moving.

Was your relay connected to anything? or contacts shorted to ground? or connected to ground by perhaps 1M?

[Chuck]

R.G.,

Yes, thats what I'm thinking now -- that I'm there I just don'e know it yet. 

In the test bed, there were four 1/4 jacks.  Their sleeves all connected together, with a 1 M Ohm resistor to ground.
The tips of the jacks went to the relay.  The two jacks for the effect were normalled.  I didn't actually use a affect and let this closed bypass the signal.

Aside from the four tips, and a jumper to bypass the signal when the relay is unpowered -- I had a regulated 12 volt dc power supply.  The ground went to pin 10, the 10 k audio pot went between 12+ and pin 1.  I should have just drawn a picture.

So anyway, I'm moving forward with some etched boards.  I also ordered some 5 volt EA2s which should be here Monday.  I figured I'd try them just for grins since I've gone this far.  I'll use sockets so they'll swap in easy.

When I get these boards done, I'm going to try a live test with sound and effects.  I expect the results to be good.

thanks for the help r.g.

Chuck

[R.G.]

I was just thinking that if you connected the signal path  to signal ground with a resistor, perhaps 100K or so, it might clean up the remaining click.

[Chuck]

Thanks.
There was no noticible difference between 1m and 100k ohms.

Could it be microphonic at all?  I swear I can actually hear a metalic tink in the amp speaker when going to the on position.  There is a slight tick/sound change when turning off.

Chuck

[R.G.]

There are a host of very tiny influences in the make and break of all mechanical switches. Generally these are too tiny to be concerned about, but it is possible that you are hearing them under the conditions of quiet background and high amplification.

There are indeed some piezoelectric and microphonic effects. There is also a tiny voltage generated when dissimilar metals touch. The dissimilar metals can be solder/copper/ and relay contact metals, and the precious metal contacts merely connect the dissimilar voltages. But this kind of thing is a concern for microvolt sensing setups and not usually a problem in audio.

Since the EA2 is intended for low level switching, I'd guess that most of the secondary effects are pretty small.

Here's an interesting question. We have eliminated the electronics driving the relay as a source of noise entirely by the very slow pot turn on/off. Let's eliminate the relay, too. Install a mechanical switch in parallel with the relay and see if you can hear that pop when you flip it and when the relay is off. Perhaps try making the same contact with a copper wire and see if you can hear that as well. These are all interesting tests I never had the time to do.

[Chuck]

Yes, that is my motivation for etching a board.
Which is what I'm about to do. 
My original plan was moduler, so I need the board anyway.

I figured if I used an etched board I remove all the breadboard connections which IMHO are shakey when compared to good solder.

I'll also try the switch idea soon -- as well as use another amp.  I have been testing with a cheap Fender practice amp on the workbench.
I have to dig up a switch. 

I read in another post somewhere that putting a 1m ohm resistor across the ring and sleeve of the jacks would quiet down noise, but I would think that could affect the sound.  I haven't tryed it yet.

Chuck

[QSQCaito]

I read in another post somewhere that putting a 1m ohm resistor across the ring and sleeve of the jacks would quiet down noise, but I would think that could affect the sound.  I haven't tryed it yet.

Chuck

This is the same as a pulldown resistor, it lowers noise. You can ussually see them on effects, check some Layouts at tonepad, the first thing you'll see will be a 1M pulldown resistor from input to ground(ring 2 sleeve) and then the cap..

bye bye!

DAC

PS: I think they won't affect noise, they just(if not wrong) discarge capacitor, which ussualy causes pop.

[Chuck]

R.G.,
Ok, we can put this to bed. 

I drew up a pc board, etched it, drilled it.  Used board mounted 1/4 jacks.  Socket mounted a EA2-12NJ.
By passed everything except for the relay diode.  Used a 100K ohm resistor between the audio ground (sleeve) and the power ground.
I connected the + side of the relay to a 9v battery.  I used a 10k Audio pot between the neg side of the battery and the neg side of the relay.

I placed a Boss DS-1 Distortion Pedal on the effects loop with 6 inch pedal patch cords.
I connected the switcher out to the Marshall TSL-100 above with a 15 foot Planet Waves cable.
I connected a Gibson SG to the switcher in using another Planet Waves cable..

Perfection.  Switched off there is zero noise.  I had to strum the guitar to make sure the amp was even powered on.  Total silence.  No humm, hissing, nothing.  When I turn the pot there is no click, just the slightly distortion hiss when the DS-1 comes online.
When I strum and rotate the pot back and forth -- there is no sound except the guitar and the distortion coming on and off.

If I disconnect everything but the cord to from the switcher out to the amp, the click is so low you can't hear it unless you listen hard for it.  If I crank the Marshall, I can hear it.  But only slightly and this is a cranked Marshall.  Its totally acceptable.

I think the problem was a combination of the loose Bread Board connections and the cheap Gibson practice Amp.  When I use the quality cables and solder it to a board it sounds great.

Next I'll put the transitor/cap/resistor circuits in and connect it to the Schmitt Trigger.

I really appreciate the guidance.  Thanks.  I think this one is going to work well.  I plan to fit the whole switcher into the empty bay on the right side of the above rack, with 10 effect loops and use a 25 pin serial cable to the connect to the foot switches.

See you around.  Let me know if you need anything.

Chuck