OMAP LCD signal muxing

Hello,

I'm planning to use the exposed DSS signals at conectors J4 and J5 on Beagle rev C to connect to an LCD. The display I'm planning to use has 18 bit RGB interface thus utilizing only DVI_DATA0 to DVI_DATA17 plus control signals on those connectors J4 and J5.

I'm under impression that as almost every ball on OMAP3 can be muxed to several configurations, they are also independent from the mux values of each other. Is this correct? I'm asking this because I also need to use SPI3 which is available not only on the expansion connector but also on LCD headers. If I could use SPI3 on the LCD headers it would free many valuable GPIOs on the expansion connector (and one bonus signal on J4). Before trying this I'd be glad to know if there aren't such features as "If you plan to use OMAP DSS you must configure all the DSS pins to DSS functionality".

Btw, what happens if two pins have such mux values that they both would have the same signal? Say, for example McSPI3_CLK. Not that I'm planning to do such, just as a curiosity.

Thank you,
  Antti

Each pin can be set independently based on the mod eused on each pin. The DSS pins are not in a mass function where they all are DSS.

It is my understanding that If you map two pins to the same signal, then it will show up in both locations, so I suggest that you not do that.

You can use SPI from the LCD connector if you choose, but do not set the expansion connecotr to the same signal. You cse them as GPIO.

Gerald

Hi Antti,

I'm under impression that as almost every ball on OMAP3 can be muxed to
several configurations, they are also independent from the mux values
of each other. Is this correct?

Correct - Information about the different possibilities for each ball is
easiest extracted from the Data Manual
(http://focus.ti.com/lit/ds/symlink/omap3530.pdf)

I'm asking this because I also need to use SPI3 which
is available not only on the expansion connector but also on LCD
headers. If I could use SPI3 on the LCD headers it would free many

valuable GPIOs on

the expansion connector (and one bonus signal on J4).

Agree - Just go ahead and check - On the other hand, you would just have the
extra GPIOs on the LCD connectors :slight_smile:

Before trying this I'd be glad to know if there aren't such features as

"If you plan to use OMAP

DSS you must configure all the DSS pins to DSS functionality".

No such requirement can exist, since the multiplexing is done completely
independently from the IP blocks => As long you are only speaking output
pins the OMAP IPs wouldn't even notice.

You of cause might have to deal with some source code trying to change to
mux-mode of a pin, but at least as I know the system for now, muxing is
mainly dealt with within U-boot, so I don't think you will get into trouble
- Secondly this is only SW - no HW requirements exist...

Btw, what happens if two pins have such mux values that they both would
have the same signal? Say, for example McSPI3_CLK. Not that I'm planning

to

do such, just as a curiosity.

The signal would be routed to both balls - This is not intended, but might
give you some pretty smart debugging possibilities once in a while :slight_smile:
However TAKE CARE: Do not connect two balls to the same input signal in the
OMAP - Do only use the trick for output pins :slight_smile:

Best regards
  Søren

Thanks Gerald and Søren for your responses!

Wed, 6 May 2009, Søren Steen Christensen wrote:

I'm asking this because I also need to use SPI3 which
is available not only on the expansion connector but also on LCD
headers. If I could use SPI3 on the LCD headers it would free many valuable GPIOs on
the expansion connector (and one bonus signal on J4).

Agree - Just go ahead and check - On the other hand, you would just have the
extra GPIOs on the LCD connectors :slight_smile:

Heh, indeed :slight_smile: Let's see how and what I end up using.

Btw, what happens if two pins have such mux values that they both would
have the same signal? Say, for example McSPI3_CLK. Not that I'm planning to
do such, just as a curiosity.

The signal would be routed to both balls - This is not intended, but might
give you some pretty smart debugging possibilities once in a while :slight_smile:
However TAKE CARE: Do not connect two balls to the same input signal in the
OMAP - Do only use the trick for output pins :slight_smile:

That really would create weird situations. Better to follow Gerald's advice and not to try it :slight_smile: