Question about proper grounding

[Trellot]

Hello everyone,

I ran across a strange situation the other day while helping a friend with his PC; the unit was experiencing sporadic lock-ups. Part of the problem was that it regulary would not boot without freezing during POST (no BSODs). However, if he kept at it, the PC would boot-to-desktop after maybe 4 or 5 attempts. When it did boot, all would be well for an hour or so and then another lock-up (and again, no BSODs; no error messages, etc). When I took a look at it I started by running my good friend "Malwarebytes-in-Safe-Mode" and it dispensed with something like 56 various viral nasties. That took care of the booting problem and every reboot afterward went through straight to the desktop. Things were looking up. Fast forward a couple of days and I'm speaking to my friend about his PC and he told me that it locked up again two hours after I had left. At this point, I began questioning him further about his setup and he mentioned something about proper grounding. More questions from me and I finally discovered that the wall outlet he is using does not have a slot for the ground prong! Apparently the home was built a loooooong time ago! My buddy uses a surge protector and I asked him how in the world did he plug it in!!? He replied, rather sheepishly, that he thought he remembered having cut the ground prong off ...lol!

Anyhooo, my question is, since I have never heard of something like this before...how might this have affected his power supply and, further, could it be the cause of his computer lock-ups? Also, might other components be damaged as well? I'm no electrical engineer, I just know that you don't cut the ground prong off a PC surger protector's plug-end just to make it fit in the wall outlet, hehehehe!

Thanks for any comments,

Trellot

[TwistedKestrel]

I've run into this myself - the apartment I moved into has only two-conductor wiring, but there are grounded receptacles installed everywhere. The ground is not connected on most of them.

Grounding on a PC serves two purposes. The first and obvious one is safety - the regular DC voltages inside a computer are not very dangerous, but the 120/240 AC inside the power supply is. If there is an AC short inside your power supply, it will normally be shunted to ground. With no ground, the whole chassis will become energized - not a good situation. This kind of failure is rare, but under the right circumstances could kill you. It probably won't kill you, as someone will immediately point it, but it's possible.

The second purpose, less obvious but having a more immediate effect, is to turn the chassis into a grounded RFI/EMI shield. This has an effect on all kinds of interference, from your monitor, to the mains, to the very components inside your computer. It is quite possible for a computer to become less stable due to a lack of ground.

Incidentally, cutting the ground off a surge protector renders it 100% completely useless. Power surges get shunted to ground; a lot of surge protectors have a warning light if they don't detect a ground connection.

[Captain Ned]

Incidentally, cutting the ground off a surge protector renders it 100% completely useless. Power surges get shunted to ground; a lot of surge protectors have a warning light if they don't detect a ground connection.

Dig deep into any current US-standard breaker panel and you'll find a bus bar that ties the ground terminal block to the neutral terminal block.

[StefanVonS]

Incidentally, cutting the ground off a surge protector renders it 100% completely useless. Power surges get shunted to ground; a lot of surge protectors have a warning light if they don't detect a ground connection.

Dig deep into any current US-standard breaker panel and you'll find a bus bar that ties the ground terminal block to the neutral terminal block.

Yup, still grounded however. If it wasn't tied to the neutral, fault current might not be enough to trip the breaker. This doesn't negate the need for a case ground.

[westom]

Dig deep into any current US-standard breaker panel and you'll find a bus bar that ties the ground terminal block to the neutral terminal block.

Dig deeper into any computer to discover those two 'grounded' wires connect to things electrically different. A neutral and safety ground wire are electrically different at the receptacle end. All wires are electronic devices - not pure conductors. Electricity is different at both ends of any wire. Sometimes that difference it trivial. Sometimes significant.

Accurately noted is the primary and first reason for safety ground. Human safety. Most EMI/RFI features are over exaggerated. In reality, a ground wire can become an antenna to transmit EMI/RFI noise. Make the problem worse. Because that wire is too long for EMI/RFI soutions.

