[rescue] let's revisit the whole FC-AL thing

[nick at snowman.net]

I was thinking 8 disks per board for a total of 24.  it simplifies
addressing tremendously.  (since it's done the same way scsi is, just 8
bits instead of 3 or 4).  6 power pins, a bunch of ground, activity led
shouldn't need to leave the board.  I don't Belive that discrete
components will be required per board, but the "host" board (that three of
these trays connect to) will require several, and a bunch more if the
sleds for the PSs are "stupid", forceing it to do the power handling.  I
can send you my prelim schematics, but they're... erm. Messy.
	Nick

On Tue, 26 Feb 2002, George Adkins wrote:

> > > > Dude... Do you have any idea how bitchy these would be to hand
> > > > assemble?  I'll help if it happens, but gah.  We're talking ~1600
> > > > solder points ber disk board give or take.
> > >
> > > Umm, I don't think so.
> > >
> > > how about 40-60 solder joints per card, max.
> 
> > Erm, you're talking tcards.  I'm not terribly intrested in tcards.  I can
> > provide the required connections though.
> 
> Okay, your SSA boards need to host what, 10 drives?
> 
> that's :
> 
> tx+, tx-, rx+, rx-, 8 ID pins, what, 8 power pins? 2 for activity LED, what 
> else?  that looks like about 20 per connector + discrete components, maybe 
> 40+ the interconnects to whatever the backplane boards connect to?
> 
> Okay, maybe 300 per board at the most...  I've done more wire-wrap than that 
> on one board...
> 
> I don't think it'll be that difficult.  all I need is the physical layout 
> parameters for the board and time to lay out the traces so that they're all 
> the right length (signal) or width (power).
> 
> George
> _______________________________________________
> rescue list - http://www.sunhelp.org/mailman/listinfo/rescue