Runoffgroove Splitter Blend with simple crossover?

[CamH132]

Hey guys,
Basically I'm planning on building a splitter / blend pedal which has a simple variable crossover inside so that the guitar signal can be split into low and high frequency bands which can effectively be processed through their own individual FX loops.
Here's my current schematic:


I used the runoffgroove splitter-blend schematic as a starting point (http://www.runoffgroove.com/splitter-blend.html) and replaced the blend section with a variation on figure 7 here: http://sound.westhost.com/pots.htm

The one thing that i'm really not sure about is the crossover section. In my schematic I've just used a dual gang pot to simultaneously adjust the resistance in the respective hpf and lpf RC networks for each band. To achieve a lowest crossover point of around 100Hz and highest of around 1kHz, I've calculated for the pot to be 1k, the cap 1.5uF and the resistor in series with the pot to be 100Ω. The series resistor is there so that when the pot is fully open there will still be at least 100Ω of resistance in the RC network to limit the highest crossover frequency to around 1kHz. Would this all work or am I completely wrong here?

Another thing that I would like to implement is wiring the pedal so that the sends / returns are half normalled so that if nothing is plugged in, the dry signal will still pass through to the blend / output stage. I'm assuming this can be done fairly easily with stereo jacks but I'm not really sure how to go about wiring it up?

Thanks heaps in advance

[antonis]

A schematic of your crossover would be helpful...

[electrosonic]

Two things jump out at me. The series impedence is effectively the input impedance, which is low enough to warrant an input buffer. U1A is biased around ground so it will distort. (Your input cap is blocking the bias voltage)

Andrew

[PRR]

> calculated for the pot to be 1k, the cap 1.5uF and the resistor in series with the pot to be 100r

As Andrew is saying: then the input impedance can be as low as 100 Ohms, which is far too low to stick in a guitar chain. Or even for convenient buffering. (TL072, even 5432, won't drive 100r well.)

And 1.5uFd is an awkward value. It is either a small loose-tolerance electrolytic, or a huge costly film.

And the opamp specified is fine in circuits up to 1 Meg. So why go low?

Pick 100K pot, 10K stopper, cap is now 0.015uFd (15nFd).

The input can still be as low as 10K, which is kinda low for most guitar chains, and suggests a buffer (TL072 will be running light and easy).

The output mix has some loss (also depending what you pick for that output pot). Depending what FXs you put in the loop, you may wish for more unity gain.

Seems to me there must be a simpler way to do a 2-way 1-pole variable crossover. But I need to think on it.

[PRR]

> must be a simpler way to do a 2-way 1-pole variable crossover.

One frequency pot.



Derive the high-pass.

Subtract the high-pass from the total signal.

Result is a perfectly complementary crossover.

(Don't get carried away. This gets screwy with 2-pole networks.)

I have also set-up the DC bias for single 9V supply, showing where you need a 4.5V "Vref" which is also firmly AC-grounded. And the necessary in and out caps. (The 10K at the outputs are not strictly necessary, but my simulator can't see what's outside the box, anyway a bleed is good form.)

You will still want a buffer before all this (even tho it means an odd number of opamps).

[CamH132]

thanks guys, i'm still fairly new to this stuff so i'm going to need a bit more guidance...

Paul, can your crossover schematic replace the whole first half of my original schematic (with the addition of an input buffer) or will it still required the op amp output buffers? Also, the amplifier labelled '1' just after the high pass filter, what exactly is this?

For the output put i was planning on just using a standard 100k log pot. The FXs that i'll be using with this will most likely only be distortion pedals (tube screamer / big muff / rat etc). Would you recommend adding in an output booster? 

[blackieNYC]

Paul - this is great. I'm building a fuzz around this immediately. Blended back together. Had good luck with a primitive BMP tone split as a xover. This should be better.
Would the best way to give it a notch be to simply inset a notch filter before the split? I don't think you could work a notch into the circuit elsewhere.

[PRR]

I did not bother with your output buff/mix; I assume this was known-good.

The "1" triangle is a buffer (gain of 1).

[CamH132]

In that case how does something like this look for the first half?

[slacker]

You don't need U2A and U2B, you can just take the sends from the outputs of U1A and U1B like in Paul's schematic above, "High" and "Low" become green and red send.
You might want a resistor in series with the pot to set the upper limit of the filter, like you had in your original drawing.

[CamH132]

Alright so how's this look?


i'd also still like to wire the send / returns up half normalled so that if nothing is plugged into the sends the dry signal for that respective band will still get through to the output / blend stage. all wiring diagrams for half normalled sends / returns that i can find are all TRS balanced though so i'm not really sure how to go about implementing this since the pedal sends / returns are unbalanced / TS.

[slacker]

That looks good to me, I'm not sure what your balance control adds compared to the way the splitter blender does it, you could add a volume control to that by replacing the 1M resistor after the blend pot with a pot wired the same as your out pot. Your way should work fine though.

To normalise the send return loops you want sockets like this http://www.parts-express.com/Data/Default/Images/Catalog/Original/090-972_HR_0.jpg These have two contacts in each position, as you can see the one on the right hand side of the photo sits on top of the left hand one. when nothing is plugged in they connect together. When you insert a jack the connection breaks and the right hand side one carries the signal from the jack. Easier so see when you have one in your hand than to explain

[CamH132]

I was originally under the impression that the added balance control blended the two paths a lot more smoother, but if the original splitter blender log blend pot works fine then I'd prefer to just stick with that.

The switching jacks make perfect sense, cheers! - and also a big thanks to Paul for your help

EDIT: the frequency pot in the crossover section, should that be log or lin?

[PRR]

> pot in the crossover section, should that be log or lin?

Reverse audio.

[thehallofshields]

I like this.

Any way to adjust the Crossover Frequency?

[blackieNYC]

The one pot Paul has there is the frequency pot.

[thehallofshields]

The one pot Paul has there is the frequency pot.

R5 just looks like a HPF Trimmer to me. What am I missing?

[slacker]

U1A is a HPF, R5 sets the cut off frequency. U1B then subtracts the high passed signal from the input signal which gives a low pass filter with the same cut off frequency.

[Gus]

A link http://www.diystompboxes.com/smfforum/index.php?topic=99697.msg875121#msg875121

[thehallofshields]

U1A is a HPF, R5 sets the cut off frequency. U1B then subtracts the high passed signal from the input signal which gives a low pass filter with the same cut off frequency.

Thanks Slacker. You've always got my back!