popping on diode clipping switch

[fuzzy645]

I know there are a ton of threads on the dreaded "popping" when enaging / disengaging the stomp box via a true bypass switch. I found the info on GEOFEX site and a bunch of threads here.

My issue (i think) is slightly different in that I was playing around with diode switching (from LEDs to 1N914's) to vary the distortion sound and it works great.  The only problem is you hear a healthy "pop" when switching from the LEDs to 1N914's (in either direction).

Grasping at straws, I tried adding a 1M resistor from each leg of this switch to ground, and that did not help.

Ideas?

Thanks in advance!

[anchovie]

Will you ever switch the diode type in the middle of playing something? If not, switch the diodes with the pedal in bypass.

[fuzzy645]

Will you ever switch the diode type in the middle of playing something? If not, switch the diodes with the pedal in bypass.

Yes, I agree that I will never switch the toggle while playing, and a easy solution is to (as you say) bypass the pedal before switching.

That being said, I just wanted to make sure I'm not missing some easy/known solution for this type of thing.

Thanks.

[Fender3D]

...
Grasping at straws, I tried adding a 1M resistor from each leg of this switch to ground, and that did not help.

Ideas?

Thanks in advance!

You don't need a resistor to ground.
You may have 2 way to connect your diodes:
either from signal to ground ala dist+
or in the feedback of any gain stage ala ts9/BMP
So you'll end up having a common node for diodes, and 1 node-per-diode-set going to switch.
Simply connect a high value resistor (1M or more) from each node-per-diode-set and the switch out going to PCB.

Edit:
something like this:

[fuzzy645]

Thanks for your reply.  I will show you what I tried below.  The 1st image is the "before" schematic, and the 2nd image is the "after" schematic showing where I attempted to put the 1M resistor. It did not work though, as the popping is still there.

[amptramp]

Remove the resistor you added, revert to the original circuit, then place the resistor from the centre contact of the switch (left side of the feedback capacitor) to the base of the transistor.  When the switch is between the outer contacts, there is no control of the voltage on the left side of the capacitor connected to the switch.  If it was charged up to the voltage on the LED and you switch to the silicon setting, the voltage on the left side of the capacitor was a LED diode drop from the base voltage (about 1.8 volts for red LED's) and then switches to the silicon diode setting, this charge conducts through the silicon diode until it reaches one silicon diode drop (about 0.6 volts) from the base.  With the resistor from base to capacitor, the capacitor goes to the base voltage and the diodes do not conduct when switching occurs.

Another possibility would be to use a make-before-break switch so the capacitor left side is never disconnected from the circuit.  This at least stops the noise when going from the silicon to the LED setting.  If you are not using a centre-off switch i.e. the switch is never held in a disconnected state, connect the LED's to the centre contact of the switch and just open the switch for LED or close it for silicon.  You can get away with an SPST switch that way.

[fuzzy645]

Remove the resistor you added, revert to the original circuit, then place the resistor from the centre contact of the switch (left side of the feedback capacitor) to the base of the transistor. 

Thanks for your reply, I appreciate your effort.   I tried it, and unfortunately it did not help with the popping, and also had the effect of diminishing the distortion clipping a bit.  At this point I have tried experimenting with a breadboard by connecting that 1M resistor from the common of the switch to just about every point in the circuit with no success.   

[Fender3D]

did you try my advice?

[fuzzy645]

did you try my advice?

Fender3D -

Thanks for your advice too. I meant to post a response to you next but got called away at work.  I did not try your solution yet, as I'm not exactly sure how to implement in this Easy Drive circuit.  Would the 1M resistors go between the .1 uf cap and the switch (to the right of my switch in the diagram), or rather directly to the left of the switch but before each diode pair?

Thanks!!

[R.G.]

I believe the problem is that when you switch diodes, there is an instant change in the voltage on the feedback path around the transistor. This forces the transistor to generate a pop.

