Band pass filter pedal

[Bastis]

Hey guys!

I finally had time to get back into this jungle! I've made a new attempt at getting a schematic that would work. I switched the filters to sallen key filters and redid the switches as PBE6 suggested.

Thoughts? Was it right to set use Vref instead of ground in the filters? And is the Vref set up correctly?



edit: Forgot to attatch the image

[PBE6]

I think you need to switch the order of the HP and LP sections. The capacitor C3 won't pass DC voltage through to the IC1 opamp circuit so it will be mis-biased. Switch the order so IC3 comes first and you should be fine.

You can almost certainly get away with a single input buffer as well. Instead of using 4 JFETs, use one half of a dual opamp for the input buffer and the other half for the mixer.

[Bastis]

Ah cool! I'll hopefully make a new (and final?!) version tomorrow! Do the input and output cap values make sense btw?

[garcho]

Do the input and output cap values make sense btw?

yes, but you may want to add something like a 1MΩ/2.2MΩ resistor to ground after the final output cap to help prevent popping during switching

consider using pots instead of switches for the mixing section. that op amp arrangement is a "summing amplifier", for all practical purposes it's just an inverting amplifier. you have four 10kΩ resistors in parallel with a 10kΩ feedback resistor, which will roughly give you a gain of 4 (V_OUT = -(V1_IN+V2_IN+V3_IN+V4_IN). maybe use four 50kΩ pots instead of switches and then you'll be able to dial in how much of each band is in the total mix and how much it's amplified. might as well throw some coupling caps between the switches/pots. if you stick with the switches, you might want to put them somewhere earlier and ground the unused band's input, or just use SPST and don't ground the outputs, to 0V (or Vref). were you planning on using only one band at a time? because if you were, there are ways to "tune" a multiple feedback band pass filter (it greatly affects the Q, so no cool synth stuff, but still sounds like a wah-wah BPF effect) with only one dual gang pot, for instance. or if you don't want to mess with that, you could use one rotary switch to select which band you want; Small Bear has a nice compact 1P6T that wouldn't take up too much space.

[PBE6]

Adding a resistor to the end of the circuit to drain any excess voltage from C10 is a good idea if you plan on switching the circuit in and out. A 47k resistor should be high enough, the frequency cutoff is 34 Hz and the time constant is about 5 ms. Anything higher will work too, but very high values might slow down the discharge rate appreciably.

With regard to the summing amplifier, the noise gain will be 4x with all the channels connected but the output signal will still be Vout = -(v1*R1/Rf + v2*R2/Rf + ...) = -(v1 + v2 + ...), or unity inverting gain of the sum of all connected channels, if the input resistors are the same value as the feedback resistor. These channels are bandwidth limited and should add up to Vin when they're all engaged. Pots, as implemented on the Elliott Sound mixing page, are a good idea if you want to control the blend more precisely. Switches will be fine though.

With regard to switch placement, disconnecting the input resistors from the -ve input on the mixer completely will reduce noise gain. Other options will allow mixing as desired, but won't have that (minor) advantage. I would suggest leaving them where they are.

[Bastis]

The plan is to use more than one filter at the time, and for now I won't need any pots, switches will do the trick.

I tried to implement the changes you suggested. I don't however quite understand where coupling caps would go between the switches. Would it be between vRef and the switch when the switch is set to "off"? And PBE6, when you write "disconnecting the input resistors from the -ve input on the mixer conpletely will reduce noise gain", is it the solution I already have that you suggested, or a tweak? I also hope I managed to get the input buffer set up correctly.

Edit: I just reread the replies in the post, and I saw the handdrawn picture PBE6 posted. So the coupling caps should go between the op amp and the resistor before the switch, right?

[PBE6]

I think you can get away without using coupling caps around the switches. The switches looks fine as drawn, no changes are necessary.

You should remove all the C6 capacitors however. The C4 caps do double duty as DC coupling caps and filter components. Unless C6 is very large, it will interfere with the filter operation by reducing the apparent value of C4 (caps in series add in parallel).

Once that change is made, you're ready to build!

[PBE6]

EDIT: I just noticed that the frequencies in your first two sections are 130-250 Hz and 350-1042 Hz. Did you want to de-emphasize the 250-350 Hz region on purpose, or is that a typo?

[Bastis]

Oh, that's a typo! I'll double check the numbers in a filter calculator to make sure I didn't make a mistake there though.

Awesome, I'll start working on a perf board layout then! Thanks a ton, and I'll keep this updated once I start building it!

[PBE6]

Sounds good, looking forward to the results!

[garcho]

add a pull down resistor on the front, too. when you switch this pedal in and out, you might hear a 'pop' when you toggle the bypass. if you add a 1MΩ resistor between your input jack and the C1 (and one at the output, which you've already done) this will prevent that popping, or at least make it tolerably low-level.
Capacitors store charge, and while the low-value input and output caps don't store a lightning bolt's worth of current, they do store enough to make an audible pop when they go from "hanging" (not connected to anything on one side) to "connected" (what happens when you engage the bypass switch).
looks like a cool project, sorry if my musing confused you.
glad to see a saxophonist around here!! cheers

[PBE6]

Good point, garcho. An input ground resistor will definitely reduce pops when switching in and out.

I wonder if we've been recommending appropriate impedance input and output values for this project? Bastis, what connections will you making? SM58 into the circuit, then into the board? If so, the front end pull down resistor should be quite small, somewhere around 4.7k (or less) for best signal/noise ratio.

EDIT: actually much less, for a 150 ohm mic it seems the ideal value for high SNR is somewhere around 320 ohm, but that assumes a perfectly quiet amplifier. I'm sure the standard 1.5k-3k value would work ok here.

[Bastis]

I've changed my setup a bit, so now I'm using a JTS wireless microphone with a line level output from the reciever. So it should have the same impedance as a guitar I think.

I made the changes you guys suggested, and I am super grateful for all the advice, encouragement and your patience! You guys are great teachers!

[Bastis]

I just started breadboarding this thing, and stumbled upon a question. Should the Vcc- pin of the TL072 (pin 4) be connected to Vref or ground? Or something completely different that I had no idea existed?

Edit: Ground seems to be right... at least for the first opamp...

[PBE6]

The Vcc- pin should be connected to the lowest voltage available. With a single battery, that is the battery's negative terminal.

[Bastis]

I'm getting heavy distortion from the op amp in the mixer section... Any thoughts on why? I've only connected one filter now, if that should matter...

[PBE6]

That's very odd. The circuit looks correct as drawn. What diagnostic tools do you have available? DMM? Scope? Audio probe?

[garcho]

I'm using a JTS wireless microphone with a line level output

if it's getting line level instead of "guitar level", maybe you'd need to attenuate the signal up front?

[Bastis]

All I have is a multimeter. I'll try playing a guitar though it and putting a pot before the opamp tomorrow!

[Gargaman]

I just like to thanks and praise you, this post has been a lesson to me.
Best regards