USB port blew -- why?

[larrymoencurly]

One of the ports of my NEC-based USB 2.0 card quit working because the
National Semiconductor LM3526 power controller chip failed. I'm
pretty sure that I didn't zap it with static (I always touch the
outher metal shell to the computer case before plugging it in) or try
to plug it in backwards. The data sheet for the LM3526 says that it's
protected against voltage, current, and even temperature, so what
happened?

National Semiconductor says that each USB port must have a 120uF or
larger tantalum capacitor between +5V and ground for protection
against transients that occur during hot-plugging, and NEC's example
schematic for their uPD720100A USB 2.0 chip shows the LM3526 using
150uF aluminum in parallel with 0.1uF ceramic. But my USB card has
only a 100uF aluminum capacitor and maybe a ceramic capacitor in
parallel for this. Could this be why the USB port got zapped? Would
it help to solder a tantalum in parallel as well? Will it hurt to add
it?

 

[NSM]

....
| National Semiconductor says that each USB port must have a 120uF or
| larger ....................................... But my USB card has
| only a 100uF aluminum capacitor and maybe a ceramic capacitor in
| parallel for this....

Close enough!

N

 

[daytripper]

...
| National Semiconductor says that each USB port must have a 120uF or
| larger ....................................... But my USB card has
| only a 100uF aluminum capacitor and maybe a ceramic capacitor in
| parallel for this....

Close enough!

Except the tantalum cap would have much lower esl & esr figures than an
aluminum cap, making it more effective.

otoh, tantalums can have a brilliant failure mode ;-)

 

[James Sweet]

One of the ports of my NEC-based USB 2.0 card quit working because the
National Semiconductor LM3526 power controller chip failed. I'm
pretty sure that I didn't zap it with static (I always touch the
outher metal shell to the computer case before plugging it in) or try
to plug it in backwards. The data sheet for the LM3526 says that it's
protected against voltage, current, and even temperature, so what
happened?

Stuff just fails occasionally with no good explaination, could be it was
just a defective part that eventually quit.

 

[NSM]

|
| | > One of the ports of my NEC-based USB 2.0 card quit working because the
| > National Semiconductor LM3526 power controller chip failed. I'm
| > pretty sure that I didn't zap it with static (I always touch the
| > outher metal shell to the computer case before plugging it in) or try
| > to plug it in backwards. The data sheet for the LM3526 says that it's
| > protected against voltage, current, and even temperature, so what
| > happened?
|
|
| Stuff just fails occasionally with no good explaination, could be it was
| just a defective part that eventually quit.

Or as I say, "Everything works until it breaks". Then there are the three
brand new tubes I took off the shelf one at a time, each of which was bad.
Still can't get over that one.

N

 

[David Maynard]

|
| | > One of the ports of my NEC-based USB 2.0 card quit working because the
| > National Semiconductor LM3526 power controller chip failed. I'm
| > pretty sure that I didn't zap it with static (I always touch the
| > outher metal shell to the computer case before plugging it in) or try
| > to plug it in backwards. The data sheet for the LM3526 says that it's
| > protected against voltage, current, and even temperature, so what
| > happened?
|
|
| Stuff just fails occasionally with no good explaination, could be it was
| just a defective part that eventually quit.

Or as I say, "Everything works until it breaks".

HEY!! That's MY line.

 

[larrymoencurly]

Stuff just fails occasionally with no good explaination, could be
it was just a defective part that eventually quit.

I am eternally grateful for your answer, which not only completely
solved my problem with the USB port but will also magically turn the
rest of my life into bliss. So please give me your address so that I
can send you a $500 check, no, make that a blank check -- your advice
was that good.

 

[Tom MacIntyre]

I am eternally grateful for your answer, which not only completely
solved my problem with the USB port but will also magically turn the
rest of my life into bliss. So please give me your address so that I
can send you a $500 check, no, make that a blank check -- your advice
was that good.

Believe it or not, it was a good and reasonable response, given the
information he had to work with.

Tom

 

[keith]

Except the tantalum cap would have much lower esl & esr figures than an
aluminum cap, making it more effective.

otoh, tantalums can have a brilliant failure mode ;-)

So do engineers when they're bringing up a system where all
of the tantallums were inserted backwards. You want to see fireworks!
(well,it was 25 years ago - I've mostly recovered and the tinninus
isn't so bad. <twitch>)

 

[David Maynard]

So do engineers when they're bringing up a system where all
of the tantallums were inserted backwards. You want to see fireworks!
(well,it was 25 years ago - I've mostly recovered and the tinninus
isn't so bad. <twitch>)

LOL. Yeah, I'll bet.

You get a similar result plugging non keyed circuit boards into a vertical
card rack 180 degrees reversed.

