Questions about MMC2+UART2 / daughterboard ?

Hello

I'd like to experiment with my beagleboard, but considering the
various USB problems, to increase its connectivity options and
reliability, I am mostly thinking about using the 2 other UARTs
instead.

So I've got some very naive hardware questions. Any help is appreciated.

I was first thinking about putting the SD port in SDIO mode (pretec
wifi) and moving the memory card to MMC2. From what I've read in the
SRM, RevC makes MMC2+UART2 possible (MUX0).

Any report on SDIO status then? The only messages I could find dated
from ~1 year ago and weren't very enthusiastic.

I am also wondering if anyone made a daughter board for the extension
connector to have at least UART2 and MMC2 (MUX0), with the proper
voltage reg. to interface the 1.8V signals to 3.3 or 5V - and in that
case, if they are for sale?

If there is no daugtherboard, I am thinking of wiring directly a
MMC-plus or MMC-mobile, which operate at 1.8v. Will it work as good as
I believe, without requiring anything else? For the UART2, I could
always put a max3218, but some premade solution would be better.

I'm also thinking about UART1, but it would then require using both J4
and J5 to have RTS/CTS along with RX/TX (MUX2). Beside making it
impossible to connect a LCD panel, wouldn't that also interfere with
the DVID port? Could the 4 pins be on either J4 or J5 in a future
release? (like UART2 MUX0 changes on revC)

A final question about UART3: there doesn't seems to be CTS/RTS
signals (nothing about them in the SRM). Any reason why UART3 is
exposed then, instead of say UART2?
Could a future release expose a UART with RTS/CTS on the IDC10 instead of UART3?

It'd be easier than having to dig UART2 or UART1 and wondering about
how to do that and the consequences it may have.

Hi Guylhem,

I was first thinking about putting the SD port in SDIO mode (pretec
wifi) and moving the memory card to MMC2. From what I've read in the
SRM, RevC makes MMC2+UART2 possible (MUX0).

Agree - No problem - You can set the Mux mode individually on each pin. It's
not a requirement, that all pins need to be in Mux 0. You can have some in
mux mode 0 while others are in mux mode i.e. 2 or 4...

If there is no daugtherboard, I am thinking of wiring directly a
MMC-plus or MMC-mobile, which operate at 1.8v. Will it work as good as
I believe, without requiring anything else?

I would expect so, but haven't tried :slight_smile:

I'm also thinking about UART1, but it would then require using both J4
and J5 to have RTS/CTS along with RX/TX (MUX2). Beside making it
impossible to connect a LCD panel, wouldn't that also interfere with
the DVID port? Could the 4 pins be on either J4 or J5 in a future
release? (like UART2 MUX0 changes on revC)

You are right, this would interfere with the DVI-D port as well. You might
be able to set the DVI framer in power down state and thereby be able to use
UART1 though. But of cause with no DVI-D port available...
I think the pin-out is unlikely to change since this would leave current
made LCD daughter cards outdated. Better accept the current pinning and use
two connectors... :slight_smile:

A final question about UART3: there doesn't seems to be CTS/RTS
signals (nothing about them in the SRM). Any reason why UART3 is
exposed then, instead of say UART2?
Could a future release expose a UART with RTS/CTS on the IDC10 instead
of UART3?

One guess might be, that you can do peripheral booting (using the ROM code)
on UART 3 only. Furthermore UART 3 supports IrDA and CIR (again as the only
UART). Instead of breaking backward compatibility I would instead
prefer/recommend that balls H18 and H19 (UART3 CST/RTS) are made accessible
on the connector or on test points in order to avoid the extra cost of level
shifting.

That being said - AFAIK there is a lot of Rev C boards in Stock.
You therefore shouldn't expect any recent change... Unfortunately :slight_smile:

It'd be easier than having to dig UART2 or UART1 and wondering about
how to do that and the consequences it may have.

UART2 is directly accessible on the expansion connector with no limitations
or drawbacks...

Best regards - I hope this helped?
  Søren

Hello

Agree - No problem - You can set the Mux mode individually on each pin. It's
not a requirement, that all pins need to be in Mux 0. You can have some in
mux mode 0 while others are in mux mode i.e. 2 or 4...

Didn't know that - even better.

I think the pin-out is unlikely to change since this would leave current
made LCD daughter cards outdated. Better accept the current pinning and use
two connectors... :slight_smile:

I just thought the multiplexing could be changed, so that UART1 would
go somewhere else and keep both J4 and J5 for DVI

But of cause with no DVI-D port available...

So UART1 is better left alone. I'll use UART2 with a max 3128, and
directly wire a mmc plus. I'll let you know what happens.

While I'm at it, any suggestion to easily convert the MMC2 pins to
3.3V so that a SD could also be used ?

Instead of breaking backward compatibility I would instead
prefer/recommend that balls H18 and H19 (UART3 CST/RTS) are made accessible
on the connector or on test points in order to avoid the extra cost of level
shifting.

Even better. Didn't know they existed. They should really be exposed
on the connector.

Best regards - I hope this helped?

Det hjælper!

Med venligt hilsen

Hi Guylhem,

I just thought the multiplexing could be changed, so that UART1 would
go somewhere else and keep both J4 and J5 for DVI

OK - Got your point. UART1 is as well available on balls AA8, AA9, W8 and
Y8.
Unfortunately only balls AA9 and W8 (RTS and CTS) are routed out from the
balls.
Currently these are controlling the two user LEDs on the board. One idea for
the future might be to route AA8 and Y8 out from the chip as well in order
for people to use UART1 in case they like...

