Random Power Off or Crashing (Typically after static shock)

[Throttle]

There are NO docking stations or port replicators involved.

Users have a battery securely installed and an AC adapter plugged in.

Users will be working and it just powers off or reboots right in front of their eyes. Other users come back to their desk to find their computer powered off.

I have powerline monitoring equipment testing AC input, and have never seen anything power-related.

The problem seems to go away when we switch from the ethernet port to wireless. We experience this in multiple locations on multiple brands of ethernet switches.

We're working with Lenovo, but they can't duplicate the problem on one of our machines. Out of about 350 machines, it happens to 30 or so. All of our machines are running the same standard load. We can take a machine that's never experienced the issue and give it to someone who has and they will have the same issue with a known good machine.

A user could be fine for the last 9 months and then all of a sudden they start having the issue for a few days and then it can go away. We have some users who have the issue every time we swap them back to ethernet from wireless. Might happen 2 or 3 times a day.

I suspect it's a power management issue and probably related to the system protecting itself from the battery. We recently had one user report that the Power Manager Gauge turned yellow right as the computer powered itself off. Yes, we replaced all of the batteries related to the recall already.

Does anyone else experience this issue?

[hoya]

go into device manager and disable all the power mgmt properties for both the wireless and wired connections via the Advanced Tab for each respective card. be sure to disable deep smart power down on the wired card as well. please let us know whether that works.

[Throttle]

Thanks for the response and thoughts.

All of those settings are already set either manually (advanced tab on NICs) or via registry (Deep Smart Power Down and Smart Power Down) when we build the machines. We also tested with all different versions of the BIOS.

Any other thoughts? Anyone else have these issues?

Thanks,

ajh

[hoya]

try running the problematic machines on a UPS w/ a good surge protector. switching power sources seems to be highly problematic on these T60's so I would rule that out.

what application(s) are the users running when this occurs? are they in a dry climate with any significant static? (you mentioned static in the title but not the body of your post). Which BIOS are they running?

I've had lots of freezes on several T60's but I did have one that would randomly shut off (only happened a few times). strangely enough it happened once when I closed Windows Media Player during playback of an mpeg file while running on battery power.

[jdhurst]

Throttle penned: "The problem seems to go away when we switch from the ethernet port to wireless. We experience this in multiple locations on multiple brands of ethernet switches."

I just reimaged an IBM desktop machine at a client on Saturday. Random failures were causing hard drive corruption. I even changed the computer and had the same problem on the different computer and different drive.

Saturday, I put my laptop down on the users desk, plugged into the nearby switch and could not get a steady connection. Now my T41 is as near to flawless as one can get, so I automatically assumed there was something wrong at the specific client workstation. After a bit of work, I was ready to condemn the switch (and had not yet gotten to the desktop which needed reimaging).

I started to pull out the switch and decided to check that the cable from the wall was plugged into the correct port. I then discovered the cable was exceedingly badly kinked (to the point of being broken, but not open).

I replaced the cable, got a solid connection on my T41, imaged the machine, checked today and all is well.

I now believe that the bad cable was causing bad data from the server that was in turn causing the machine to freeze and the then the data on the drive to corrupt. It seems somewhat unlikely, but I have no other cause at this point.

Unlikely this is your problem, but it is something else to think about in your circumstances.

... JD Hurst

[Throttle]

We've checked/replaced all of the ethernet cabling and tried a UPS and even swapped the AC adapters, batteries, etc. I have a powerline monitor on one system that keeps track of sags, surges, etc., and that has never gone outside seeing perfect power.

Applications vary, but users are typically creating an MS Word document at the time or the machine is just sitting idle and powers off. The user doesn't even need to be near the machine, however, in a lot of cases the user puts their hand on the keyboard or on the side near the screen and a they feel a shock when the machine powers off.

We are in a climate where there is a lot of static and right now the air is very dry.

ajh

[hoya]

I owned a Toshiba A75 which eventually had a full recall and a class action settlement due to static causing the entire system to freeze. hard to say... if I think of anything else I'll post here.

[claudeo]

It seems that the title of the original post contains a good hint: Static shock. Nothing should be expected to survive that uscathed. The next good hint is that the issue seems to be tied to users rather than machines. Give a good machine to one of the affected users and the problem seems to occur. The third hint seems to be the wired vs. wireless connection. The AC adapter itself is double-insulated, but the Ethernet connections share a (weak) common ground somewhere. IMHO, you need to either switch to static abating carpeting or get those users to use best practices in a high static electricity environment, e.g. touch a *good* ground *before* they touch the machine. And do it *every* single time they sit down at their computer. Now, if they walk to their desk with computer in hand, having built a nice static charge, and they plug it into the Ethernet cable or into the docking station, that is a disaster waiting to happen. Again, they should get into the habit of touching a good ground at the same time as they are touching a ground point of the laptop before plugging in the laptop. It looks like the T60 hinges are tied directly to the chassis of the machine, so that's a good touch point on the laptop. But finding a good touch point on the desk may be more difficult. Remember the AC adapter is probably double-insulated, so that does not make a good ground. Neither does any 2-wire device, like a desk lamp or pencil sharpener. You may have to run a green wire from the ground on an AC outlet to something that can be touched on the desk. Granted this is nasty. Getting rid of static building carpet may be cheaper than all this, what with the training, hacks, repairs and computer replacements.

[joester]

Look for anti-static pads to put under the chair. Add a grounding point on the desktop like a metal strip the worker can't avoid touching with bare skin. Ground it with a 1 MEG resistor, so the charge is bled off slowly. Straight wiring to gound will give the familiar doorknob jolt, and result in unhappy workers.
Is the network cable shielded? Add a ferrite loop somewhere near the computer. We have a "wild run" at work that acts up similarly when cold weather presses in like it has lately. Everyone is required to wear ground straps and faraday jackets, so further investigation found that unwanted signals were being transferred through the UNshielded network cable. This particular run is quite long. Plans are set to replace the cable, but adding ferrite loops seems to have done it for now.
http://www.internet-security-inc.com/ca ... eClamp.gif if you haven't a clue what ferrite I'm talking about. This link is for the picture only, not an advertisement or sales pitch. Get the right size for the cable, or loop the wire through the middle, around the outside and back through the middle again.

Joe