FreeLists / si-list / [SI-LIST] Re: Resistors to use for DDR termination

[Bill Wurst <billw@xxxxxxxxxxx>]

Dimiter,

You might want to consider embedded resistors (see 
http://www.ohmega.com/).  It is not necessary to add layers to the 
stackup as series terminations can be added to a signal layer and shunt 
terminations can be added to a plane layer.  Advantages:  no stub and 
lower parasitics.  A few disadvantages:  cost of material, additional 
processing cost, higher tolerance (±10% to ±20% or more, depending on 
design, which you must do yourself), and the inability to change the 
value once the board comes back.

Otherwise, I'd use 0603 or if space is a real concern, 0402.

     -Bill


       /************************************
      /         billw@xxxxxxxxxxx         /
     /                                   /
    / Advanced Electronic Concepts, LLC /
   /           www.aec-lab.com         /
   ************************************
=============================================================
Dimiter Popoff wrote:
> Scott,
> I have done more designs than I care to rememeber and none of them needed a 
> redesign.
> By just looking at the board I have more simulation results available than a 
> simulator could
> provide in most cases - and this is one of these most cases. For the few ones 
> where
> simulation is needed, my organic simulator will alert me on the spot.
>   It is just because I am going low risk that I have put 4 resistors per 
> line, even though
> all these lines are < 2 inches long, no connectors between. Now tell me I 
> need to simulate
> that... I am not a customer of consulting firms, in fact I have been/am doing 
> designs for
> other companies myself. All I am wondering about is what other people who 
> have _designed_
> actual borads do use as resitors - a handfull of 0603, 0402, what? Where do I 
> get them?
> Any suitable arrays? I did try searching without much luck, otherwise I would 
> not have
> asked this...
> 
> Dimiter
> ------------------------------------------------------
> Dimiter Popoff               Transgalactic Instruments
> 
> http://www.tgi-sci.com
> ------------------------------------------------------
> 
> 
> 
> -------Original Message-------
> 
>>From: "Scott McMorrow" <scott@xxxxxxxxxxxxx>
>>Subject: Re: [SI-LIST] Re: Resistors to use for DDR termination
>>Sent: Jul 12 2005 18:23:22
>>
>> Dimiter,
>> 
>> Choosing not to simulate this interface is a choice you can make, butit 
>> certainly is not anything I would call engineering.  It is, however,not 
>> clear to me whether you are talking about just series terminationresistors 
>> or both series termination and Vtt parallel termination.  Ifyou already have 
>> Vtt parallel termination, then your bus may very wellwork. But again, it 
>> depends.  It depends on the particular devices andpackages that you use.  It 
>> depends on the Vtt termination resistorvalues that you are using.  
>> (Freescale, by the way, has an applicationnote where they recommend the use 
>> of 28 ohm termination to Vtt, whichis generally just a bit too stiff.)  It 
>> also depends on how much riskyou are willing to take in your design.
>> 
>> The simple answer to your question is "it depends,"  just as Mikeindicated.  
>> After many simulations of many DDR interfaces over manyyears, I've not see 
>> one yet that would work over PVT extremes withoutany termination whatsoever. 
>>  I've seen a few that would work with 1 or2 memory devices with just Vtt 
>> termination, and a majority that requireboth Vtt termination and series 
>> termination.  DDR devices havesmokinFastTM drive strengths and edge rates, 
>> which have atendency to overdrive the inputs of sensitive devices.  
>> Generallyseries termination can be removed when the controller has a 
>> heftydynamic overshoot tolerance, and can sustain massive amounts of 
>> currentdue to SSI on the IO ground ring, without causing threshold 
>> violationsat the receivers.
>> 
>> regards,
>> 
>> scott
>> 
>> Scott McMorrowTeraspeed Consulting Group LLC121 North River 
>> DriveNarragansett, RI 02882(401) 284-1827 Business(401) 284-1840 
>> Faxhttp://www.teraspeed.comTeraspeed®; is the registered service mark 
>> ofTeraspeed Consulting Group LLC
>> 
>> Dimiter Popoff wrote:
>> 
>> thanks for the reply. Here is some more info:I will actually be soldering 
>> two DDR chips (x16 each) as close as possible to the 5200 chip.This is why I 
>> afforded at least the thoughts of no termination at all.My termination 
>> voltage is designed conservatively enough, no problems expected 
>> there.Simulating a DDR design at the conditions above sounds like a waste of 
>> time to me,and please don't go into suggesting more simulation, I am only 
>> looking for theright package/vendor/distributor of the resistors I have to 
>> use, I know the restof the 
>> story.Thanks,Dimiter------------------------------------------------------Dimiter
>>  Popoff               Transgalactic 
>> Instrumentshttp://www.tgi-sci.com-------------------------------------------------------------Original
>>  Message-------
>> 
>> From: "Mike Greim" Subject: [SI-LIST] Re: Resistors to use for DDR 
>> terminationSent: Jul 12 2005 17:16:26 Hi Dimiter,  Unfortunately the answer, 
>> as it almost always is, is it depends.  How many dimms=20 you are supporting 
>> on the chanel, supported configurations, if the design is embedded, etc all 
>> impact the termination solution space.  Simulation is really the answer.  I 
>> would also take care in the design and decoupling of my VTT island as I=20 
>> have seen this be the source of channel failure on a number of designs.   
>> You also need to figure out how much margin you are comfortable dealing 
>> with.   This will also impact the=20 various length matching requirements 
>> you have for the interface. Leverage your vendor and they may be able to 
>> give you an idea regarding your degrees of flexibility.  They may even be 
>> able=20 to simulate a given topology to assess viability.  Let me know if I 
>> can be of further assistance.  I just finished=20 fixing one of these 
>> interfaces for someone wh
o
> 
>   cut too many=20 corners.  Best Regards,  Michael Greim Platform Engineering 
> Newisys, Inc. 512-340-9050, ext 232    (V) 512-740-9097                  (C)  
> We will either find a way or make one     -Hannibal  And all this science 
> they don't understand is just my job six days a week........  The time is 
> gone, the email's over Thought I'd something more to say.......   
> -----Original Message----- From: si-list-bounce@xxxxxxxxxxxxx = 
> [mailto:si-list-bounce@xxxxxxxxxxxxx]On Behalf Of Dimiter Popoff Sent: 
> Tuesday, July 12, 2005 8:41 AM To: si-list@xxxxxxxxxxxxx Subject: [SI-LIST] 
> Resistors to use for DDR termination   I am routing a board with an MPC5200 
> (freescale/mot) which has a DDR interface (133 MHz clock). What sort of 
> termination resistors = (resistor arrays?) do I put? Thera are just so many 
> of them, I begin wondering how forgiving the 133 MHz interface is and could I 
> get away by just routing without any terminating restors at
> -------Original Message-------
> ------------------------------------------------------------------
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
http://www.freelists.org/webpage/si-list

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List FAQ wiki page is located at:
                http://si-list.org/wiki/wiki.pl?Si-List_FAQ

List technical documents are available at:
                http://www.si-list.org

List archives are viewable at:     
                http://www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu