SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : Advanced Micro Devices - Moderated (AMD)
AMD 203.50-0.6%3:53 PM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: combjelly who wrote (65676)12/14/2001 1:16:05 PM
From: Ali ChenRead Replies (1) of 275872
 
"I dunno, AMD seems to understand the problems. Now either they fail to communicate the solutions to the board manufacturers, or the board companies don't have the proper engineering resources to understand what they are being told."

Unfortunately, known facts and experience do not support
your assessment. Very unfortunately.

"Couldn't figure out what kind of slew rate would be important here, but now it occurs to me that the setting could be for changing the impedance of the bus terminators."

It is impossible to figure this out by looking at the
board:-) It is equally impossible if you are using a
300-MHz scope, especially if you do not specify what
kind of probe do you use (and possibly have no clues
about differences in probes at all). [by "you" I do
not mean you personally, just in general]. It is equally
impossible if the board has no properly grounded
low-proximity probe pads.

Also, things are "sligtly" more complicated than old
"bus terminator".
In modern peer-to-peer bi-directional connections
there is no bus terminators, they usually use so-called
"reflected wave switching" technique, so in receive
mode the line is unterminated, only some weak package
capacitance/inductance.

However, to drive the line effectively, you need to
match the driver impedance with the line impedance,
to have a nice symmetrical signal swing. To control the
driver impedance, a small built-in analog-to digital
converter is used, and the value of the impedance of
a group of drivers is determined by an external resistor,
40Ohms, or 56 Ohms, or whatever, to match particular
board design. There could be separate resistors that
controls pullups and pulldowns independently. Usually
the settings are not very critical (if the design
has postive margins<g>).

The slew rate is completely independent parameter,
programmed usually via a separated register, and
separately for different signal groups. The value
depends, too much can cause bad reflections and high
ground bounce on certain data patterns.

There must be certain basic discipline and background
to handle the signal integrity issues.

- Ali
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext