Okay, so maybe the thread title has an obvious answer.... but this is actually the problem I am having, and am yet to succeed in finding the cause.
My HTPC build is tucked away under the stand for my bedroom LCD, so it's never really been out in the open since first being finished a couple weeks ago. I had felt a tingle when brushing up against it to plug peripherals in before, but assumed it was just the hair on my arm getting snagged in a seam of the case. Well, last night I was going to turn the fan up slightly using the PCI control panel for the first time, and I got a legitimate shock when I first touched the steel thumbscrew for the side panel.
I immediately shut down the rig and have been attempting to figure out where on earth the case could be contacting a power source. There are no bare wires I can find within the PSU bundle. There are unused molex connectors that were tucked away and perhaps could have touched something, but would have been unlikely to have had their tips form real contact with the case.
I don't even know if there would be enough power flowing to an empty molex connector to deliver what was a fairly hard shock to a human. Even moreso, I don't understand how I have been running this thing for weeks now without a single issue while apparently having every metal surface of the build electrified. Has anyone else got any ideas before I simply put things back together and see if I get tagged again?
I'm thinking its a bad power supply, try another power supply! could be something in side that power unit, that's not just right.
I am thinking that is entirely possible, as it's being run with a cheap Antec Basiq 350w I had sitting around. Newegg is taking forever to deliver my Black Friday Corsair CX430 unit. However, it is worth noting that I was concerned with the power supply to begin with, so I monitored the voltage with OCCT for a short CPU test to begin with, and there was nothing unusual in terms of power spikes....
The volt meter is a very good idea I should have thought of myself. I must be tired from being up half the night putting around with this thing.
Definitely was not static discharge. Did you ever touch a 9v battery to your tongue as a kid and get the zap of the current? Well, touching the case screw was the same feeling in my hand, except about five times as strong. I even felt a slight zap when I brushed the top of the case, which has a black coating that I imagine is very low in conductivity.
I just tested it using every ground I could find nearby, including myself (don't try that one at home kids). The peak AC voltage level is 56v, as can be reached by touching any part of the back panel of the case, the PSU exterior, or case screws.
The really interesting part is that those readings are with the PSU switch turned on, but the PSU/computer actually turned off. So, that's a live wire providing the power, not a molex adapter as I had suspected.
Well, I've decided that I am concerned enough about the voltage levels and not being able to detect the source of the issue, my only real option is to just replace both the PSU (already was anyway) and case (it's a cheap Elite 343- thinking maybe it's causing the point of contact).
It would be nice to know what happened, but I guess not vital. Now the only question is if I buy a better case for my main system and put the HTPC into a Thermaltake Commander-Snow, or if I just buy a new micro tower. I'm hoping to drop less than $75 given all the other projects that I have going.
If you breadboard the entire thing (put it on cardboard or something non-conductive), can you detect a voltage by touching the PSU exterior and a ground? Heck, if you even plug in the power supply by itself (nothing connected to it), can you detect anything?
Wow. I just disassembled two entire builds to figure out that NEITHER of them was actually the source of the issue. I did, however, pinpoint what was going on.
I took out the Antec Basiq 350w and it was giving off high readings without anything connected to it. Next I made random move and took out the OCZ ZT 550w supply that I knew was fine from my other rig. It gave off even higher readings. This started a "WTF" all-night marathon of testing possible scernerios. Just solved it:
My home was originally built in the late 1950's, but the office space was a new addition built onto the back around ten years ago. I wasn't aware that 1) the outlets in the room were not grounded at the time of construction, and 2) the outlets were installed with REVERSE POLARITY.
Now, the first issue is my fault for not considering, since parts of the house have still not been updated to have grounded 3-plug outlets. However, flipping the positive and negative sides of the receptical is just unimaginable carelessness. So, while an ungrounded computer in itself would not have caused the issue, the lack of a ground wire left the discharge from the incorrect power flow with no place to discharge. I just confirmed this by taking a 3-pin to 2-pin adapter and using it to plug the PC cable upside down at the wall. Not a volt of power on the meter. Now I just need to figure out how to fix an entire room that needs grounding and new outlets....