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: grok who wrote (26118)	8/2/1999 8:51:00 AM
From: John Walliker	Respond to of 93625

KZNerd, RE: <Yet another issue is that conventional clocking arrangements fail when the wavelength of the clock signal is of the same order as the dimensions of the conductors. Rambus overcomes this by using two clocks, one for each direction of signal travel.> This is the first I've heard of this wavelength issue. Sounds like something worth checking into. I had two effects in mind: 1) When the electrical length of a transmission line is half a wavelength and the ends are open circuit, it will resonate strongly. It will take a long time for the ringing to die down and it will radiate interference very efficiently. The same will happen if it is a quarter wavelength long and one end is shorted ie being driven. Terminating the open end with a resistance equal to the characteristic impedance of the transmission line causes the signal to disappear into the resistor and become heat rather than being reflected. 2) When the length of half a clock cycle (for DDR) becomes shorter than the electrical length of the transmission line more than one bit of data will be propagating along it. If both are going in the same direction the clock will be in the right place to latch the data. When they are going in opposite directions the wrong data bit may be latched. This is why Rambus used two clocks travelling in opposite directions. As Scumbria has pointed out earlier, there is a turnaround issue with any bus system where signals travelling in one direction must die down before a new signal can be transmitted in the other direction. This can be overcome using ring architectures. To put some of these things into context, light travels in a vacuum at 30cm per nanosecond. In wires and printed circuit tracks it is around half that. A 400 MHz clock has a half-period of 1.25 ns. Therefore any transmission line longer than about 20 cm will have more than one data bit travelling along it at any time. This sets a fundamental distance related upper limit on the speed of any system which does not have its clock either embedded like Ethernet or travelling along side the data in the same direction like Rambus. So although it has not been reached yet, DDR and derivatives have a fundamental speed limit getting close, while Rambus does not. (Its limits are more to do with imperfections in the implementation and at extremely high frequencies they will occur when the WIDTH of the printed circuit tracks start to be close to a wavelength and multiple propagation modes become possible. That is a LONG way away. John