Changing diodes causes a change in how much voltage is on the capacitor in series with them. The transistor base is at a relatively fixed voltage, so the collector must supply enough current to charge the capacitor in series with the diodes. That is probably the pop.  Note that this happens regardless of any parallel resistors, resistors to ground, etc., as you've found out.

One way to stop this is to use a slow switch. Replace the mechanical switch with two P-channel JFETs. put a diode-R-C network on the gates and use your mechanical switch to pick which JFET gate is pulled up to +9V. The R-C network can be set to do this slowly enough that no audible pop gets created. The pop is still there, just slowed down so it's below the pass band of the circuit.

There are other, but more complicated ways.

[fuzzy645]

I believe the problem is that when you switch diodes, there is an instant change in the voltage on the feedback path around the transistor. This forces the transistor to generate a pop.

Changing diodes causes a change in how much voltage is on the capacitor in series with them. The transistor base is at a relatively fixed voltage, so the collector must supply enough current to charge the capacitor in series with the diodes. That is probably the pop.  Note that this happens regardless of any parallel resistors, resistors to ground, etc., as you've found out.

One way to stop this is to use a slow switch. Replace the mechanical switch with two P-channel JFETs. put a diode-R-C network on the gates and use your mechanical switch to pick which JFET gate is pulled up to +9V. The R-C network can be set to do this slowly enough that no audible pop gets created. The pop is still there, just slowed down so it's below the pass band of the circuit.

There are other, but more complicated ways.

Thanks so much RG....much appreciated!!  I want to try that for sure.  I will first try and draw something up and post for you to look at to see if I have it correct.

[fuzzy645]

One way to stop this is to use a slow switch. Replace the mechanical switch with two P-channel JFETs. put a diode-R-C network on the gates and use your mechanical switch to pick which JFET gate is pulled up to +9V. The R-C network can be set to do this slowly enough that no audible pop gets created. The pop is still there, just slowed down so it's below the pass band of the circuit.

Thanks again RG.  OK, I tried to draw something up.  I am wondering the following:

1.  Did I interpret this correctly as you described?

2.  I assumed I would go from the outer lugs of the switch to the "drain" of each JFET, and then connect the "source" of each JFET to ground.  Is this right?

3.  Do I have the R-C network correct (one end to the gate, and other end in series with the corresponding diode pair)?

4.  Does it matter if R is to the right of C (as I have drawn it), or should R be to the left of C?

5.  What values of R & C should I try?  Isn't it R x C = time constant so in theory (for example) a 2.2 uf cap and a 150K resistor would slow it down by 1/3 of a second (that is 0.0000022 x 150000 = 0.33 seconds)....or perhaps that is too long?

6.  Is there any particular brand/type/rated JFET I should be using, or will any P-Channel JFET work?

Thanks!!

[amptramp]

I think he intended to have the drain go to the capacitor, the source go to the diodes and the gate controlled through an R-C filter and the switch selecting which gate was off (at high potential).  Most JFET switches turn on when the gate is open (at least in a lot of commercial pedals) such as the Boss pedal here:

http://www.generalguitargadgets.com/boss/BossBD-2.gif

Note that they use n-channel JFETs for switching and they switch off when pulled down.  The gates are allowed to float to switch on.  P-channel JFET;s would switch off when pulled up.

[Mark Hammer]

Here's what I think.

The .1uf cap has to be connected at all times.  The way it's drawn, the connection between the end of the cap that is tied to the switch common is momentarily interrupted.

So here is what I had originally suggested to the OP (and a picture would have helped circumvent this).

Each diode pair remains connected to the cap, and has a 1M resistor in series with it (I imagine smaller values, like 100k, would work too),  So, from the cap we have two parallel diode/resistor paths.  The common of the toggle goes to the junction of the two added resistors, and essentially bypasses the one added resistor, or the other.  So, no paths are interrupted, nobody has to find JFETs or worry about the pinouts, and hopefully there is no pop as a result.

Reasonable?

[R.G.]

Here's what I came up with:

Mark has the idea for switching - leave them connected, but shunt a big resistor. But I think this will still cause pops because of the sudden change in clipping level on the diodes. I could be wrong, but I suspect if you happen to switch at an instant when the signal is not near zero, you still get a pop.

