alll possible permutation switching device

is there an IC that allows one to take in N inputs and permute them to N outputs based on other inputs? (I guess if there are N inputs then there are N! permutations and one would need log[2](N!) control lines?) (and I'm not talking about in software though. I might need upwards of 40 inputs or more (though maybe one could cascade smaller ones to make larger ones?)

i.e., I just want to have a "controlled" permutation of N lines. Maybe control the lines using a serial interface instead of having all those control lines.

Just wondering,

AD

--
"Permutate" them? Yes.  Here on Earth, we call it a
"microcontroller".  Or, possibly, a "ROM".

Digital or analog? Each output is mapped to one and only one input or something else? Each input is mapped to one and only one output or something else?

If the second statement above is true, then you need log2(N) bits for each output, so N*log2(N) bits. These could be fed over a serial bus with only a few wires.

Sometimes this function is called a "crosspoint switch"-- you can imagine an array of N wires horizontally and N vertically, with switch contacts in the appropriate spots.

I have one recent design that feeds out the switch value for almost

1,300 "switches" from a handful of pins on a microcontroller.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

without that info, cant recommend anything specific. One approach is to use cmos invertor ICs. For digital, open circuit, for analog, with feedback. Add dc to the input to switch them off, via a diode, and OR or mix the outputs. This is good for some apps.

NT

Could it be a crosspoint IC like they used in telephone installations??

--
Maybe.

Ok, I think this might be what I'm talking about or very close. I'm not so much interested in the logic part of it but just the switching(i.e. for analog).

The way I see it is that one has a matrix

1 --+---+---+---+---+---+--- | | | | | | 2 --+---+---+---+---+---+--- | | | | | | 3 --+---+---+---+---+---+--- | | | | | | 4 --+---+---+---+---+---+--- | | | | | | 5 --+---+---+---+---+---+--- | | | | | | 6 --+---+---+---+---+---+--- | | | | | | 7 --+---+---+---+---+---+--- | | | | | | 8 --+---+---+---+---+---+---

Then by breaking certain connections and jumping over others one could route different pins to others. I think this is similar to the idea of FPGA or something like that. The idea is very similar to how memory works but instead of holding a value that value is just actually used to control a line(using a transitor I guess).

Its hard to show cause one has to jump over lines that I can show here but you probably get the idea anyways.

I'll look more into the FPGAs and see what I can come up with

Thanks, AD

Yep, thats basicaly the idea. I actually need a digital and analog one. I figure that at each point one would just use a transistor so it could control digital or analog...

;/ I looked online and found some. They are quite expensive it seems... Although I'm seeing ones used for video and that could be the difference... it definately seem to look like what I want. I'm looking specifically at the AD8116 right now and the FBD seems to be what I want more or less. (though I would need something with about 100 inputs and 100 outputs). Just for 100$ its not worth it ;/

I suppose one can accomplish the same thing with using multiplexers. One would need one Nx1 multiplexer for each of N lines ;/

I might have to find some other way as this seems to be way to expensive ;/

Thanks, AD

Well, I don't need all the extra stuff that comes along with a MCU. I just need the switching and only the switching... and many inputs. While maybe I could easily accomplish it with just 2 parallel MCU's(say each with about 50 pins since I need about 100) I figure that surely there must be devices that are strictly for switching that are pretty cheap... I might have been wrong though and might end up having to use MCU's.

Thanks, AD

Yep, that seems to be pretty much exactly what I want. From a very brief search on the web though it seems these things are pretty expensive, I might end up having to find another alternative then.

Thanks, AD

Reading between the posts, I think you are getting seriously hung up on creating a universal do-everything programmer.

If you are interested in working with microcontrollers, pick some microcontroller family - any family - that can be programmed in-system, and buy or build the ISP programmer for that, and get on with playing with microcontrollers. Forget making a do-everything programmer until (or unless) you find that you really need one.

--
Peter Bennett, VE7CEI  
peterbb4 (at) interchange.ubc.ca  
new newsgroup users info : http://vancouver-webpages.com/nnq
GPS and NMEA info: http://vancouver-webpages.com/peter
Vancouver Power Squadron: http://vancouver.powersquadron.ca

--
What is it, exactly, that you want to do?

Delete my suggestion. FPGAs by thenselves can't switch analog signals. Although depending upon the complexity of your switching task you could use CPLDs or FPGAs to conrol an analog multiplexer.

Howard

Abstract Diss> > Sure a FPGA. However, see "The Design Warrior's Guide to FPGAs" for a

If you're working on this universal device programmer, either you have to restrict the universe or deal with some voltage/current levels that are more extreme than you seem to expect. Old UV EPROMS will have at least 25 volts, and old bipolar PROMS and PALS can have currents in the

100 milliamp range that have to switch from 0 to about 12 volts in 10 to 100 nanoseconds (the old AMD PAL programming spec, in this case).

You'll find that you only really have a few special voltages for any one part, (most are either logic, static power, or ground). So you can get by with, at each pin, switches to a few programmable voltage supplies along with a logic low, a logic high (both also settable, but not dynamic like the other voltages), and a really stiff ground.

Hardest to deal with are various specs that specify a source resistance, a limited slew rate, or timing down to a microsecond or so. And an analog comparator to read the outputs back that is both fast and won't get fried.

Mark Zenier snipped-for-privacy@eskimo.com Googleproofaddress(account:mzenier provider:eskimo domain:com)