PC won't boot with network cable connected

[LosOjos]

This is a really strange one, wondering if anyone has seen it before.

 

I came in to work yesterday to a dead PC. Wouldn't boot at all, didn't even try to POST. Opened up and started diagnosing and found that it booted again when I removed the network card - great I thought, problem solved, just use the on board network port.

 

This morning, same problem. Straight away I went for the network cable and it booted straight away. Plugged the cable back in once it had booted and have had no problems with it.

 

I can't get my head around how the mere presence of a network cable is preventing the PC POSTing, a bad PCI card made sense but the fact that on board LAN is acting the same now has me stumped. Any ideas?

Edited September 10, 2013 by LosOjos

[hallb15]

That is odd.

 

What about if you boot it up without the ethernet cable connected, can you then plug the cable back in and connect to the network as usual?

Sorry just re-read your post and see you have answered that!

 

Have you tried a different ethernet cable? different port on the switch/router?

Are there any other new devices on the network?

[LosOjos]

Have you tried a different ethernet cable? different port on the switch/router?

Are there any other new devices on the network?

 

Different cable/port made no difference, so far as I know the switch hasn't been changed...

Could be new devices on the network, I believe one of the IT suites in this block has just been re-fitted.

 

To add to the strangeness - it's working fine today! I'm getting a new PC in the next couple of weeks, hopefully this one will hold out til then!

[latenitetech]

Old post, but I just had a similar problem, so thought I'd share my finding in case someone else stumbles on this post. In my case, same symptoms - was a working (older) PC, then one day it wouldn't power on. By somewhat random troubleshooting, I figured out that if the ethernet cable was unplugged, it would power up and appear to work fine, but with ethernet cable attached, it wouldn't power up. Googling the symptom got me to this original post (and several others) but none of them had a solution.

 

I finally figured out it was the Standby 5v output of the PSU that was bad. I put a meter on that pin of the ATX PSU connector (typically the purple wire), and it read 5v as expected, but when I hit the power switch, it dipped to ~1v. (I'd also get a blink on the keyboard lights, but the PSU wouldn't "start.") Then I unplugged the ethernet cable and repeated the test. The standby lead dipped to ~4v, but the PC started. Apparently the NIC card draws enough current from the standby 5v line to pull that rail down only when the ethernet cable is connected. The NIC card using standby juice makes sense as it needs to be powered for WOL to function.

 

A related observation is that in this faulty condition of the PSU, the sleep function of the PC stopped working as well. The PC would appear to sleep properly, but when trying to awaken it, the PC would always cold-boot, which is akin to an "unexpected shutdown"/crash. This also makes sense as the PC needs standby 5v to keep the memory alive during sleep. On this motherboard, there is a small LED next to the memory sticks that indicated the memory is powered. That light was going out during sleep, indicating it wasn't being kept alive.

 

A bit more troubleshooting, I attached my bench supply (set to 5v) to the purple lead of the PSU as a "helper," and the PC worked properly. Started right up, sleep function worked (and memory LED stayed lit), awoken by WOL, keyboard, etc. Clearly I found the problem!

 

I pulled the cover off the PSU expecting to find the typical bulging/leaking capacitor, but alas, everything looked normal, so no quick fix there. Fortunately I had a spare PSU that I was able to swap into this PC and it's working fine again. The bad PSU is back in my [way too big] "to be repaired" pile, which may or may not ever get my attention again.

 

Here's hoping someone else stumbles upon this posting and finds it useful to this rather perplexing problem.

[mrhcarlin]

stuff

 

Small bump just to say great troubleshooting. We have a lot of older (circa 2006-2009) PCs at my school and some are naturally more worse for wear due to their usage or location, and I've seen this issue exactly as you described a few times. I never had the time to get stuck in and troubleshoot it myself, so I've bookmarked your post for future reference.

[Roxasaurus]

Old post, but I just had a similar problem, so thought I'd share my finding in case someone else stumbles on this post. In my case, same symptoms - was a working (older) PC, then one day it wouldn't power on. By somewhat random troubleshooting, I figured out that if the ethernet cable was unplugged, it would power up and appear to work fine, but with ethernet cable attached, it wouldn't power up. Googling the symptom got me to this original post (and several others) but none of them had a solution.

 

I finally figured out it was the Standby 5v output of the PSU that was bad. I put a meter on that pin of the ATX PSU connector (typically the purple wire), and it read 5v as expected, but when I hit the power switch, it dipped to ~1v. (I'd also get a blink on the keyboard lights, but the PSU wouldn't "start.") Then I unplugged the ethernet cable and repeated the test. The standby lead dipped to ~4v, but the PC started. Apparently the NIC card draws enough current from the standby 5v line to pull that rail down only when the ethernet cable is connected. The NIC card using standby juice makes sense as it needs to be powered for WOL to function.

 

A related observation is that in this faulty condition of the PSU, the sleep function of the PC stopped working as well. The PC would appear to sleep properly, but when trying to awaken it, the PC would always cold-boot, which is akin to an "unexpected shutdown"/crash. This also makes sense as the PC needs standby 5v to keep the memory alive during sleep. On this motherboard, there is a small LED next to the memory sticks that indicated the memory is powered. That light was going out during sleep, indicating it wasn't being kept alive.

 

A bit more troubleshooting, I attached my bench supply (set to 5v) to the purple lead of the PSU as a "helper," and the PC worked properly. Started right up, sleep function worked (and memory LED stayed lit), awoken by WOL, keyboard, etc. Clearly I found the problem!

 

I pulled the cover off the PSU expecting to find the typical bulging/leaking capacitor, but alas, everything looked normal, so no quick fix there. Fortunately I had a spare PSU that I was able to swap into this PC and it's working fine again. The bad PSU is back in my [way too big] "to be repaired" pile, which may or may not ever get my attention again.

 

Here's hoping someone else stumbles upon this posting and finds it useful to this rather perplexing problem.

 

 

Just want to say THANK YOU! I was fiddling with this machine for a few hours checking every BIOS setting I could, you saved me a ton of troubleshooting time with my fluke meter. You're a genius!

 

Thanks so much man,

 

Greg

[MrCoffee]

First can I thank latenitetech, this is spectacular faultfinding.

 

Workaround:

We have managed to get around this problem by plugging the NIC into a seperate POE switch. (We had a couple of unused 8 port netgears around.)

Some people might get around this with a POE injector if it's a one off.

 

This all seemed to happen to our 10 year old small form factor machines that are used as spares. (PSU model TFX-250SAV).

 

Thanks again.