Fundamentals of designing your own overdrive?

[canman]

OK guys, I pieced together a little schematic.  The values are ALL WRONG, I haven't done any calculations yet to see what values I should use.  I just wanted to see if I have the right pieces in the right places.  Basically, I have a low pass and high pass filter right after the input.  Then a gain stage with a dual opamp where the gain control is located.  After that I have clipping MOSFETs (no idea if this is where they're supposed to go) followed by a tone stack, pretty much cloned from the Triple Wreck.  Treble, Mids, and Bass.  After that I have the recovery stage with the second half of the opamp, and then a volume control.



I know I'm missing pieces, and I haven't calculated out any values or impedance or anything like that.  But with this schematic, would I be able to go in, calculate things out, tweak a few things, and have something that might sound decent?  Or did I forget something crucial?

[GGBB]

The first op-amp is biased to 9V - it is normally +4.5V for +9V devices. The second op-amp has no bias and no feedback. Your clippers are positioned before your DC blocking cap and they are directed to ground, so they are biased (the same as the first op-amp) which means only one will actually clip (recent discussion here). Move the clippers to after C5.

This is very close to an OCD. You might be better off just starting with the OCD circuit and adding your tonestack and any other changes you might want.

[canman]

Right, I'm a little hung up on the biasing.  I know it needs to be 4.5v, so should the 4.5v be coming in at the + leg on the first opamp?  I have both listed at 9v, but I thought with the voltage divide on the left there it would be around 4.5v (assuming I use the right resistor values)?  With the second opamp, I totally forgot to bias it...but why does it need anything in the feedback loop?  Or is the feedback loop was causes the amplification?  Would I just put in some sort of filter there to amplify certain frequencies?

With the DC blocking cap...I did not know you needed them in multiple spots!  Very good to know.  I thought you just needed one at the input to prevent DC from going to your guitar.  I'll move the clippers for sure.

I was following the OCD schematic a little bit when I put this together, mostly referring to it to see where filters were needed, etc...seems like most of the circuits are pretty similar after looking at them all for too long, haha. 

[GGBB]

I just read the 9V part and didn't look closer - its actually biased at 6V as you've got three 220ohm resistors between 9V and ground (R6->R2->R1) and the op-amp input is after the first one. Of course, as you said, the values are way off. Look up noiseless biasing (or have a look at the OCD).

Second op-amp doesn't have to be biased in the typical way. It can be biased by following another biased op-amp without any caps in between (there's a name for that which escapes me at the moment). The DC bias for the first op-amp passes through and biases the second op-amp (look at the OCD).

The negative feedback loop determines the gain for a non-inverting op-amp. Without something there it uses open-loop gain which is sort of like infinity.

You don't need multiple DC blocking caps - you need just as many as the circuit design requires (the OCD uses two only - for input and output). 

[canman]

Yeah I just stuck with the stock values for now...I want to get everything in the right spot before I calculate everything so I don't get quite as confused!

With the DC blocking caps...I thought that in this case, C1 would be the coupling capacitor, and also functions as a "pre-EQ" so to speak...part of the high pass filter.  Is C5 the coupling capacitor in this schematic?  And if so, is this because the voltage is supplied after the initial filters (basically, at the opamp instead of at the input)?

Looking at the schematic again, I'm having another thought.  Instead of moving C5 to before my clipping MOSFETs, could I put the MOSFETs in the feedback loop of the first stage of the opamp?  Or would that essentially be the same thing as what I've got right now?  I figure if the MOSFET clipping is in the feedback loop, it might give a different flavor then clipping the gain after the opamp.  But I know next to nothing about all this, so that idea may be redundant.  But I'm thinking maybe putting the clippers between R3 and R4 (gain pot)?

Thanks for all the help, this is really informative for me!!  I'm looking forward to breadboarding something!

[GGBB]

All capacitors block DC. A coupling capacitor is merely one that connects two circuits or circuit sections together. Most pedals have at least two coupling capacitors - at the input and at the output - in order to block DC from leaking out of or into the pedal circuit (you have them - C1 and C13). They often have more to block DC (more to it than just that but that is the main reason for their use in pedals) wherever the circuit requires.

C5 is a coupling capacitor because it connects the first op-amp with your tone stack and the second op-amp. You want to block DC from your tone stack because adjusting pots that have both audio signal and DC current present makes unpleasant noises.

Where the op-amp gets it bias voltage from doesn't technically matter, as long as it gets the right voltage, but there may be other factors as to why you might put it at a certain place and usually the determining factor is making sure it is the right value.

Moving the clippers to the feedback loop would make for very different clipping characteristics - you may like it, you may not (have a look at the tube screamer, and look up hard-clipping versus soft-clipping).

