13 Mar
2002
13 Mar
'02
4:57 a.m.
--- Gunther Schadow <gunther(a)aurora.regenstrief.org> wrote:
Ethan, I'm glad to hear you are making headway in
fixing your
fried board...
I was up to my elbows in the 8200 last night. I think I've identified
the target pin on the DD11DK that is the destination of the damaged pin
in the first place - BB2 which is defined as a ground. This is odd because
on the paddle card, the signal is individually routed over pin 30 of JP2 and
is not tied to anything else on either end.
Having inspected the cable carefully, multiple times, I'm still scratching
my head as to how the damage could have happened in the first place. I
had notes written right on the Unibus cables where they stick onto the
back of the VAXBI - the notations match the docs (I have the DWBUA technical
manual with installation instructions). I can't even see how putting the
paddle card in backwards would have caused a massive problem. I'm afraid
to hook everything up and power it on again before I identify what was
wrong.
OTOH, I did drop the T1010 card in the VAX and look for its presence on
the bus. With the cables disconnected, the card fails self-test (Duh!)
but when I go to read the ident register of the DWBUA (E 20000000), I
get back FFFFFFFF, not the expected value (as documented in the manual).
Having written VMS VAXBI drivers in the past, I can say that this is not
good. The card "shows up" - the POST shows a "-0", so the VAXBIIC of
the DWBUA is detected, but it's exceedingly unhappy. I am suspecting
that the card is more fried than one chip. I'll keep working on it to
see if the card needs a rudimentary amount of attachment to the UNIBUS
to get far enough into its self-test to initialize the ident register,
but I'm not hopeful about it - I think it should show the world what
it is, even if it's unhappy.
Fortunately, I have docs for all of this. At this time, I suspect the
M9313 UET, the DD11DK, the cables and the T1010 card. I just have to narrow
down the list of suspects. I suppose I could drop the card in with the
chip at risk removed and a lead hanging out so I can monitor what's
happening on that pin. The pin itself happens to go to the outer-most
pad on the VAXBI bus, looking at the board, it's the farthest pad
from the VAXBIIC on the solder side. It goes
diagonally in more-or-less
a straight line to a pin of a DEC DC021C, 4 or 5, IIRC
(it's not in front
of me here). Either that pin sources a whopping amount of current
from the DC021 and it was shorted to ground by a bad
cable or a problem
in the BA11, or somehow a very wrong voltage came from the BA11
into
that pin. It's the only way I can see how a trace can get cooked and
the chip melted at that pin.
Back to the basement...
-ethan
=====
Visit "The Seventh Continent"
http://penguincentral.com/penguincentral.html
__________________________________________________
Do You Yahoo!?
Try FREE Yahoo! Mail - the world's greatest free email!
http://mail.yahoo.com/
13 Mar
13 Mar
5:50 p.m.
And here I thought it was a post about Bush Jr. George DWUBA.
--- Gunther Schadow
<gunther(a)aurora.regenstrief.org> wrote:
lgwalker(a)mts.net
bigwalk_ca(a)yahoo.com
Ethan, I'm glad to hear you are making
headway in fixing your
fried board...
I was up to my elbows in the 8200 last night. I think I've identified
the target pin on the DD11DK that is the destination of the damaged pin
in the first place - BB2 which is defined as a ground. This is odd because on
the paddle card, the signal is individually routed over pin 30 of JP2 and is not
tied to anything else on either end.
Having inspected the cable carefully, multiple times, I'm still scratching
my head as to how the damage could have happened in the first place. I
had notes written right on the Unibus cables where they stick onto the
back of the VAXBI - the notations match the docs (I have the DWBUA technical
manual with installation instructions). I can't even see how putting the paddle
card in backwards would have caused a massive problem. I'm afraid to hook
everything up and power it on again before I identify what was wrong.
OTOH, I did drop the T1010 card in the VAX and look for its presence on
the bus. With the cables disconnected, the card fails self-test (Duh!)
but when I go to read the ident register of the DWBUA (E 20000000), I
get back FFFFFFFF, not the expected value (as documented in the manual).
Having written VMS VAXBI drivers in the past, I can say that this is not
good. The card "shows up" - the POST shows a "-0", so the VAXBIIC
of
the DWBUA is detected, but it's exceedingly unhappy. I am suspecting
that the card is more fried than one chip. I'll keep working on it to
see if the card needs a rudimentary amount of attachment to the UNIBUS
to get far enough into its self-test to initialize the ident register,
but I'm not hopeful about it - I think it should show the world what
it is, even if it's unhappy.
Fortunately, I have docs for all of this. At this time, I suspect the
M9313 UET, the DD11DK, the cables and the T1010 card. I just have to narrow
down the list of suspects. I suppose I could drop the card in with the chip at
risk removed and a lead hanging out so I can monitor what's happening on that
pin. The pin itself happens to go to the outer-most pad on the VAXBI bus,
looking at the board, it's the farthest pad from the VAXBIIC on the solder side.
It goes diagonally in more-or-less a straight line to a pin of a DEC DC021C, 4
or 5, IIRC (it's not in front of me here). Either that pin sources a whopping
amount of current from the DC021 and it was shorted to ground by a bad cable or
a problem in the BA11, or somehow a very wrong voltage came from the BA11 into
that pin. It's the only way I can see how a trace can get cooked and the chip
melted at that pin.
Back to the basement...
-ethan
=====
Visit "The Seventh Continent"
http://penguincentral.com/penguincentral.html
__________________________________________________
Do You Yahoo!?
Try FREE Yahoo! Mail - the world's greatest free email!
http://mail.yahoo.com/ 
14 Mar
14 Mar
7:43 p.m.
Ethan Dicks wrote:
OTOH, I did drop the T1010 card in the VAX and look
for its presence on
the bus. With the cables disconnected, the card fails self-test (Duh!)
but when I go to read the ident register of the DWBUA (E 20000000), I
get back FFFFFFFF, not the expected value (as documented in the manual).
Having written VMS VAXBI drivers in the past, I can say that this is not
good. The card "shows up" - the POST shows a "-0", so the VAXBIIC
of
the DWBUA is detected, but it's exceedingly unhappy. I am suspecting
that the card is more fried than one chip. I'll keep working on it to
see if the card needs a rudimentary amount of attachment to the UNIBUS
to get far enough into its self-test to initialize the ident register,
but I'm not hopeful about it - I think it should show the world what
it is, even if it's unhappy.
Well, not necessarily. I remember having seen such behavior with
the KDB50 when I didn't install it quite properly. The new DWUBA is
coming along, in case you need it.
Fortunately, I have docs for all of this. At this
time, I suspect the
M9313 UET, the DD11DK, the cables and the T1010 card. I just have to narrow
down the list of suspects. I suppose I could drop the card in with the
chip at risk removed and a lead hanging out so I can monitor what's
happening on that pin. The pin itself happens to go to the outer-most
pad on the VAXBI bus, looking at the board, it's the farthest pad
from the VAXBIIC on the solder side. It goes diagonally in more-or-less
a straight line to a pin of a DEC DC021C, 4 or 5, IIRC (it's not in front
of me here). Either that pin sources a whopping amount of current
from the DC021 and it was shorted to ground by a bad cable or a problem
in the BA11, or somehow a very wrong voltage came from the BA11 into
that pin. It's the only way I can see how a trace can get cooked and
the chip melted at that pin.
What scares me about it is that you put in the new DWUBA and it
might just fry as much as the old one.
good luck,
-Gunther
--
Gunther Schadow, M.D., Ph.D. gschadow(a)regenstrief.org
Medical Information Scientist Regenstrief Institute for Health Care
Adjunct Assistant Professor Indiana University School of Medicine
tel:1(317)630-7960 http://aurora.regenstrief.org
8:10 p.m.
--- Gunther Schadow <gunther(a)aurora.regenstrief.org> wrote:
Ethan Dicks wrote:
> OTOH, I did drop the T1010 card in the VAX and look for its presence on
> the bus...
OK... moving right along, I have checked everything out step by step and
a) nothing smoked :-) , but b) nothing worked. :-(
> Fortunately, I have docs for all of this...
According to the docs, when I read the general purpose register at
200000f0, it tells me that the card failed test 3 which is an internal
RAM test. I swapped gate arrays with the other card (socketed) and
there was no change in the test behavior. Not surprising, really; the
original fault seems to have roached both boards, so the same parts
are at risk. Only one PCB was damaged, so that's fortunate.
There's something deeper wrong with this board that simple visual
inspection does not reveal.
What scares me about it is that you put in the new
DWUBA and it
might just fry as much as the old one.
Nope... I did a systematic test and had it all plugged in and not
smoking earlier tonight. I started by ensuring I knew which pin
on the Unibus cable and paddle card was connected to the fried trace -
it was Unibus pin AU2 - one of the grant signals. I checked for wierd
voltages at the T1010 end with the BA-11 powered on - nothing. Next,
I checked for direct shorts to ground. Nope. I then powered up the
BA-11 and not the BA-32 (well... the BA-32 was in standby which activates
the DEC power bus to turn on the BA-11) with all cables and boards in
place - monitoring the affected pin showed TTL-compatible voltages (between
1.8 and 3.5VDC). Finally, I did a quick on-check-off test... no odd
readings on AU2. I inspected the board, no warmth where there had been
problems before. I did a longer on-check-off test and verified that
I still got no yellow LED and a "-0" on the BI POST. Board still did not
show odd warmness.
From this point, I was confident that I wasn't
going to roach anything.
I stepped through the diagnostic procedure in the DWBUA
Technical Manual
for the section that covers "light does not come on". It was there that
it mentioned checking the base address + F0 (a BIIC register that is
available for any use the module designer sees fit - there are 4 total;
we used one in the COMBOARD-BI as the "window" register we implemented
in hardware ourselves with previous COMBOARD designs). In the case
of the DWBUA, it is a status register that indicates if the Unibus is
powered up or not, and what the results of the selftest were. Since I
read a 00030081, it tells me that test "03" was where it failed, and
the rest of the register is undefined (it should read 00000001 when
all is well - Unibus powered up, no snags with the diagnostics).
The new DWUBA is coming along, in case you need it.
Cool. I think I do. Looks like I have either an address fault or
a data fault with the "IRAM" on the card. Could be a RAM chip, could
be a logic chip. None of that stuff is socketed except a PAL or two.
At least I can drop it in my box without fear of letting out the smoke.
I wish I knew why the previous ones smoked. Perhaps the cable got
crimped. Perhaps I was using a different M9313 card on the bus and
that one had a short to ground on AU2 (I have more than 4 UETs and I
don't know which one was in the BA-11 8 years ago).
-ethan
__________________________________________________
Do You Yahoo!?
Yahoo! Sports - live college hoops coverage
http://sports.yahoo.com/
8166
days inactive
8168
days old
3 comments
4 participants
participants (4)
-
erd_6502＠yahoo.com -
ethan_dicks＠yahoo.com -
gunther＠aurora.regenstrief.org -
lgwalker＠mts.net