VooDoo Overdrive mods for more metal edge.

[biggy boy]

Hi there.

I just finished building this circuit using Paul's Eagle cad layout
http://gaussmarkov.net/wordpress/circuits/voodoo-lab-overdrive/

I made the board without any of the modes Paul suggests and found the sound to be an OK mellow distortion.
But I was looking for a more edgy sound closer to metal, so I started doing mods one at a time.

The first one was to change the R5 from 47K to 4.7K as per Paul's suggestion, this help some to increase the gain, but still not metal sounding yet.

The next change was to add on the 2 up pair of diodes to the second stage of the Op-amp, as suggested by Mark Hammer.
Info of mod found here http://www.diystompboxes.com/smfforum/index.php?topic=64272.0
Adding the additional diodes made no difference in sound, other then attenuating the signal (lowering the volume).
But once I changed R8 from 150K to 220K that made a noticeable difference and allowed the second set of diodes to clip the signal.
The combined mods really give this circuit more edge. It's now close enough to a metal sound for me, I'm really enjoying the pedal
I noted the changes on the digram below, that was originally created by Paul.



Thanks
Paul and Mark for the mod suggestion's.

[punkin]

Looks like a great build...any possibilities of hearing a sound clip? Thanks for the detailed write up of the mods.

[anchovie]

You could take out the gain pot and replace it with a jumper straight to ground which will put the first stage at its maximum gain. Then replace R8 with a 500k pot (hey look, you've got one spare from the first stage!) as a variable resistor in series with the original 150k resistor and your range of distortion will start with the previously-maxed-out sound, through your 220k setting and beyond!

[aziltz]

interesting.  looks like a Rat/OCD/Voodoo Lab Clipping stage running into something similar to a TS clipping stage.  Hard Clipping to Vbias into Soft Clipping.

Kudos for mucking around with the circuit!

[biggy boy]

You could take out the gain pot and replace it with a jumper straight to ground which will put the first stage at its maximum gain. 

Hi anchovie
I like the gain pot on the first stage, with It I can dial it to any level of distortion from very mild to close to metal.
the second set of diodes only add a bit more distortion. The first set are doing most of the grunt work. I added a switch to take the second set of 2 up diodes in and out of the circuit.
I'm thinking of lowering the value of  R9 from 47k down to 4.7K to see if that pushes the second set of diodes harder!?!?!?!?

interesting.  looks like a Rat/OCD/Voodoo Lab Clipping stage running into something similar to a TS clipping stage.  Hard Clipping to Vbias into Soft Clipping.

Kudos for mucking around with the circuit!

Thanks aziltz
I'm really enjoying using it.  Was playing some  Black Sabbath stuff Iron Man....
If you dial back on the gain it gets more tubish sounding.

Looks like a great build...any possibilities of hearing a sound clip? Thanks for the detailed write up of the mods.

Oh man sound clip, never posted one before any tips on posting clips?
where would one place/store the clip?
I keep my pictures at Photobucket?

Glen

[punkin]

Here's a great place to post your samples...just upload them in MP3 format and you'll get a URL link in your email.

[Mark Hammer]

I posted a mod and explanation for you last night, but I'm not sure what happened to it.  So here goes again.

Take a 27k resistor and .0047uf cap in series, and place that pair in parallel with C6/R9.  That will up the gain for content over 1.2khz, in a similar fashion to what gets done for the Rat, although not quite as extreme (x9.1 vs x5.7 for lower content).  Should make it a little easier for the unwound strings to zing as hard as the wound ones, though.

[biggy boy]

Hi Mark
Thanks for the added info, will give it a try tomorrow night after work.
I guess with the two resistors R9 and the new one in parallel, this will drop the total resistance to ground to 17.1 K
and the two caps in parallel would increase the total capacitance Right?
I'll wire it to the switch so it comes into the circuit when the 2 up diodes are switch on.

For the soft clipping section, If I was to make them just one diode pair instead of two in series this would lower the headroom and cause more clipping,
but it would also attenuate the signal more is this correct?

Here's a great place to post your samples...just upload them in MP3 format and you'll get a URL link in your email.

Oops no link


Glen

[punkin]

 

Oops...try this;

http://www.lightningmp3.com

[Mark Hammer]

Hi Mark
Thanks for the added info, will give it a try tomorrow night after work.
I guess with the two resistors R9 and the new one in parallel, this will drop the total resistance to ground to 17.1 K
and the two caps in parallel would increase the total capacitance Right?

Not exactly.  Remember that, in a non-inverting op-amp, the amount of gain applied is an inverse function of how much negative feedback is applied from the output back to the '-' pin.  The op-amp "wants" to go full tilt, and everything you do simply adjusts how close to that maximum it gets. 

The feedback resistor and the path to ground from the inverting (-) pin act like a volume pot in terms of how they regulate the amount of negative feedback that is applied, with the inverting pin being like the wiper of that pot.  If very little negative feedback is applied (feedback resistance is MUCH bigger than path to ground), than it's like taking your foot off the brake of a runaway vehicle heading downhill.  That is, of course, why sticking a cap in the feedback path rolls off high end - you have provided an efficient path for lots of negative feedback in the upper frequencies to go from output back to inverting pin.

When two parallel paths to ground are supplied, the cap in each of those paths determines the point/range above which negative feedback will be reduced.  By having that second path with the 27k resistor, you are providing a way to "dump" more of the feedback to ground.  Just like calculating the attenuation provided by a pot, we see that 27k is a smaller fraction of the feedback+groundleg resistance  [220k+27k]/27k  than 47k is ( [220k+47k]/47k ).  However, the .0047uf cap in series with the 27k resistor means that this "feedback dumping" change can only be applied to content that easily passes through that .0047uf cap to ground.  What easily passes through it?  That would be content above 1 / (2*pi*R*C) = 1/(6.28*.027*.0047) = 1254hz.

In contrast, the other path (47k/2.2uf) provides a fixed amount of gain for all content above 1.5hz, so essentially flat.  Injecting a bit of animism into the picture here, if YOU were a 200hz wave, where would YOU choose to go to ground through?  remember that caps work like frequency-dependent resistors, providing a higher impedance path the higher the frequency.  The .0047uf cap makes it hard for you, so you'll "choose" to find your wayto ground through the 47k/2u2 network.  However, if you were a 2khz wave, you'd have a "choice" between the 27k/4n7 path and the 47k/2u2 path.  Following the proverbial "path of least resistance", you'd take the 27k/4n7 path.

So, in this way the dual network provides fixed gain which is flat below a certain point, flat above another point, and sloped/graded, between those two points.  In the Proco Rat, this strategy is used to provide substantially more gain for content above 1.5khz.

For the soft clipping section, If I was to make them just one diode pair instead of two in series this would lower the headroom and cause more clipping,
but it would also attenuate the signal more is this correct?

Yes, you've got it right.  Although it would be fuzzier, the final output would not be any greater than what is seen at the output of stage 1, simply because the diodes set the same absolute ceiling in each stage.  That's why you needed to increase the gain in stage 2 to get additional clipping with the dual-diode pair.  A compromise is to use a 2+1 diode complement in stage 2, such that you get the added headroom for one half-cycle, but the enhanced double-clipping for the other half cycle.  This gets you fuzzier-but-louder.  Another compromise is to use a dual pair of Si+Ge for a clipping threshold that is higher than a 1+1 Si complement, but lower than a 2+2 Si complement.

Keep in mind that overall output volume is helpful for extracting additional clipping from the amp, whether in the preamp or power-amp stage.  So, depending on your tastes/needs, you can place the emphasis on pushing the pedal to sizzle more, or letting the pedal sizzle a bit less but "hurt" the amp more.

[biggy boy]

OK
I think I'm learning something here!

The first set of hard clipping is clipping the output from the first stage.
The second set of clipping the "soft" clipping is clipping the negative feedback loop of the second stage.

Now not positive about this one but will give it a try:

-The cap and resistor R9 act like a tone control a non adjustable one. It controls tone similar to the tone control on a guitar.

-If you were to take the cap out it would drain a portion or all of, all the frequencies in the feedback  loop to ground and lower the gain of all of those frequencies.

-The cap size determines which frequencies get drained to ground and not sent to the - pin as negative feedback. Is this correct?

-Increasing R8's value causes more of the feedback signal to travel through the clipping diodes?

The part that still puzzles me is the R5 and R9 resistors. When I lowered the value of R5 to 4.7K it increased the gain of the first stage.
When you think of resistance One might think that lowering R5 would cause more of the signal to drain to ground and less of it would go through the feed back into - pin thus lowering the gain and distortion???

[Mark Hammer]

OK
I think I'm learning something here!

The first set of hard clipping is clipping the output from the first stage.
The second set of clipping the "soft" clipping is clipping the negative feedback loop of the second stage.

Yes and no.  The clipping is every bit as "hard", but yes, they are each clipping the signal at different points along the path.

-The cap and resistor R9 act like a tone control a non adjustable one. It controls tone similar to the tone control on a guitar.
-If you were to take the cap out it would drain a portion or all of, all the frequencies in the feedback  loop to ground and lower the gain of all of those frequencies.
-The cap size determines which frequencies get drained to ground and not sent to the - pin as negative feedback. Is this correct?

More or less.  You have to ask yourself "Is signal at frequency X going to make its way fully to the inverting pin, or will much of it be lost to ground.  Remember that if you have a simple wire link from the output to the inverting pin (from 7 to 6 or from 1 to 2), it doesn't really matter what the resistor/cap combination to ground is because ALL the feedback signal goes to the inverting pin and the gain is 1x.  If the feedback resistance is greater than 0 ohms, then you will lose a certain amount of that feedback via the path to ground.
As correctly deduced, the cap value does determine what freqs drain off to ground.

-Increasing R8's value causes more of the feedback signal to travel through the clipping diodes?

The diodes do nothing until the signal reaches the forward voltage set by the sum of the two diodes running in each direction.  So, if D6 has a forward voltage of 532mv and D7 has a forward voltage of 516mv, then nothing will flow through D6/D7 until that half-wave exceeds 1048mv (532+516).  At that point the diodes conduct.  Of course, once they conduct, then they provide a more efficient pathway than the 220k resistor.  When that happens, gain is reduced via the application of negative feedback from pin 7 to pin 6.  At that moment when gain is reduced via the diodes, R8 once again resumes its role of determining the gain of that stage.  Think of R8 like the typical TV sitcom dad going "Hey, let's party!".  When the diodes conduct, they poke their head in like the TV sitcom mom and say "What's all this racket in here?", at which point things temporarily quiet down.  But once mom closes the door, things go crazy all over again.

The part that still puzzles me is the R5 and R9 resistors. When I lowered the value of R5 to 4.7K it increased the gain of the first stage.
When you think of resistance One might think that lowering R5 would cause more of the signal to drain to ground and less of it would go through the feed back into - pin thus lowering the gain and distortion???

That's the part that a lot of folks have a hard time wrapping their head around, because its a double negative.  Again, think of the op-amp as being like a car with the gas pedal permanently floored.  The only way for you to adjust the speed of the vehicle will be to use your brakes.  So, if you either make the feedback resistor greater, or make the resistor to ground smaller...or both, you will be losing feedback signal, and essentially applying the brakes less.  That's why making the path through R5 more "efficient" increases gain - you're losing feedback that would nromally suppress gain.

[biggy boy]

OK good stuff thanks, this is great.
I love the analogies!
Hope I'm not driving you nuts!!

OK lets see if I understand.
-The op-amp's second stage is the floored gas pedal wanting to let as much signal through as possable.
-The negative feedback loop is acting like a brake. We don't necessarily want to slow down on the gas pedal but to get distortion we need to apply the brakes.
The brakes (negative feedback) is what creates the distortion along with the clipping.

