From: Steve on
Hi,

After reading over the documentation for DDR2 and the SSTL signalling
standard, I have a question about the role of termination in DDR2.

It appears to me that in addition to the usual termination function,
the resistors provide some sort of biasing function around the Vref
(1.8v/2) point. Is it the case that the IO on the DDR2 modules can
"swing" around that midpoint without the presence of termination
resistors, or are these necessary for correct operation (even using On-
Die Termination). What I would like to know is whether, in addition to
the unidirectional signals (Address, etc...) that do not have ODT, it
is necessary to terminate the Bidirectional signals (even with ODT)?

Thanks for your assistance,

Stephen
From: Symon on
Steve wrote:
> Hi,
>
> After reading over the documentation for DDR2 and the SSTL signalling
> standard, I have a question about the role of termination in DDR2.
>
> It appears to me that in addition to the usual termination function,
> the resistors provide some sort of biasing function around the Vref
> (1.8v/2) point. Is it the case that the IO on the DDR2 modules can
> "swing" around that midpoint without the presence of termination
> resistors, or are these necessary for correct operation (even using
> On- Die Termination). What I would like to know is whether, in
> addition to the unidirectional signals (Address, etc...) that do not
> have ODT, it is necessary to terminate the Bidirectional signals
> (even with ODT)?
>
> Thanks for your assistance,
>
> Stephen

Hi Stephen,

Have a look at Sheet 13 of this:-

http://www.xilinx.com/support/documentation/boards_and_kits/ml50x_schematics.pdf

There are no terminations on the ODT signals to the DDR2 SODIMM, because the
ODT takes care of it.

HTH., Syms.