Initial state of DDR RAM pins?

Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

Threaded View
Hello all,

I was wondering if any one knows the state of the DDR RAM's
bidirectionnal pins upon power-on. And only upon power-on  - no clk and
all other pins to ground. I am not interested in making the DDR RAM
operate properly, I just want to power it on in order to measure the
input pin leakage current.
I scoured the DDR's datasheet. I did not find if the bidirectionnal
pins are or aren't hi-z upon power-on.

All help is appreciated.

Kind regards,

Re: Initial state of DDR RAM pins?

Quoted text here. Click to load it

This text is extracted from a Micron DDR datasheet. Not from the
datasheet listed below, but just the first one I had on disk.

  DDR SDRAMs must be powered up and initialized
  in a predefined manner. Operational procedures other
  than those specified may result in undefined operation.
  Power must first be applied to VDD and VDDQ simultaneously,
  and then to VREF (and to the system VTT). VTT
  must be applied after VDDQ to avoid device latch-up,
  which may cause permanent damage to the device.
  VREF can be applied any time after VDDQ but is expected
  to be nominally coincident with VTT. Except for CKE,
  inputs are not recognized as valid until after VREF is
  applied. CKE is an SSTL_2 input but will detect an
  LVCMOS LOW level after VDD is applied. After CKE
  passes through VIH, it will transition to a SSTL_2 signal
  and remain as such until power is cycled. Maintaining
  an LVCMOS LOW level on CKE during power-up is required
  to ensure that the DQ and DQS outputs will be
  in the High-Z state, where they will remain until driven
  in normal operation (by a read access). After all power
  supply and reference voltages are stable, and the clock
  is stable, the DDR SDRAM requires a 200Ás delay prior
  to applying an executable command.

That would suggest keeping CKE grounded while establishing
VDD/VDDQ, and then bringing up VREF/VTT, should place
DQ and DQS tristated.

The whole box of chocolates...
A 32Mx8 DDR400 TSOP...

An analog simulation file of the input stage, would be as
meaningful as what you are trying to do. But the vast
majority of manufacturers, don't give real info, but
stuff like IBIS. I've only managed to get Spice decks
that were accurate for conditions outside the rails a
couple of times, and those days are over. (And you have
to be real careful, and check their handywork, because
usually the decks will be wrong the first time.) It is
too much fun for most companies, to hide behind their
IBIS pseudo info.

I think 10 minutes on the phone with an apps engineer,
would tell you as much as a whole week spent in the lab.
Perhaps you should be attacking the problem from the comfort
of your desk :-) Either have the apps engineer verify
the proper leakage measurement technique (if you don't
want to reveal your secret project), or describe to
him/her exactly what crazy scheme you have in mind.


Re: Initial state of DDR RAM pins?


Quoted text here. Click to load it

Great suggestions.
I just sent micron a list of technical questions. Unfortunately, I only
could find an [online form to complete]. I couldn't seem to find a
support phone number. That would have been to much to hope for :-( .


Site Timeline