[canman]

Oh yeah OK I know what you mean...kind of like how the Box of Rock circuit has some scratchy pots (either on gain or volume, maybe both, don't remember)...I don't know if that's from DC getting to the pots, but I think I know the sound you're talking about.

So when you add a coupling capacitor, you need to choose a value such that the resulting frequency won't affect your tone, right?  Little bits here and there are starting to make sense now, haha.  I spent the past couple of hours putting together a little reference spreadsheet of RC frequencies for bass-mids-treble.  I hope this makes it easier to choose a capacitor value that won't mess with the guitar tone, and for sure help when choosing low and high pass filters throughout the circuit.

(If anyone wants a copy of the frequency chart or would like to double check my work, let me know, I'm happy to send it around)

Clipping is starting to get a lot more interesting for me after looking at hard vs. soft clipping, and all the possibilities you can choose.  Heck, I may even throw FETs in the loop and then put some after the loop, just to see what it sounds like.  Might not sound good, but the idea of testing out different ways to achieve clipping sounds like a blast.

[antonis]

So when you add a coupling capacitor, you need to choose a value such that the resulting frequency won't affect your tone, right? 

We've discussed it before but you have to undestand some item "properties"...

It's rather unsual for a capacitor to "stand alone" (meanin that almost allways it interacts with some neighbor resistance - plus it's own frequency depentable one) so its also a good opportunity to create a desirable filter - and save some extra items...

If you want the capacitor for strictly decouping propose you'll have to compromise between size (value) and cut-off frequency...
(you don't have to worry about the cut-off frequency of a 47000μF but you have to find where to place it and how much to pay for it..)

[canman]

Where's the "smack myself in the face" smiley when I need it??

For whatever reason, I keep forgetting that caps always interact with resistance and create filters, and I can't kick the idea of "using cap values at a frequency that won't affect tone."  It must have been from some of the reading I did...it wasn't specifically related to guitar pedal circuits, so I can see how in other audio applications, that might make sense. 

So essentially, we always want to be as efficient as possible and use the capacitors for as much as we can (ie, coupling capacitor AND low/high pass filter).  Let me ask this then...say at the input of the circuit you have the coupling cap, and a high pass filter with a cutoff frequency of say 82 (low e frequency on guitar)...this means that all frequencies above 82 will pass through the filter and get amplified at the first gain stage.  Whether or not this translates into a "transparent" portion of the circuit, I don't know...we can discuss that later.  But, in my VERY rough and seemingly OCD-cloned schematic above, C5 needs to block DC from the tone stack.  No problem there, but since C5 WILL make a filter whether I like it or not, is it best to set the filter to be the same as the initial high pass filter (assuming I don't want the input of the tone stack to be some different frequency)?  I hope that makes sense...it does in my head, so if y'all can't understand what I'm saying let me know and I'll clarify.

[GGBB]

is it best to set the filter to be the same as the initial high pass filter?

"Best" is entirely dependent on the goals of the circuit - in other words how you want it to sound. It can be summed up by saying you want to either let all frequencies through or let only some frequencies though. In an overdrive, often the input cap is used not only for decoupling but also to let only some frequencies through (i.e. cut some bass) in order to define how the first stage will sound. The same technique can be used with the coupling caps between distortion stages in order to define how the following stage will sound. Tone changes via filters sound different when applied before distortion than the same filters after distortion. This has to do with harmonic content being added through the distortion process.

As you are probably realizing now, designing an overdrive with a certain sound in mind is actually fairly complicated even though the circuits themselves aren't that complex. There are many places to effect a change and more than a few ways to do it, and each change interacts with all the others - it quickly adds up to a seemingly infinite number of possibilities. All this is why there can be so many variations on basically the same circuit - each one having its own distinct flavor.

Truthfully, the only way to determine what is best is to experiment - you need to breadboard the circuit and try it out.

[canman]

I'm getting more and more excited to start breadboarding stuff!!  I can see the need to bust out the breadboard...frequencies may look good on paper, but like you said, things change before and after distortion.  I don't know if you can calculate harmonic content (I'm sure you can) but it seems like breadboarding things may be easier and give you a better idea of what frequencies to shoot for.

So breadboarding...can I breadboard just a simple high-pass filter, hook up my input, and just audio probe it until I find a combination I like?  And likewise, can I add the first gain stage, audio probe it, make changes I see fit, and continue in similar fashion until the circuit is "complete?"  Or do I need to breadboard the entire circuit and then go through and audio probe each stage to see where changes should be made?

Heck, I don't even know if high/low pass filters need power...I guess not, since it seems like that's more or less what an electric guitar circuit is..?

