Pedal Matrix

[zerohero]

Hey guys,

I haven't posted in a while, Ive really just been soaking it all in. I am a music engineering student at UM (Miami fl), and i have been given the opportunity to have a semester (i might be able to stretch it to 2), of independent study, the idea for my project is much like what is discussed here < http://www.diystompboxes.com/smfforum/index.php?topic=42518.0 > and on R.G's site, however i want to be able to not only choose the order of effects but have some in a parallel. so if you had 8 effects, you can choose your guitar signal to go through effect 1 then to 5in parallel with 2 , then to amp. or any combination. now i realize this is a huge task, it might be easier to convert the audio signal to digital and then route it all with a program, that would require as many A/D's D/A's as there are effects but it might be the way to go. however i want to keep it as close to R.G's idea as possible, its actually funny cause i thought of the idea and then i came on line and searched here only to find it had been thought up before lol.

so i guess what i am asking for is just a push in the right direction, there is so much out there its hard for me to see where to start. I'm not asking for nor do i want the answer cause that takes away the fun for me.

thanks guys

[puretube]

http://www.diystompboxes.com/smfforum/index.php?topic=43104.0

[Peter Snowberg]

Take a look at the Zarlink MT8816 cross point switch.

I was all excited to build a crossbar switcher but life got in the way and there was very little interest here. The jacks are sitting in a 1U rackmount case waiting for a little free time. Someday....

[zerohero]

thank you for pointing me towards teh Zarlink chip, that gonna make my life a lot easier and because of all the chip select feature it seems easily expandable.  my only question at the moment is, at the moment im doing all my designing with block diagrams, just a rough sketch of the signal flow and where to commands come from, im not even thinking about the physical layout of the circuit, is that a wise thing to do or will i end up screwing myself over?

thanks again

[Peter Snowberg]

I'm not totally sure I understand, but I think I have the picture.

I generally start with "chicken scratch" designs where the basics are covered and then work towards a complete design with all the chip selects etc.

With the crossbar design, I started with a matrix representing the crosspoint switch chip and than asked a whole bunch of what-if questions. Eventually I came to a point where I was happy with how the ins and outs all related and from there I added the details that actually make the thing run.

Start with a list of what you hope to accomplish. From there jump to rough sketch block diagrams. Go back and forth between those to things until the list contains everything that it need to and the block diagrams show that it can be done. When that looks good, think about the actual schematic.

That's how I do it anyway.   

[Dave_B]

You might also want to consider laying out the hardware before you get too far on the software.  By making a few compromises in software, I've eliminated a lot of jumpers and silly looking traces.

[David]

Take a look at the Zarlink MT8816 cross point switch.

I was all excited to build a crossbar switcher but life got in the way and there was very little interest here. The jacks are sitting in a 1U rackmount case waiting for a little free time. Someday....

Peter:

I did look at it.  That is one cool piece of circuitry.  I actually got light-headed thinking about what that could do.  The next thing I did was try to figure out where I could buy it.  Umm...  I can't find it.  Do you know who has it?

[puretube]

http://products.zarlink.com/where_to_buy/

[David]

http://products.zarlink.com/where_to_buy/

Well, that explains not finding it at DigiKey or Mouser.  Yup, there it is.  Pricey little bugger!
However, if it keeps me from having to make a crossbar switch the hard way, it might be worth it...

[Dave_B]

Looks pretty cool.  I found it for $7.00 in small quantities.  David, was it more expensive where you called?

[R.G.]

It's probably worth mentioning that I looked for suitable crossbar ICs when I first did the footswitcher stuff.

They exist, and it is possible that new ones have been announced that are cheaper or otherwise better than I found, but I didn't find any crossbar that had all of the characteristics that it was:

1. under $20
2. rated for at least 2V pk-pk signal
3. suitable for making up at least an 8x8; there are a few 4x4's and 4x6's that are possibles, but you need two to four of them
4. available in through hole packages, not surface mount only.

Maxim came closest. The possibles I found were primarily baseband video switches. The Zarlink is an interesting new device that I didn't find.

Frankly, for the degree of difficulty and prices I found, making an 8x8 out of eight CD4051's is just pretty darned attractive. Sure it's a bit bigger PCB space, but the parts are cheap and easily available, and once you do the PCB (which I did, by the way) it's no more complicated than building one with a bigger chip, and a lot less complicated than building with a surface mount device.

So I just presented the low-level integration version made from CMOS. By the way - the CMOS multiplexer version works well.

I used a trick to make sure that it did not pop. I used an H11F3 LED/photo-FET module driven by the microcontroller to "short" the signal to ground at the output of the matrix from just before I started changing switches until just after I started changing switches in the uC programming. I think it was about 48uS of "silence" altogether. That low impedance completely ate any switching transients from the switch array.

I've seen that trick after I came up with it, but not before. I haven't seen a patent on it, and I didn't want to spend the money doing it.

[David]

Frankly, for the degree of difficulty and prices I found, making an 8x8 out of eight CD4051's is just pretty darned attractive. Sure it's a bit bigger PCB space, but the parts are cheap and easily available, and once you do the PCB (which I did, by the way) it's no more complicated than building one with a bigger chip, and a lot less complicated than building with a surface mount device.

This is a telling point, and worth being brought up more than once.  I should know better, too.   
Sometimes, a monolithic be-all-end-all solution may not be a solution.  Not in software (no way), and not in electronics.

[puretube]

page 12

[R.G.]

page 12

Interesting, but limited.

The chips let you select one of 10 stereo inputs. That's the equivalent function to two CD4051's in parallel with ten inputs instead of eight.

You could use these to do a 10x10 matrix, but you'd need ten of them to do it.

I pored through the Zarlink datasheets. They do the necessary stuff, all right. The only issues are price and availability.  A PCB for one of these should take about 30 minutes to do, and after that, it's ASMOP.

Same comment applies to an 8x8 CMOS array. I'm guessing that the only real differences are secondary ones - chip cost, chip availability, and programming. The CD4051 solution only requires you to put out three bits of address to each of eight chips to point each input (or output) to its destination. The Zarlink chip requires you to individually write each switch. In payment for that bit of frustration, the Zarlink has the ability to be a non-blocking crosspoint since you can use all of the switching individually. That's made less valuable by the fact that an any-order FX switch doesn't need non-blocking - all of the reasonable configurations can be done by a blocking crossbar.

It's always something, isn't it?     

[TELEFUNKON]

>1.200 pages
.pdf about switching by TI  (20Mb).

[puretube]

Zarlink @ futurlec