a chip which do the same as the mux/demux When they are put together

[erfanbas]

(i mean an alternative for this circuit)

 

[KlausST]

Hi,

You want us to decode the function.
Isn't it more simple if you give us a description or a truth table?

Klaus

 

[dick_freebird]

It looks to me like you could replace it with an octal buffer and
ignore the addresses altogether.

 

[FvM]

Not an octal buffer. The circuit blanks all not addressed signals. But the left circuit halve can be omitted.

 

[Akanimo]

Hi,
Unfortunately, this is something you might need a programmable logic device to implement.

 

[KlausST]

Hi,

From what I see ... one can omit the right side completely.
Isn't the output of the 8 left side AND gates the same as the output of the right side AND gates?

Klaus

 

[Akanimo]

Hi,

From what I see ... one can omit the right side completely.
Isn't the output of the 8 left side AND gates the same as the output of the right side AND gates?

Klaus

Hi Klaus,

I thought so at first. Having taken a closer look, I realized it's a combination of mux and demux like the OP said. The left side is the mux and the right side is the delux.

I think he wants to implement something that can at one time, data parallel data from the inputs and pass them through a single selected output, at another time it can take a serial data from any of the inputs and channel it through a single selected output and at other times it can take a serial data from any of the inputs and broadcast it to the outputs.

However, if this is the case, I'm concerned about why the inputs and outputs are sharing a common address bus. I'd think the address ports should be separated

 

[KlausST]

Hi,

yes, but the control for both sides is identical. The control is not independent.

Thus if D is selected on the left side ... it will be fed through the one line, but then gets output at Q
The channel on both sides us the same.
If number (3) is chosen on left side .... also number (3) is chosen on right side...
Am I wrong in this?

What about more information from the OP?

Klaus

 

[Akanimo]

You're right, the OP definitely needs to give more information so we can reason along.

I don't know of any such chip in existence so I suggested the use of a PLD for the implementation. That should get it done.

 

[FvM]

I think he wants to implement something that can at one time, data parallel data from the inputs and pass them through a single selected output, at another time it can take a serial data from any of the inputs and channel it through a single selected output and at other times it can take a serial data from any of the inputs and broadcast it to the outputs.

We can either discuss the given circuit or any similar looking circuit that might be implemented.

As for the given circuit, you can either omit the left or the right circuit halve and get the same function, as already mentioned.

 

[Akanimo]

You're right, all of us are saying the same thing. There's actually no point for the part of the circuit on the right. He mentioned mux+demux in one package. This circuit can't implement that function. It can't do more than a mux can. However, there something additional to the mux function that the circuit could do if the control inputs were separate. That's what we were pointing out to the OP.

I think he wants to implement something that can at one time, data parallel data from the inputs and pass them through a single selected output ... and at other times it can take a serial data from any of the inputs and broadcast it to the outputs.

This circuit can't do these two. I'm sorry I made that blunder. The schematic actually does not tally with the subject. Quite misleading though.

 

[erfanbas]

THANKS for answers.you are right.i made a mistake.the control in two side shouldnt be the same.

 

[erfanbas]

thanks everybody.you are right i should change the circuit .if the controls dont be the same in both sides,is there any alternative for this one?(i mean for example the control line be 111 in the left but in the right for example it is 100?

 

[Akanimo]

If such a chip exists, then it would not be something to easily come by. Have you searched the internet? Major parts dealers like digikey? I suggested earlier that you could use a programmable IC. It will save you a whole lot of headache and time.

By the way your circuit is a lot better now. Can you tell us how you intend to use it? We might have something more to add or modify.

 

[KlausST]

Hi,

You may program this into a PLD.
You may use a DEMUX/MUX combination.
You may use a crosspoint switch.

Klaus

 

[ads-ee]

The circuit is just an 8 channel router. With 8 of these you could make a cross point switch. If you are looking for a chip that does this, you should try looking for a multi-channel switch of some sort, but this particular (single bit) implementation seems rather unlikely. There might be a better chance if you are looking for a byte or nibble wide implementation as that could be used for something like a multi-master to multi-slave SPI switch.

I've used FPGAs to do this kind of thing in the past.

 

[Akanimo]

...
I think he wants to implement something that can at one time, take parallel data from the inputs and pass them through a single selected output, at another time it can take a serial data from any of the inputs and channel it through a single selected output and at other times it can take a serial data from any of the inputs and broadcast it to the outputs.
...

If you need to implement all of these three functions, then you need to include registers and you'll need at least two clock domains. You can even do parallel to parallel data transfer in addition to those three, but at several clock cycles.