In message <200208272126.g7RLQ8t24829@ragno.ewranglers.com>, "Johan A. van Z
anten" writes:
>
>---In message <20020827173622.6686211D1D1@mail.bullers.net>
>>A few quick questions for the experts on the list...
>
>[smip]
>
>>So last night, I put all the stuff together (except for the speaker
>>box) and tried to get it to talk to my pc532 (since it has some
>>free serial ports).  Headless, and with no keyboard, I figured I'd
>>see something come out of the serial port at 9600 N81.  I didn't.
>>My breakout box tells me the lines are correct, but no data that
>>I can see.
>
> If the CPU was dead or not properly seated, that would fit the symptoms
>you describe. I believe i have seen improper (or bad) RAM configurations
>also cause this sort of failure.  The idea being if you don't have one
>good CPU, or a legal and functional RAM config, you are not even going to
>see the OpenBoot PROM (OBP) banner.

Hmm. Could be I didn't push on it hard enough, but I don't think
that's it.  I'll check again tonight just in case.  It has to be
in MBUS0 which is the lower connector closer to the back, right?
The upper module (if I had one) would be in MBUS1 then?

And something I didn't understand from the service manual *at all*
is how the keyboard lights can tell you if the MBUS, DRAM, or mother
board are busted.  Isn't that keyboard just connected to a Z8530
serial chip (or clone)?  Wouldn't all the parts to be tested be
needed to run the tests and report the results?  Does the POST
stuff just check for more subtle things and assume that the major
parts are at least 90% working?  That implied to me that maybe they
hid some small controller to run the tests in one of those big
ASICs somewhere.  But why would they bother?

> Do you have another SPARC-10 or 20 to test the CPU in? (Some CPUs work in
>both machines, but not all.  If in doubt, check the part number.) Or do
>you have another CPU to test? 50 MHz SuperSPARCs are pretty cheap now.

Unfortunately, no.  I only have one of each item at the moment.
I'd test the CPU in my LX, but it has a microSPARC chip soldered
in, not an MBUS module... (I'd add a smiley here, but it needs to
frown at the same time too.)  Can't test the memory out that way
either, although I have a pair of those, so in theory, swapping
them might help.

I was thinking about buying another keyboard and if necessary a
motherboard or power supply to get a working machine.  One thing
I just noticed while looking for possible replacement parts on
EBay, do SS10 power supplies have 3 fans on the outside of the
power supply frame, and SS20s have 2 on the inside?  If so, that
would make this an odd sort of beast.  An SS10 board with an SS20
power unit.

Along those lines, is the connector for the floppy, hard drives,
and power led the same on the SS10 and SS20?  If so, I could replace
the board with an SS20 board if I ever want a 20% boost in memory
bandwidth.  (Assuming the power supply connector is the same too.)
Although, given the price and availability, it would probably be just
as cheap to get a new stripped machine.

> L1-N (Stop-N) should reset NVRAM to defaults, although i would expect you
>to have seen something on ttya when you did L1-D. However, i do not
>remember what version of the OBP L1-D and L1-N were introduced (or if they
>were there as long as L1-A), and it's possible they are disabled because
>OBP security has been turned on.  If OBP security is enabled, i believe
>your only two choices are:

Didn't know about L1-N.  And since it was free and stripped clean
on arrival with no real info about it, it could easily be in secure
mode.  I should have a cg6 arriving in a couple of days, and that
may help in the debugging as well, I hope.

> Be advised that lots of things can go wrong when you start pulling chips
>off the motherboard, such as bent pins and ESD.  So if you aren't ready to
>buy a new motherboard or NVRAM chip, you probably don't want to do this.
>(FWIW, a bare SPARC-10 motherboard or the NVRAM chip should each cost
>about the same now: ~$20)

In 1983 I spent a tour of duty putting die in 68 pin PGAs.  Trust
me, I will never forget the fun of ESD and bent pins. 8^)  But at
the same time, it doesn't strike me as that big a deal either,
provided you know a couple of rules and follow them.

(Make sure your hands have enough moisture to conduct, and always
keep at least one finger on the metal case parts while the other
hand moves the parts into contact.  Only put stuff in or on
anti-static bags.  I know, I should have a ground strap, but I keep
losing them, and a proper anti-static work area is too expensive.)

Jon