Box of rock, clean blend, baxandall

[marcelomd]

Hi,

I posted this a few days ago, but looks like the post was lost with the hosting changes so I'm posting this again.

This is a thing I designed myself, then I found something like that already existed. The pedal is an AMZ mosfet booster followed by two gain stages I got from the Box of Rock, in parallel with a clean blend, into a Baxandall tonestack I got from Xotic. JHS Charlie Brown is the same thing, minus clean blend, plus mids in tonestack.

It's a nice low-mid gain overdrive. Just wanted to share.



Questions:
- The tonestack sounds a little muffled and thin. Anything I can do? Someone said to try changing some coupling caps, but I can't remember the values suggested.

- Do I really need the buffer in the clean side? I put it there because the wet side was bleeding into the booster and causing feedback. Edit: Looks like I need something there. Is there a simpler way to avoid feedback? With less parts?

Thanks!

[Danich_ivanov]

Hey, there it is again. Yes I suggested replacing 1uf and 10uf caps with something smaller, like 100nf , since big caps in this case can make things "muffled", there might be something else though.

This buffer doesn't seem necessary since you'r dry signal is already amplified, i doubt buffer adds much to the picture.

[marcelomd]

Hi, it was you  Danich, Thanks. I'll try smaller values as soon as I can.

[Danich_ivanov]

But there is something else, which is that at the moment how dry you'r dry signal is, depends on how much gain you have, it also ends before eq section, so unless the idea behind this is that dry is not entirely dry, I would suggest to rearrange some things, by splitting the input into the first gain stage and a buffer and mixing two signals at the very end.

[ElectricDruid]

What do the two 1M resistors on the ends of the Mix pot do? R23 and R24?

They're connecting the signal to earth, which doesn't seem right. If you removed them and C22 as well, that whole section would be biased to Vb by R30 - so no problem.

I'd also ask why there's separate Va and Vb bias networks. Are they different voltages? Do they really need to be? With a bit of luck, one of those could go too.

HTH,
Tom

[marcelomd]

Hi,

The dry signal is not really dry because I wanted less difference in volume between channels. Also, this way I can use the pedal as a cleanish-sparkly boost, a la SHO, as a slightly dirty boost, with some dirt mixed in, up to full distortion.

The tonestack location is something I keep changing every time I rebuild this =) The current thought is to be able to shape the tone of the pedal as a whole. I may try a simple tonestack on the dirt side in addition to the main one.

Resistors R23 and R24 are leftovers from when I was having feedback problems. I think I wanted to load the signals. I'll remove them. Good catch.

VA and VB. Duh, I did not put the values on the bias networks, sorry. The AMZ booster calls for a 100k/62k pair. I left the reference if I needed a buffer at the end. I'll try self biasing too.
The opamp uses symmetrical values to get half supply.

Thanks guys. I really appreciate the input. I'll do my homework, brush the schematic and get back to you.

[rankot]

I had a similar problem recently, trying to run two pedals in different parallel/serial combinations. One of them causes heavy feedback, so I concluded that simple potentiometer blend makes the trouble. I wanted to have two independant volume controls mixed at the end (not single blend pot). So I came to this solution, but I still didn't try it in reality:

[marcelomd]

I had a similar problem recently, trying to run two pedals in different parallel/serial combinations. One of them causes heavy feedback, so I concluded that simple potentiometer blend makes the trouble. I wanted to have two independant volume controls mixed at the end (not single blend pot).

Someone correct me on this. But signal bleeding happens because of an impedance mismatch between sources, right? The buffers provide low enough output impedance that one side can't drive the other.

The key to avoid feedblack/signal bleed is to use active mixing.

This is a good guide: http://sound.whsites.net/articles/audio-mixing.htm
The first circuit here works as well: http://www.geofex.com/Article_Folders/panner.pdf

For the specific circuit in this thread, I went with passive mixing because I don't mind a bit of dirt in a... dirt box... =) A compressor or modulation needs to be squeaky clean though.

Thanks!

[ElectricDruid]

Someone correct me on this. But signal bleeding happens because of an impedance mismatch between sources, right? The buffers provide low enough output impedance that one side can't drive the other.

I wouldn't have described it like that. If you didn't have that buffer in the clean channel, you've effectively got a resistor (ok, two resistors and a pot) between the output of the drive side and the input. That's clearly going to cause feedback. The buffer breaks that path, which is why it helps.

In the case of passive mixers like the ESP example, where you various sources fed via a 10K resistor to a single point, the only thing stopping bleed through from channel 1 to channel 2 is 20K. That's not much.

The key to avoid feedblack/signal bleed is to use active mixing.

Absolutely. Mixing the currents at a virtual ground node means that there is no voltage there, so there is no signal that can "go backwards" down the mixing resistors to the inputs and cause bleed through.

This is a good guide: http://sound.whsites.net/articles/audio-mixing.htm
The first circuit here works as well: http://www.geofex.com/Article_Folders/panner.pdf

I'd say they're both excellent articles on the topic.

For the specific circuit in this thread, I went with passive mixing because I don't mind a bit of dirt in a... dirt box... =) A compressor or modulation needs to be squeaky clean though.

You adapt the solution to fit the requirement. Entirely sensible and fair enough. Not everything needs a full-blown active mixer.

Tom