 

[NSM]

| So do engineers when they're bringing up a system where all
| of the tantallums were inserted backwards. You want to see fireworks!
| (well,it was 25 years ago - I've mostly recovered and the tinninus
| isn't so bad. <twitch>)

I recall being told of a large TTL circuit board which was powered up with
reverse polarity. This was noticed due to the considerable amount of heat
being given off.

Everyone was mystified when the correctly rewired board worked OK. I guess
TTL is a little more tolerant than is generally believed.

N

 

[Craig Hart]

forgive me for not immediately recommending you open the controller chip and
place it under your scanning electron microscope to search for signs of
electromigration...

 

[James Sweet]

"James Sweet" <[email protected]> wrote in message

I am eternally grateful for your answer, which not only completely
solved my problem with the USB port but will also magically turn the
rest of my life into bliss. So please give me your address so that I
can send you a $500 check, no, make that a blank check -- your advice
was that good.

Just what sort of answer did you expect? I seem to have misplaced my crystal
ball and I can't find my spell book either. How is anyone supposed to tell
you why a chip failed? It just happens.

 

[NSM]

|
| > Stuff just fails occasionally with no good explaination, could be
| > it was just a defective part that eventually quit.
|
| I am eternally grateful for your answer, which not only completely
| solved my problem with the USB port but will also magically turn the
| rest of my life into bliss. So please give me your address so that I
| can send you a $500 check, no, make that a blank check -- your advice
| was that good.

Assuming you are being sarcastic (it's not clear), I refer to this as a
CatRan question, i.e., "I was typing away on my computer and I was holding
the keyboard above my head and I was watching the screen through my toes
when all of a sudden the cat ran over my stomach and I noticed the screen
blinked twice. What causes that"?

N

 

[keith]

LOL. Yeah, I'll bet.

You get a similar result plugging non keyed circuit boards into a vertical
card rack 180 degrees reversed.

A good reason to hang the mechanical designers by the short things. One
*should* have to go to great lengths to pluch dangerous things in
backwards. In this case they did just that. ...and even complained about
how hard it was!

 

[keith]

| So do engineers when they're bringing up a system where all
| of the tantallums were inserted backwards. You want to see fireworks!
| (well,it was 25 years ago - I've mostly recovered and the tinninus
| isn't so bad. <twitch>)

I recall being told of a large TTL circuit board which was powered up with
reverse polarity. This was noticed due to the considerable amount of heat
being given off.

Everyone was mystified when the correctly rewired board worked OK. I guess
TTL is a little more tolerant than is generally believed.

TTL protection diodes are as strong as moose! Unless the chips got hot
enough to let the magic smoke out, they'll likely survive, though perhaps
somewhat injured.

 

[David Maynard]

A good reason to hang the mechanical designers by the short things. One
*should* have to go to great lengths to pluch dangerous things in
backwards. In this case they did just that. ...and even complained about
how hard it was!

Hehe. Well, with the one I mentioned it was an experimental prototype built
by the design engineer.

 

[Rich.Andrews]

|
| > Stuff just fails occasionally with no good explaination, could be
| > it was just a defective part that eventually quit.
|
| I am eternally grateful for your answer, which not only completely
| solved my problem with the USB port but will also magically turn the
| rest of my life into bliss. So please give me your address so that I
| can send you a $500 check, no, make that a blank check -- your advice
| was that good.

Assuming you are being sarcastic (it's not clear), I refer to this as a
CatRan question, i.e., "I was typing away on my computer and I was
holding the keyboard above my head and I was watching the screen through
my toes when all of a sudden the cat ran over my stomach and I noticed
the screen blinked twice. What causes that"?

N

Reminds me of the time an aquaintance went to lunch with me and a few of
the other engineers. He was in a rather "know it all" mood. I started
plying him with questions about how water softeners work, what causes
gravity, etc. He was doing good until I asked him, "Ever wake up on a
wednesday and think it was tuesday all day and then wake up the next day
and everything is ok again? What is that called?" He shut up after that.

r

 

[Lionel]

TTL protection diodes are as strong as moose!

Well, the individual diodes aren't all that strong, but there's one on
every single I/O pin on every chip, so on a big PCB, the load will be
spread over a *lot* of diodes.

Unless the chips got hot
enough to let the magic smoke out, they'll likely survive, though perhaps
somewhat injured.

Yup. The best I've personally seen was the time I accidentally plugged a
2716 EPROM backwards into a programmer. It lit up like a xmas tree
through the UV window, but worked fine when I turned it around the right
way.

 

[Grumble]

Reminds me of the time an aquaintance went to lunch with me and a
few of the other engineers. He was in a rather "know it all"
mood. I started plying him with questions about how water
softeners work, what causes gravity, etc. He was doing good
until I asked him, "Ever wake up on a wednesday and think it was
tuesday all day and then wake up the next day and everything is
ok again? What is that called?" He shut up after that.

Groundhog day?