[canman]

In addition to my last post/questions...is this schematic right?  I traced it from the boost portion of the Triple Wreck on tagboardFX, just to see if I could trace a layout into a schematic...



And here's the schematic I traced:



Did I do this correctly?

[MaxPower]

I hate vero because it's a pain to convert to a schematic. In any case, it looks like C3 should be connected to the collector of Q1 (below R5) and... there may be other errors.

I just wanted to say, if you are interested in learning how to design your own circuits and so on, I highly recommend Malvino's Electronic Principles and Transistor Circuit Approximations.

Also, check out LTSpice and/or Tina-ti. They are free circuit simulation programs. LTSpice is pretty popular. I found Tina-ti easier to learn.

[highwater]

I hate vero because it's a pain to convert to a schematic. In any case, it looks like C3 should be connected to the collector of Q1 (below R5) and... there may be other errors.

Yes - you have C3 connected to the wrong end of R5, and you have a jumper drawn across the "Cream" pot. The rest of your trace is correct.

Also, check out LTSpice and/or Tina-ti. They are free circuit simulation programs. LTSpice is pretty popular. I found Tina-ti easier to learn.

+1. There is even a sub-forum here for circuit simulation, the existence of which is a far better endorsement than I can give.

SIMetrix is another good one - it's what I started with. I'm transitioning to LTSpice at the moment, because it can use .wav files as input/output, but I have already found several things that would have confused the crap out of me if I hadn't learned an easier program first. I haven't used Tina-ti myself, but I probably would have tried it first if I hadn't already somewhat-learned SIMetrix by the time I heard of it. If you already have experience with CAD in general, and with command-line-software/scripting/programming, LTSpice would probably be manageable to learn... but otherwise I would recommend something else. They're all based-on the same underlying simulation software (SPICE) though, and will ultimately have about the same limitations/accuracy/foibles.

[canman]

Thanks guys, I'll for sure check out those links.

Regarding a jumper across the cream pot...that's technically not connected.  I couldn't figure out how to make that jumper go over the jumper connecting the resistor and cap to lug 3 of the pot.  I know that lug 1 and 3 shouldn't be jumpered, that's just my bad on the schematic cuz it looks like it is connected.  So C5 goes between the collector and R5...dang.  I was really close.  That's the only part of the layout I had a hard time with. 

The two 47k resistors that come after the collector are throwing me off...

[highwater]

Whoops, my bad! I see it now - the re-sizing/anti-aliasing of the image made a little dot appear there... the actual "these wires connect" dots are much bigger. I guess that's the price you pay for using a free, moderately intuitive schematic program. You could avoid the crossing if you move the ground-point for R8 to above the output jack... this would also make it more apparent that it is a tone control.

Like this... aside from the incorrect values:


I rather prefer the schematic style where crossing but unconnected wires have a little hump drawn in one of them - using a dot, or a little zig-zag to indicate when they *do* connect can leave the reader scratching their head on a schematic that has a crossing, but no four-way junctions. The "hump" approach can clutter-up a busy schematic though, so I can understand why CAD software uses the "dot" approach instead (which is also easier for the software developer).

[duck_arse]

eagle is the only program you need to draw circuits, and as a bonus, it does pcb layouts as well. there is a free version limited to single sheet circuits. it even puts the connecting dots automatically, and only where they are meant.

[canman]

Oh OK, I see what you mean with drawing it like that...I was set on having ground going down, I wasn't sure if it was kosher to put ground wherever.  I guess you just do whatever it takes to keep the drawing clean!

I did download Eagle, but I couldn't get the libraries figured out.  Everything I see online looks so good, but I'm too dumb to figure out how to work the program!  I spent a couple of hours trying to make it work without luck, so Fritzing is my next best option.  For learning purposes, it does just fine.

Thanks for all the tips, I really appreciate them.  I'm sure I'll be back with more questions and schematic sketches!

EDIT:  OK I redrew the boost schematic, just to see if I got it right this time, and making the tone control a little friendlier to look at:



Better?

[PBE6]

Almost! Previously you had C3 connected to the top of R5 (directly connected to the 9V supply). You've changed it so that it is now in series with R5 and the collector of the BJT, which is incorrect. The correct method is the go back to your previous diagram, disconnect the left end of C3 from the 9V supply and reconnect it to the bottom of R5 (directly connected to the collector of the BJT). It will look like this:



(Don't worry about Ce at the bottom of the diagram, your C3 should be connected like Cout.)

[canman]

Oh, wait so do I have the whole tone control hooked up to the wrong spot as well?  Because C3 connects to R7, so should I have just moved that whole section (with C3 attached) down to between the collector and R5?

Actually I just drew it up again, should make more sense with a picture...