In non analogise terms:
We use a negative feedback loop to enhance distortion but the loop decreases the gain and output level.

So I guess it's sort of a balancing act, a give and take situation between output and distortion levels.


Glen

[Mark Hammer]

1) What I described is how ALL noninverting op-amps work.  They ALL want to go full tilt.  We set the "gain structure" of a circuit that uses multiple op-amps by applying a little more brakes here and a little less there.

2) If we apply NO brakes whatsoever, the op-amp will produce what is referred to as "open-loop gain", which takes us into the range of gains of thousands - a recipe for oscillation and other lunacy.  So, we apply the brakes fairly hard in the vast majority of instances.  For instance, take your average JRC4558.  The datsheet says that if there is nothing in the feedback loop between output and input, there will be a default gain of around 3100x for 1khz and just under 32,000 for 100hz!  You can find this information by looking at the "gain-bandwidth product" or the "open-loop voltage gain" chart in the datasheet for most op-amps.

3) The distortion is produced by the "hold on a minute there, not so fast buddy" action of the diodes.  The signal itself may continue higher than the peak permitted by the diode/s, but once the diode/s start to conduct, that potential peak is held at the level dictated by the diode/s.  In other words, what you hear as distortion is simply the manner in which the diode/s limit the headroom for the signal. 

If you had NO diodes whatsoever, and goosed the daylights out of the signal, you would still hear distortion, but from other sources that also constrain headroom.  For example, if you power an op-amp with a 9v battery, and feed it a 100mv p-p signal, how much gain can you apply to that signal before the op-amp says "Nah.  Can't do it."?  For a JRC4558, the datasheet says that with a +/-6v supply (i.e., 12v with a Vref midway), the maximum voltage swing from one peak to the other would be limited to about +/-4v.  Extrapolating from the graph, it looks like a 4558 powered with a 9v supply (Vref at 4.5V) craps out at around +/-2.5v.  In other words, our 100mv p-p input can't get any bigger than 5v p-p without suffering some headroom limitations.  Dividing 100mv into 5v, we have a practical limitation of a gain of around 50x before the chip says "Nope.  Can't give you more signal than that.".

We use a negative feedback loop to enhance distortion but the loop decreases the gain and output level.
So I guess it's sort of a balancing act, a give and take situation between output and distortion levels.

4) The gain merely brings the signal up to the absolute level where the diodes can be brough into conduction.  If I replaced the 2+2 diode pair with 3 LEDs, end to end, running in each direction, pretty much all the clipping you would hear would be due to the chip and the clipping in stage 1.  Why?  because there is no way in hell that a 9v supply and any amount of gain is going to bring the signal up to the needed level of around +/-4.5 for those LEDs to do their thing.  They're like the cymbal player in the orchestra who sits around waiting for their brief moment that never comes.

If the clipping threshold (the point where the diodes conduct) is low enough, then you don't need to have much gain applied to the signal to bring it up to that point.  If you have less gain applied, of course, the maximum output will be lower.  If you set the threshold higher using more or different dioes, you can still apply more gain to produce clipping at whatever threshold is dictated by those diodes, but only insomuch as the chip and supply voltage permit.

But in general, yes you are correct, it IS a tradeoff between how much volume you want at the output jack, and how much clipping you aim for.

5)  The perceived intensity of diode-based clipping is a function of how much of the note's lifespan falls above the clipping threshold, and how much of the "period" of the waveform falls above that threshold for that sensitive portion of the lifespan.  Plug a guitar into a scope, play a note, freeze the image, and draw a horizontal line across the screen.  Some of the peaks will be above the line andmany will be below.  In general, the height of the peaks is maximum at the point where you pick, and decline quickly after that.  If your clipping threshold is high enough, only the first couple of wave peaks will actually be clipped, and the remainder of the sustaining note will be just fine.  Lower that line and the clipping will take place not just at the outset of the note, but a little while longer too.  Generally, what we think of as "fuzz" is something where the clipping is applied across almost the entire lifespan of any held note, and what we call overdrive refers to that situation where only the initial peaks are clipped.

