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: John Walliker who wrote (31412)	10/2/1999 5:24:00 PM
From: Bilow	Respond to of 93625

Regarding the emissions from that system shipped by one of the major PC manufacturers using PC100 memory. The thing was only just below the limit at the CPU clock frequency (450 MHz I think). 450MHz is consistent with radiation from PC100 memory (see note later), but it is also consistent with radiation from a 100MHz FSB, for instance. But I agree that the 450MHz you saw was likely associated with SDRAM, for reasons I will explain later. The worst radiation from any machine I've ever tested always came from the highest frequency signals present on the system. I've seen more problems from the clock that goes to memory than from the data bus of the memory, for instance, even though there are far more data lines than clock lines. But the engineers probably terminated the clock lines, and left the data lines running free. This would be consistent with design practice similar to the one that I used to estimate DDR power consumption (when I got figures that were a tiny fraction of the ones you calculated for it.) Suffice it to say, that the duty of the engineers is to produce the cheapest machine that will meet its specifications. It would be great if they gave you extra pay for over-exceeding FCC compliance, but this is not something that the consumers care much about. The objective is to build the cheapest machine possible, not gold plate everything. For this reason, it is natural to not terminate the PC100 data bus, but terminate its clock. I think that this is probably how your 450MHz machine was designed, and that this is standard industry practice. No surprises. DDR memory designs use SSTL_2, rather than LVTTL_3. There are several differences between the technologies, and these differences result in SSTL_2 radiating considerably less energy (as the engineers who wrote its specifications intended.) The first is that SSTL_2 uses a lower voltage swing than LVTTL_3. The second is that SSTL_2 allows termination options that would reduce noise even further, if required. The thing that would be best for reducing radiation, I believe, would be for the manufacturers of memory to start putting 33 ohm series resistors on all their data lines. But this isn't an area where I am much of an expert, so WDIK. It seems that we are here discussing the seamy underbelly of electronics, the parts that consumers never find out about. I think this is great! I love this stuff! -- Carl Note on radiation signal patterning: For a 100MHz "square" wave, the frequencies that make up the wave are at 100MHz, 300MHz, 500MHz, 700MHz, etc. So there is no 450MHz radiation output from 100MHz clocks. But a 100MHz bus typically produces radiation at 50, 150, 250, 350, 450, etc. This would be consistent with what John Walliker is saying, if the 450MHz radiation was coming from the data bus, rather than the clocks.