| Title: | DEChub/HUBwatch/PROBEwatch CONFERENCE |
| Notice: | Firmware -2, Doc -3, Power -4, HW kits -5, firm load -6&7 |
| Moderator: | NETCAD::COLELLA DT |
| Created: | Wed Nov 13 1991 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 4455 |
| Total number of notes: | 16761 |
Hello,
I just wanted to pass on how a plugging a 900 series module took down an
entire hub. The hub went through a reset as if it were powered on for the first
time. We then moved this to another "test" hub and it also went into a "reset"
condition.
That's when we noticed there was a bent pin on the DECconncentrator900MX.
Holding the conncentrator module upright with the pins facing you
Large connector at bottom of module
top
x x x x x
x x x x x
x x x x x
x x x x x
.
.
.
.
x x x x x
x x x x x
x x x x x
x x x x x ----------------------------------------\
\____this pin bent and touched this pin _ |
bottom
Unfortunately this occured at one of Digital's most visible customers at one
of their revenue generating sites and they were interested as to how a module
could take down a hub.
Well, I thought I would supply this information if anyone is interested
the how's and why's, as well as possible prevention.
Bob
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 2825.1 | Mating radius angle largest at lowest pins on connector... | NETCAD::BATTERSBY | Tue Oct 03 1995 16:06 | 15 | |
Bob those two pins have gnd and +5V (the right hand most pin
is the pin with +5v on it. So when the module is plugged into
the hub slot, the bent pin is probably pushed further against
the pin with +5V on it. As soon as another matrix connector gnd pin
mates with its pin in the backplane one has a short between gnd
and +5v. This will cause the power supplies to shut down.
The very lowest pins in the matrix connector are the very first
pins to mate to the backplane These pins are subjected to the largest
radius angle of mating of all the pins in the backplane. As other
pins further up in the connector are mated, the radius angle gets
smaller and smaller.
It is probably prudent ritual to inspect modules for bent pins prior to
plugging modules into the backplane.
Bob
| |||||
| 2825.2 | similar problem | TIMABS::OBERLE | Wed Jan 24 1996 09:09 | 25 | |
Hi,
a customer here have had a similar problem. He tried to insert a brand
new PORTswitch 900CP into his DEChub 900. He inserted the module onto
the lower edge of slot 1, moved it close to the connectors, and BEFORE
he could lift the lever to allow the module to snap-in, the whole Hub
switched off ! ALL LED's went off, LCD background light, P/S LED,
everything. After a couple of seconds the hub rebooted and continued to
work as before.
The customer tried the same module on a spare hub and could insert and
remove it several times without any problems. Visual inspection of the
module showed no bend pins.
Unfortunately this happened on a very large customer site to a productive
hub. They have lots of other Hub's and are now, unnecessary to say, VERY
worried ...
b.t.w. there was no entry in the errorlog
Has anybody an possible explanation for this problem ???
best regards,
Bernd
| |||||
| 2825.3 | GRANPA::DFAUST | Bad Things, man... | Wed Jan 24 1996 09:54 | 7 | |
What slot did you plug the module into? I had a problem with a hub
where any modules plugged into slots 5-8 caused a problem like that. It
turned out that the power lugs weren't properly attached on the etch. I
put in a note about this back in october or november.
Dennis Faust
| |||||
| 2825.4 | TIMABS::OBERLE | Wed Jan 24 1996 10:08 | 12 | ||
On this hub it was slot 1.
The same customer have had this problem before on another hub with
another module, fortunately it was a spare one and he decided to
ignore it, but now it's much more serious.
b.t.w. personally I have had the same problem a couple of month ago
on another site with a DECswitch 900EF, same effect ...
cheers,
Bernd
| |||||