6) The Big Muff Pi is a double clipper such as what you're working towards here.  Instead of op-amps, it uses transistors to provide gain.  And instead of clipping differently in the two stages, it clips, then boosts and reclips in exactly the same manner.

[biggy boy]

 
WOW that was great I'm starting to get it. Cool!!
I really like learning this stuff, I'm not satisfied just building something I like to know how it works too.

Thanks Mark

If you have more to teach us (me) go for it.
I'm eating this info up like an all you can eat buffet.


Glen

[biggy boy]

Another question:

The voltage divider is used to set set a bias I take it these Op-amps like to be biased at half the applied voltage in this case 4.5 volts.

I see that the first set of diodes are connected to the V reference (half the applied voltage 4.5v) why are they connected there as apposed to ground?



Glen

[Mark Hammer]

Another question:

The voltage divider is used to set set a bias I take it these Op-amps like to be biased at half the applied voltage in this case 4.5 volts.

I see that the first set of diodes are connected to the V reference (half the applied voltage 4.5v) why are they connected there as apposed to ground?

Glen

Since it is a signal that swings positive and negative, there has to always be a middle reference point.  It could be whatever you want it to be, but for maximum possible headroom on either peak (pos/neg), you set the reference point to the middle of the two possible signal extremes: 0v and 9v.  As noted earlier, that does not necessarily imply there will be loads of headroom, since there will always be limits to how wide the signal can "swing" using supply voltage X.  But, within limits, using the midpoint of whatever the supply voltage is (9v, 12v, etc.) will optimize headroom for both negative and positive swings/peaks.  Besides, if you set the Vref for the "middle" - whatever that might be - then as your battery starts to decline from 9 to 8.5 to 9.0, etc., your midpoint will move along with it.

As a Canadian afternote, I might add that many fellow canucks have strongly recommended (tongue in cheek, of couorse) that the national capital be relocated to Winnipeg...because it is in the "middle".  The idea being that a truly centrally-located capital would be less biased in favour of the east, west, north or south of the country.

[biggy boy]

As a Canadian afternote, I might add that many fellow canucks have strongly recommended (tongue in cheek, of couorse) that the national capital be relocated to Winnipeg...because it is in the "middle".  The idea being that a truly centrally-located capital would be less biased in favour of the east, west, north or south of the country.

That would certainly make the westerners more happy, But think how that would make Quebec's feel 
Then they would really want to separate more LOL (tongue in cheek)

[biggy boy]

I have completed the additional mods the circuit by adding and extra resistor and cap to ground in parallel to R9, C6.
I can now notice a bigger difference with this latest mod, more gain.
This thing is wicked now!

I used a switch to put the second stage of diodes and the additional resistor/cap to ground into the circuit.
I can now go from close to clean, by turning down the gain, turning the switch off and lowering the volume on my guitar.
Or I can crank everything up, flip the switch and get a fuzzy sustain monster.

The switch is not mounted to the board.

Ive included the upgraded schematic called the Voodoo double overdrive.
If anyone would like to make and post a Veroboard (Renegadrian) , that would be great.





Thanks Mark for the help

Glen

[biggy boy]

OK
Finally got some sound samples of the VooDoo with the second clipping stage added.
This is the first time I've ever recorded myself OH it's scary but here it goes!

Recorded into my DigiTech JamMan with:
  VooDoo OD into clean AMP with a little reverb.

First recording is with gain set to zero

http://www.aronnelson.com/DIYFiles/up/no_gain_metal.mp3

Second recording gain set half way

http://www.aronnelson.com/DIYFiles/up/medium_metal.mp3

Third recording gain at max

http://www.aronnelson.com/DIYFiles/up/metal1.mp3

Another one with gain at max

http://www.aronnelson.com/DIYFiles/up/metal2.mp3

I hope that didn't make any one sick to there stomach 

Glen