> All: > > Hi. A colleauge and myself are in the position of defending a particular > bus standard that we chose to transfer data. I won't say what we selected > (not yet - I will in a subsequent mail message), but here are some of the > factors. > > - We need a respectable data transfer rate (>1.0 GB/s). > - We have multiple boards with different power supply generation on each > board. > - We have 2 controlled differential impedance connectors to traverse (we > can change them to single ended if need be). > - We want to minimize the number of signals (and minimize ASIC/FPGA pins, > board routing layers, etc.). > > We can do, say, 32 signals at 250 Mb/s each or 16 signals at 500 Mb/s > each (both source synchronous). We can also do, say, 4 signals at 2.5 Gb/s > each to give roughly the same bandwidth (8B/10B). The term "signal" is > also being questioned. Do we need single ended or differential (management > likes less lines)? > > As we are all aware, the faster we pump a bus, the higher BER we will > get. Our question is: if you had to make this decision, which type of bus > would you implement (2.5 Gb serial with a cumbersome link recovery > protocol due to higher BER or go much wider and slow things down to an > almost non-existent BER with no link recovery, or something in between)? > Where is the line drawn on the speed scale that determines whether link > recovery is necessary (assume an enterprise class system)? > > Also, would you prefer 16 differential signals at 500 Mb/s (32 lines) or > 32 single ended signals at 250 Mb/s (32 lines)? Who has the better bit > error rate and noise margin? Where do you think the industry is going > (faster serial lines with link recovery such as IB or wider single ended > standards such as IMB and Hublink)? > > I understand that I've left this rather vague (on purpose). I would > appreciate general discussions on where you think things break (require > link recovery), advantages of each type of bus, experiences, etc. Any and > all thoughts are welcome. > > Regards, > > Matt > > Matt Ruston > Signal Integrity Engineer > Voice: 508-249-5223 > Fax: 508-497-8010 > ------------------------------------------------------------------ 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: //www.freelists.org/webpage/si-list For help: si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field List archives are viewable at: //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