Series/Parallel effects switching matrix - concept

[G. Hoffman]

So, I've had this idea, and I think it may be too complex for my skill level, and so far it is all in my head.

I want to be able to run my time domain effects (delay, reverb, chorus, etc.) in either series or parallel, and the other night I had an idea on how to do it efficiently, and with a versitility and user simplicity which I can wrap my mind around.  The basic idea is derived from the switching for big inline studio consoles, such as SSL's or Neve's

To start with, you have 8 loops.  In my case, each loop would get a separate device.  Each device is set up 100% wet (or dry kill, or what ever you want to call it - it does not output the dry signal). 

Second, you have 8 Send and Return busses.  Each send buss sums the dry signal with the previous Return buss, and can send it to any or all loops.  Each Return buss takes any or all loops and sends them to the next Send buss, and to the output mix buss.

For example, if you assign all of the devices to Buss pair 1, everything is in parallel, getting just the dry signal in, and the output going just to the output mix.  If, on the other hand, you wanted to have everything in series, you would assign each device to it's own buss pair. 

It gets even cooler, though, because if you wanted two different series strings in parallel with each other, you just skip a buss.  So, for instance, you could assign a chorus to buss 1, and a delay to buss 2, skip buss 3, and assign a delay to buss 4 and a reverb to buss 5.  The reverb would not have the chorus, or the first delay. 

I know it's a little hard to understand it written up, but the pictures in my head are wonderfully clear.  I just haven't had the time to draw it up yet.

So, the benefits are pretty obvious - incredible flexibility achieved without too complex an interface.  Combine it with some MIDI controlled switching, MIDI volume controls, and transformer isolated outputs on the loop sends, and it is quite appealing.

Negatives - it's a lot of switches!  I'm using a Liquid Foot MIDI controller, so it would be relatively easy to control, but it involves 64 switches.  72 if you add switch able dry signals!  And I'm quite worried about noise - I'm not The worlds greatest electrical designer, and all those gains stages and switches, that could be a lot of hissing!  Also, definitely would need to be solid state switching.


Gabriel

[G. Hoffman]

Here is a quick sketch - a bit more than a block diagram, but definitely not a complete design.  Click the picture for a bigger file.  The buss switches are in pairs - one send, one return - and you can only have one pair at a time.  As I said above, the switches would be electronic rather than physical, but the idea is the same.




Gabriel

[G. Hoffman]

So, I want this to be +-9V.  Does this make sense for a JFET switch on a bi-polar power supply?




Gabriel

[Hatredman]

You should check this topic.

http://www.diystompboxes.com/smfforum/index.php?topic=112598

[G. Hoffman]

The problem with that chip is cost.  I can buy 250 jfets for less than two of those, which I would need for eight stereo loops. 

Also, if I do that I have to code, and I'm a terrible programmer.  My plan is to use a couple of Highly Liquid's MPAs (which are now open source, and I've got five of them) for the MIDI implementation.  It gives me all the logic outputs I need (16), with eight digital pots thrown in to implement volume controls. 

Still, it would keep the parts count down, and make for a smaller footprint on the board.

[G. Hoffman]

I take back what I said about cost!  I forgot the cost of the optos!


Gabriel

[SISKO]

I like it 

I would add DC block caps on the audio switch.