Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

March 28, 2024, 11:00:50 AM

Login with username, password and session length

Recent Posts

 

Switch pop blues

Started by Bajaguy, May 24, 2013, 12:25:22 PM

Previous topic - Next topic

Bajaguy

I've been screwing around with a few different preamp ideas over the last couple of months, and I decided to build a dedicated "test mule" tda2003 amplifier board to try out the different bread board circuits on. I bought a amp kit through one of the online places, and it works fantastic. I'm very pleased with the sound of the amp, but I've had other problems that are sending me into mental overdrive :duh

One of the problems that I have been unsucessuful in fixing is a switch pop issue. I have a protoboard that has a modified MXR distortion plus and a mosfet clean boost on it. I've played it as a pedal, and it works great in front of everything I've plugged it into so far. I got this idea fron one of Juan's posts and tried to add it in as the preamp for this little amp, and the switch pop is driving me nuts. I originally designed the circuit to use 2 different 3PDT switched to turn each effect on and off separately, so when I decided to add it on to the front of my test mule (I call it Pepe), I used a couple of 9V relays I got from one of the online places to do the switching so I could use mini toggles or a foot switch to switch the effects on and off. The footswitch works great (and is quiet), and the switches work, but create a nice fat pop every stinking time. I added a 1uf cap across the relay coils and a diode to keep from blowing it up, but I'm still getting that pop from the front panel switches.

I'm kind of lost here, but I figured I'd throw it out and see if the amplifier gods had another tidbit of info that might keep me from pulling my hair out :grr

Thanks ahead for any help,

Baja

J M Fahey

I *suspect* you are switching DC together with your signal, and that pops big way.
Please post the schematic of what you are *actually* switching.
At least show the output of each preamp, the relay/switch contacts and the input of the power amp.
Capacitors in theory do not pass DC, unless grossly bad, but if left "floating" they have no ground reference and any tiny tiny tiny loss current charges them to some voltage.
If you do not provide some high resistance path to ground, the floating end can and will have *some* DC in it ... which will pop when switched.

Bajaguy

I'd love to, but photobucket seems to be having a melt down with it's uploader :grr

I'll try to find another photohosting place and get it up asap.As for photobucket... :trouble

Baja

Bajaguy

#3
Ok, here's a couple of schematics that should help. The first is a schematic of the clean boost circuit I used, and the second is schematic of my relay board and how the circuits are wired up.

The only other schematic is the MXR distortion +, but I'm not sure if I can post it without getting in trouble (it's got another website's schpeel on it).



[



Baja

J M Fahey

Quote from: Bajaguy on May 24, 2013, 04:21:25 PM
The only other schematic is the MXR distortion +, but I'm not sure if I can post it without getting in trouble (it's got another website's schpeel on it).
Well, you have already *built* it, haven't you?
So if you worry about intellectual property it's sort of late, isn't it?  :lmao:

No, don't worry but anyway don't post *the image* here, just the link.

The point is that you must switch between points that are always at the same voltage, in this case ground, so each of them needs a 1M resistor to ground.

The FET booster already has R11 and R17 .

Don't have the MXR schematic, I seem to remember the volume pot is straight to the output jack but th input cap is floating. Add 1M to gtound.

And do the same at the TDA2003 input.


Bajaguy

My MXR clone already has a 1M to ground just before the cap. It's the only mod I did to this one, and it solved the pop issue when I had it in the stomp box.

I'm going to throw a .1uf cap across the switch contacts tomorrow and see if that doesn't tame the beast a bit. I had a delay pedal a few years ago that I had to solder in a 1M between the input pad on the PCB and a spot I cleared off on the ground plane to get it to stop. That one had a different pop than this one, not quite as solid but alot of that is probably the different speaker I was using then.

Baja

Roly

Well the signal path looks okay with the switch sides tied back to ground via resistors, but the relay control circuit ain't right.

Rewire C1 and C2 with a 100 ohm resistor in series with each, and across the relay coils, like the diodes.

Where they are they will charge up to full relay supply and when you close any of the activating switches will discharge with an almighty surge current and contact arc (well, "almighty" in signal terms anyway).

By rewiring them as CR snubbers across the relay coils you will suppress the inductive kick-back from the coil when the relay is switched off.

"but ... but ... I thought that is what the diodes are for..."

Well yes, but there's a gotcha.  When the switch tries to break the relay coil current the back EMF from the energy stored in the magnetic field causes a spike of kilovolts to rise instantly (that is I had to use a 100MHz CRO to see that the catch diodes have a small but finite turn on time, and until they start to conduct the voltage goes way north for an instant, many kV - causing havoc with the state logic system I was building).  The CR snubber slows the rise time of the back EMF long enough to the catch diode to actually catch the spike.

The primary need for catch diodes is to protect a driving transistor from being destructively punched through by the stored energy as it's the total stored energy, not just the voltage, that damages the transistor (which can cope with being zenered out for a few microseconds, but not for quite a few milliseconds).  The series resistor is required to limit the charging current through the switching transistor, or switch contact in this case, so it isn't damaged by the charging surge into the cap during switch on.

It is worth keeping in mind that the charging or discharging current with a simple cap and switch is only limited by the minimal resistance and inductance of the loop and may amount to many amps for a few milliseconds, and is not healthy for the switching device.  Switch contacts also do not simply open or close; when closing they will bounce causing a string of closes and opens, and when opening will strike and arc, albeit microscopic, but none the less a pinprick of super hot plasma that will melt and pit the contact surface.

If the relays run directly form the same 9V supply as the audio circuits you could also isolate them with a suitable small resistor and local bypass electrolytic.

HTH
If you say theory and practice don't agree you haven't applied enough theory.

Bajaguy

That makes sense. I didn't think about the diodes not catching that spike. I'll drop in a couple of resistors on Tuesday when I get back to the shop.

Thanks!

Baja

Roly

...and rewire the CR circuits in parallel with the catch diodes.
If you say theory and practice don't agree you haven't applied enough theory.

Bajaguy

#9
OK, sounds good! One question, how big should my 100 ohm be? I figured everything for current @ 9V, but if there's a large spike coming through, should I up my resistor size to 1/2W?

Baja

Roly

Quote from: Bajaguy on May 26, 2013, 06:01:43 PM
OK, sounds good! One question, how big should my 100 ohm be? I figured everything for current @ 9V, but if there's a large spike coming through, should I up my resistor size to 1/2W?

Baja

Big enough to fit between its leads?    ;)

Current only flows while the 0.1uF cap is charging or discharging and unless you are doing Riverdance on the stomp switches that won't be very often.  And even if you are doing Riverdance the total amount of current flowing will still be quite small.

With the cap discharged the voltage across the series resistor will be the full supply, 9 volts.  assuming a 100 ohm resistor this will result in a current of 9/100 amps, or about 10mA.  The instantaneous power is E * I or 9 * 0.01 watts, call it 100mW, but it rapidly decays as the C charges.  How rapidly?

The time constant, tau, of 0.1uf and 100 ohms is CR, 10-7 * 102 seconds, or 10-5 seconds, that is 10uS for it to fall to 33% of the initial value, or about 33mW.  After ten time constants the current is effectively zero.

You would need to stomp the control switch around 10,000 times a second to get about half the initial power, or 50mW.

So use whatever you have to hand.
If you say theory and practice don't agree you haven't applied enough theory.

Bajaguy

I got some time yesterday to redraw the schematic, but the repair shelf was full so I ran out of time before I could try the fix. I should have it done this afternoon so I'll post up the results after.

Here's a copy of the schematic with the fix added:




Baja

Bajaguy

Well, I got some time this afternoon to drop in the fix. Although there was a reduction in the volume, there is still a pop from both relays. I wonder if changing the .1uf caps to a larger value might be necessary?

Baja

Roly

Could be; you could try up to 0.47uF, but ... if you can get the cover off one of your relays and operate it by finger and see if you still get the pop.  While it appears that the switch contacts should be equi-potential it would be good to know for sure if it really is coming from the control side before ripping into it too much (or your hair out too much).  If the relays are sealed you might be able to substitute a switch for the test.  If you still get the pop then obviously it's on the signal side somewhere and there is not much point in treating the control side.  If it turns out to be quiet and likely on the control side, then you'll need to try decoupling the relay power supply with a resistor and electro as previously mentioned.
If you say theory and practice don't agree you haven't applied enough theory.

J M Fahey

^^^^^^ COOL idea.

This guy Roly should start writing in some Guitar related Forum.
.....

And why not? even someday have a Forum of his own.

Just sayin ;)