That's what the JFETs are for. They are slow-turn-on resistors. What I was trying to do was to let the JFETs slowly change resistance. The R-C-D networks on the gates let the gate be pulled up by 1.01M or down by 1M. The R-C timing can make the JFET turn on or off over several milliseconds and should reduce pop for the same reason Boss/Ibanez pedals don't pop. The 1M resistors in parallel with the JFETs and the diode pairs keep the DC level the same on both sides of the JFETs and on the diode side of the collector capacitor.

[fuzzy645]

RG & Mark -

Thanks so much for taking the time to draw out these solutions.    I will definitely try both.  For RG's I will need to order the JFETs from Mouser, but for Mark's I will try tonight since I have all the needed parts.

Mark - I misunderstood the original advise you sent me via PM, and the picture you provided makes a lot more sense.  The one thing want to make sure I am getting is how the switch is to be wired.  I have therefore redrawn it below.  Please take a look to make sure I got it right. I will be breaking out the breadboard as soon I hit send

[fuzzy645]

OK, here is an update for you.

Mark - I tried your solution.  Assuming I interpreted the switch correctly (as per my re-drawing) the verdict is the popping has been reduced by about 30% (by my estimation). I was able to A-B the original version against yours. I have the original version hard wired into an enclosure and I have yours up on a breadboard.  In the original version, the popping is invasive. In your version Mark, the popping has been tamed to a reasonable level.  The popping is still there, but not as loud.  I have also been trying to A-B if there is a tonal difference in the overdrive character, but the good news is to my ears the tone seems the same.  Perhaps a teeeny bit darker, but that might be other factors.   I do wonder though if the value of the resistor could have an impact on the degree to whch the popping gets tamed.  It looks to me like those two 1M resistors are in parallel, which would yield 500K.  I wonder if doubling the value to 2M would  help.

RG - as I have indicated I will try out your JFET switching method next, once I get the JFETS.

Thanks again!!

[Mark Hammer]

OK, here is an update for you.

Mark - I tried your solution.  Assuming I interpreted the switch correctly (as per my re-drawing) the verdict is the popping has been reduced by about 30% (by my estimation). I was able to A-B the original version against yours. I have the original version hard wired into an enclosure and I have yours up on a breadboard.  In the original version, the popping is invasive. In your version Mark, the popping has been tamed to a reasonable level.  The popping is still there, but not as loud.  I have also been trying to A-B if there is a tonal difference in the overdrive character, but the good news is to my ears the tone seems the same.  Perhaps a teeeny bit darker, but that might be other factors.   I do wonder though if the value of the resistor could have an impact on the degree to whch the popping gets tamed.  It looks to me like those two 1M resistors are in parallel, which would yield 500K.  I wonder if doubling the value to 2M would  help.

RG - as I have indicated I will try out your JFET switching method next, once I get the JFETS.

Thanks again!!

Thanks for the feedback.  It was a shot in the dark, so it's helpful to have others do the experiments for me! I'm not so sure that increasing the resistance value is the way to go since the whole intent is to provide a means for the .1uf cap to drain off quickly, and higher resistance will impede that (pun intended....or maybe just there as a freebie  )

[R.G.]

Just to clarify - on my schemo, where lines connect, there is a dot. Where they cross but there is no dot, there is no connection.

[fuzzy645]

The common of the toggle goes to the junction of the two added resistors, and essentially bypasses the one added resistor, or the other. 

JFETs will be here Monday so my test with JFET switching will wait till next week.

Mark - quick question for you about your method, purely for my understanding/education.  When the SPDT switch (as I have drawn it) is in the "down" position,  is it correct to say the silicon diodes will be the ones we are hearing at that point, because the silicon diodes really  have 2 paths at that point (one with the resistor and the other without), where as at that point the only path for the LED pair is through that big resistor.  I guess I'm just trying to clarify the exact reason as to why we can only hear 1 diode pair when both pairs are always connected.

Thanks!