While I'm at it, any suggestion to easily convert the MMC2 pins to
3.3V so that a SD could also be used ?

There are chips like:
http://focus.ti.com/docs/prod/folders/print/sn74avca406e.html
You are however missing the direction information pins on the OMAP => You
can't use it :frowning:
Without the direction pins, the transceiver would need to decode the data
send on the bus in order to decide on the direction, and to be honest I
don't think such a chip exist. I therefore think you are stuck at 1.8V.

Det hjælper!
Med venligt hilsen

Hmm - A Secret Danish Spy :-)Where did you learn this? Have you ever visited
Denmark?

Best regards - Good luck
  Søren

Hello,

Unfortunately only balls AA9 and W8 (RTS and CTS) are routed out from the
balls. Currently these are controlling the two user LEDs on the board.

Damn. I had read UART3 and I was already thinking about removing the
LEDS and adding stripping wire to have RTS/CTS on the IDC10.
:frowning:

the future might be to route AA8 and Y8 out from the chip as well in order
for people to use UART1 in case they like...

That would be great. Even if only as solder pads.

Without the direction pins, the transceiver would need to decode the data
send on the bus in order to decide on the direction, and to be honest I
don't think such a chip exist. I therefore think you are stuck at 1.8V.

I'll leave that to others and stick to the original plan for now -
soldering a MMC which operates at 1.8 V to validate my software.

Hmm - A Secret Danish Spy :slight_smile:

Nej ikke - jeg forstaar dansk! Er du selv dansker? (Søren + .dk ? I
*do* think so !)

I have visited Danmark (AArhus + Kjøbenhavn), I totally love danish
culture (from Aqua to Lars von Trier, the smørrebrod and the grøn
tuborg, the language and way of life etc!)- it was just a long time
ago and my language is falling. There's little opportunity to practice
danish outside denmark, and even less outside europe. I occasionally
stumble on websites or news- still understand most of it. I also
still have a red tshirt with a dannebrog that I wear occasionally !

I may have an opportunity to spend a year or two in Danmark in the
future for research work. Maaske. Det skal vaere tid for laerer igen?

Søren Steen Christensen wrote:
...

While I'm at it, any suggestion to easily convert the MMC2 pins to
3.3V so that a SD could also be used ?

There are chips like:
http://focus.ti.com/docs/prod/folders/print/sn74avca406e.html
You are however missing the direction information pins on the OMAP => You
can't use it :frowning:
Without the direction pins, the transceiver would need to decode the data
send on the bus in order to decide on the direction, and to be honest I
don't think such a chip exist. I therefore think you are stuck at 1.8V.

Actually, for MMC2 there are direction pins available on the OMAP (AE4,
AH3, AF19 and AE21 for data and AF3 for CMD). I just have no idea if
available on the Beagle Board...

They are not.

Gerald

Hi Søren,

Søren Steen Christensen wrote :

While I'm at it, any suggestion to easily convert the MMC2 pins to
3.3V so that a SD could also be used ?

There are chips like:
http://focus.ti.com/docs/prod/folders/print/sn74avca406e.html
You are however missing the direction information pins on the OMAP => You
can't use it :frowning:
Without the direction pins, the transceiver would need to decode the data
send on the bus in order to decide on the direction, and to be honest I
don't think such a chip exist. I therefore think you are stuck at 1.8V.

TI has this kind of chip (which also includes a port multiplexer) :
http://focus.ti.com/lit/ds/symlink/txs02612.pdf

Eric

TI has this kind of chip (which also includes a port multiplexer) :
http://focus.ti.com/lit/ds/symlink/txs02612.pdf

Cool thing - I have not seen that one before - Thanks for pointing me to it
:slight_smile:
Unfortunately it seems to be a little hard to get unless you want to by
2500pcs from Digikey :slight_smile:
  Søren

Hmm – I thought I knew the OMAP and the BeagleBoard HW pretty well, but apparently I didn’t knew it well enough J

I just double checked the direction pins for MMC2 (Thanks Alexander for pointing me to the pins), and they are actually all (except one) available on the Beagle Expansion connector if you can do with a 4 bit interface (which is totally OK for a SD(IO)-card J)? The only signal missing is mmc2_dir_dat2 which however can be wired from the SYS_BOOT4 pin J

All in all – No problem using MMC2 in 4-bit mode with level conversion using this solution. Alternatively you could try the txs02612-approach (described in the previous post) in case you can get hold on such a chip J

Best regards

Søren

Søren Steen Christensen a écrit :

TI has this kind of chip (which also includes a port multiplexer) :
http://focus.ti.com/lit/ds/symlink/txs02612.pdf

Cool thing - I have not seen that one before - Thanks for pointing me to it
:slight_smile:
Unfortunately it seems to be a little hard to get unless you want to by
2500pcs from Digikey :slight_smile:

It's actually not available from Digikey (they have a minimum quantity of 1 so no need to purchase a full reel once they get it in stock - a batch is announced for the 15 june on TI's website).

But on : http://focus.ti.com/docs/prod/folders/print/txs02612.html
click the link "Add Free Samples to Cart" which allows you to get up to 5 pieces after registration.

Eric