Silicon Investor (SI) -- The First Internet Community

STOCKTALK

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 : Rambus (RMBS) - Eagle or Penguin -- Ignore unavailable to you. Want to Upgrade?

To: Dave B who wrote (19373)	4/27/1999 3:04:00 AM
From: Dr. Id	Respond to of 93625

Message to your technical friend: Well, duh!!! :-) JB

To: Dave B who wrote (19373)	4/27/1999 3:09:00 AM
From: Alan Bell	Read Replies (1) \| Respond to of 93625

Dave, A few technical points - The signaling convention that Rambus uses (RTL) is not all that revolutionary. Similar conventions have been used in the past in the form of GTL+ which is used for the Intel processor bus among other things. There are a number of newer conventions that are also similar - SSTL and HSTL. These all have the convention that they have reduced voltage swing and use a reference voltage. A scheme like RTL might actually have better noise immunity than LVTTL with its larger voltage swing. This is because the former uses a reference voltage to compare the signals against. I don't think your comment about driving to 1/2 the level is correct (which is called series termination.) I believe that RTL is fully terminated meaning there are no reflections. DDR SDram uses SSTL_3 for the clock and data lines (and uses LVTTL for address and control.) This is a limited swing, reference voltage based signaling convention. The beauty of the Rambus scheme is that they took a systemic approach and started from scratch. They eliminated the stubs by having the signals both go into and come off the Rimm module. They use 2 clocks - one for data going into the chip and one for data sourced from the chip. The clock follows the data so is identical in length. They use a good signaling convention, etc. -- Alan

To: Dave B who wrote (19373)	4/27/1999 9:47:00 AM
From: TigerPaw	Read Replies (1) \| Respond to of 93625

no radio signals I suppose this comes down to a matter of definition. Your friends assessment of the technology agrees with mine. I used the term radio signal not because there is an underlying carrier frequency, there is not, but because the information is carried primarily though voltage changes and not current changes. information was encoded onto some underlying carrier The Rambus signal is a high enough frequency to be considered part of the radio spectrum. The protocol calls for a minimum number of transistions in each packet. This means that you never get a string of all ones or all zeros that is so long that it makes the frequency fall below a certain threshold. This allows the signal to overcome capacitance quicker and with less distortion. A carrier is only necessary if the information signal is of too low a frequency (say voice or pictures) to act as it's own radio source. Lastly, Rambus systems are still digital systems. This could be a case of splitting hairs, but the signal which is travelling between the processor and the RIMM has many analog characteristics. One of the most important is the reflections from the unterminated line. This reflection is critically dependant on the length of the line and it's impedance. The problems involved in getting this to work, especially across a plug-in connector, is more similar to the problems solved by radio engineers than digital engineers. It is true that the signal can be considered digital; so long as the correct reflections are achieved the end result is sending ones and zeros and not intermediate values. TP