Basic doubt in I2C interface

[prav]

Hi all,

I have got a basic doubt in I2C interface.
In my s/m there are 4 masters(4 video processors).
Now my question if all the 4 processors have the same device
address,so they cannot sit on the same I2C bus. I need 4 differnt I2C
buses.
But then i am not using the I2C effectively.
Any suggestions.

Thanks in advance ,
Praveen

 

[John Fields]

Hi all,

I have got a basic doubt in I2C interface.
In my s/m there are 4 masters(4 video processors).
Now my question if all the 4 processors have the same device
address,so they cannot sit on the same I2C bus. I need 4 differnt I2C
buses.
But then i am not using the I2C effectively.
Any suggestions.

 

[Apparatus]

Hi Praveen,

On most chips of any complexity the IIC address should be configurable.
Check this first.

If not, you could use a CPLD (writing your own VHDL) to modify the
address of the chips. This is really more of a hack than a clean
solution, but it should work.

Not only will you not be using the bus effectively if you have
duplicated addresses, but you will have a fight on the bus (when the 4
processors simultaneously respond to a request).

Chris

 

[Paul Rolfe]

See a priest?


S/M - thats another newsgroup...


Don't sit on the bus, take the A-train uptown.


You're welcome.

I dont understand what you are trying to achieve, by having the four
devices on the one address.

As far as hardware is concerned, true I2C is open collector, so you
wont do dammage. Normally I2C is fast enough, and I would doubt
(unless you are actually streaming video) that the bandwidth
limitations should be an issue.

I would put the devices on different addresses.

Paul

 

[Paul Rolfe]

See a priest?


S/M - thats another newsgroup...


Don't sit on the bus, take the A-train uptown.


You're welcome.

Actually I did not read your post properly. Most I2C IC's have
addressing schemes, some of which are EEPROM, and you may have to put
some option links on your board, for the very first time.

 

[petrus bitbyter]

Hi all,

I have got a basic doubt in I2C interface.
In my s/m there are 4 masters(4 video processors).
Now my question if all the 4 processors have the same device
address,so they cannot sit on the same I2C bus. I need 4 differnt I2C
buses.
But then i am not using the I2C effectively.
Any suggestions.

Thanks in advance ,
Praveen

Along with your previous question, I think you need to invent a I2C router.
It should have its own address and some writable (and readable) registers
that records which master has to be connected to which set of slaves. A
master should set the correct path before accessing the slaves. The
appropriate switches should be opened and closed to make that connection.
You will still have collision problems when two or more masters wants to
access the same set of slaves. You may need to define and implement a
protocol to avoid collisions. (I have no patent on the idea, but you will
need to pay me to develop it for you.)

petrus bitbyter.

 

[Luhan Monat]

Hi all,

I have got a basic doubt in I2C interface.

See a priest?

In my s/m there are 4 masters(4 video processors).

S/M - thats another newsgroup...

Now my question if all the 4 processors have the same device
address,so they cannot sit on the same I2C bus. I need 4 differnt I2C
buses.

Don't sit on the bus, take the A-train uptown.

But then i am not using the I2C effectively.
Any suggestions.

Thanks in advance ,
Praveen

You're welcome.

 

[Rob Gaddi]

Hi all,

I have got a basic doubt in I2C interface.
In my s/m there are 4 masters(4 video processors).
Now my question if all the 4 processors have the same device
address,so they cannot sit on the same I2C bus. I need 4 differnt I2C
buses.
But then i am not using the I2C effectively.
Any suggestions.

Thanks in advance ,
Praveen

As several people have answered you in your previous question, and as
you seem dead set on ignoring, masters aren't addressed, only slaves.
While it may be that your I2C masters are also accessable as slaves,
unless you are planning to do so (which given your earlier question you
have no reason to do), the fact that they share the same slave address
is 100% irrelevant.

That said, until you sit down, read the answers that people give you,
and read the I2C spec on the Phillips site, you're officially being
written off as a waste of my keystrokes.

 

[Robert Monsen]

Hi all,

I have got a basic doubt in I2C interface.
In my s/m there are 4 masters(4 video processors).
Now my question if all the 4 processors have the same device
address,so they cannot sit on the same I2C bus. I need 4 differnt I2C
buses.
But then i am not using the I2C effectively.
Any suggestions.

Thanks in advance ,
Praveen

You must be talking about Hs devices. They transmit an initial 'master
code', which is used during arbitration. However, the master code is
intended to be software selectable. If your devices have unchangable
master codes, they are not compliant with the specification.

Your choices are to run the device in F/S mode, which requires no master
code, or to use separate interfaces. Since devices always come up in F/S
mode, and must be switched into Hs mode, it seems possible to force them
to use F/S mode. However, I'd have to look at the datasheet for your
device to be sure that is possible.

--
Regards,
Robert Monsen

"Your Highness, I have no need of this hypothesis."
- Pierre Laplace (1749-1827), to Napoleon,
on why his works on celestial mechanics make no mention of God.