>> !!  I didn't think the GX in a LX was socketed!  [...]
> [...] I pulled the board out of this one.  It doesn't take a rocket
> scientist to realize -- once you've seen the solder side of the PCB
> -- that the socket is "not native".

Oh my. :-/

> Given my speculations, above, I don't think I want to risk toasting
> the GX chip out of the SBUS card just to determine if the alleged
> student has a promising career as a technician [...]

I can - very much! - understand that POV.

>> I have at hand now an LX and a cg6 that have SunGX chips that appear
>> basically identical.  The only difference I can see is [...]
>> However, I own multiple cg6s (of which only one is near me now), and
>> they don't all use chips this similar; [...].  I have no idea
>> whether the various "big chips" are pin-compatible.

> I have three different cg6 SBUS cards here.  P/N's  501-1645,
> 501-1481 and 501-1672.

The one I have at ready hand is a 501-1996, and like your -1672, it has
basically 11 devices on it: the Brooktree RAMDAC, the "SunGX" PGA chip,
eight 42C8128s, and an AM27C256.  (There's also a 10H116, whatever that
is, a 74ALS1034, a 74ACT08, a 74FCT245T, two clock oscillators
(105.561000 and 92.9405 MHz), a discrete transistor, six diodes, a
whole slew of discrete resistors and capacitors (mostly on the solder
side), and four two-terminal devices I can't readily identify but I
suspect are inductors.)

It occurred to me that I have access to another cg6 (it's not mine but
is in a loaner machine I'm using).  I just pulled this one and looked
at it, and it's very similar, except that the PGA chip has been
replaced by another chip of about the same pin count with some
surface-mount technology I don't know the name for.  It's also marked
differently.  It has a logo consisting of the letters LSI in a box, and
printing

	L1A9765 ICFAA
	CH02334
	NND 9717 **
	KOREA

where the ** represents two equilateral triangles, each about one
character-cell wide, positioned as if they were squashed capital
deltas, bottom line matching the baseline of the other characters on
that line.

Turning the card over, I see a whole slew of vias under the chip, in a
vaguely PGAish pattern, but no pins passing through the card (well, at
least there - the SBus and video connectors do have such).  Other than
that, it looks pretty much the same: the "big chip", 8 chips of RAM,
the Brooktree, a 27C256, two clock oscillators, and assorted small
logic and discrete components.  (I note the RAM is faster; on the
-1996, it's "-10" (100ns, IIUC), whereas on this one, it's "-8" (80ns).)

This is a 501-2922, with an additional sticker reading "-03REV.51".
When I get back to my main card collection, I can describe them in more
detail - if you're still interested then, ping me in a few months,
around mid-January.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B