More important is to have all interconnected components (monitor, printer, computer, etc) connected to a common ground. That (floating) ground need not be connected to anything else. That ground, by itself, does not make noise irrelevant. But is part of a redundant system so that failures (ie crashes due to signaling noise) do not occur between interconnected components.

It does not solve EMI/RFI for the same electrical reasons it also does not do surge protection. Best surge protection can exist even when no receptacles inside a house have safety ground. Surge protection is always about earth ground. Safety and earth grounds must be interconnected (per code) for human safety. But both are electrically different. So different as to even be defined separately in electrical codes.

A missing safety ground does not cause problems for a properly designed computer. If a missing ground does cause crashes, then other problems also exist inside the machine. Multiple anomalies (not just the missing safety ground) must exist to create that failure.

Missing safety ground can be solved by installing a GFCI as defined by code. Interconnected equipment should all be grounded to a common ground such as a floating ground inside that power strip. Protectors must be earth grounded. An effective protector has a dedicate wire to create a low impedance (ie ‘less than 10 feet’) connection to single point earth ground. Floating ground, safety ground, and earth ground are all electrically different. Perform different electrical functions.

[just brew it!]

I could imagine the lack of ground causing a problem during dry weather, when static electricity is more prevalent. Without the ground, static discharge can go through the PC's circuitry instead of being drained off by the ground wire.

Yes, the lack of a ground also likely renders any surge suppressor he may have useless. But consumer surge protectors are often close to useless anyway, so this may not be a huge loss.

Even in the absence of system instability, you really want the ground there for safety reasons. Otherwise, a catastrophic component failure or other anomaly (e.g. something gets wet) could result in a potentially lethal situation.

[DLHM]

I've never seen a floating ground cause a PC to lock up, but I suppose anything is possible. Like the previous poster stated, The ground wire connects to the neutral bar back in the breaker panel, and the lockups from a floating ground would most likely be caused by some other problem or problems with the computer build. I'm not suggesting anything, do you own homework, but tying the green/copper wire to the neutral would remove the floating ground from the situation, but I would bet that the computer would still lock up from other reasons. The best way to ground that outlet it to run a singal conductor to a grounding rod under/outside the house(if possible) or running it to the neutral bar in the panel. On old house that has an ungraded septic system with an alarm installed you will notice that during install of the septic system they will have usually grounded at least one circuit in the house to hook the septic alarm to. If you are familer with electrical work(and only if you are familer) you can move the recepticle to the grounded circuit.

[westom]

... tying the green/copper wire to the neutral would remove the floating ground from the situation, but I would bet that the computer would still lock up from other reasons. The best way to ground that outlet it to run a singal conductor to a grounding rod under/outside the house(if possible) or running it to the neutral bar in the panel.

Both suggestions are major human safety violations. Safety ground prong in a wall receptacle must not be connected to an earth ground rod. Neutral and safety ground wires must always remain separate except where they meet inside the breaker box. Because electricity on both wires at the receptacle end is different. Because code says safety ground and neutral must not be connected together at the receptacle. And because house insurance also has a blunt expression if they find out those suggestions were implemented.

A missing safety ground only causes computer problems when a computer itself is defective. Fix a problem; not its symptoms.

[Captain Ned]

A neutral and safety ground wire are electrically different at the receptacle end.

Maybe so, but the instant that plug goes into a wall receptacle, ground and neutral are the same thing. Any active ground protection has to be in the device itself ahead of the breaker panel.

Because code says safety ground and neutral must not be connected together at the receptacle.

If they're connected in the breaker panel they're effectively connected at the receptacle.

Safety ground prong in a wall receptacle must not be connected to an earth ground rod.

Every code-compliant US breaker panel has a copper ground rod pounded into the earth and connected to the ground terminal strip in the panel. How is the ground prong in the receptacle (or the neutral prong, for that matter) not connected to an earthed ground rod?

[DLHM]

Safety ground prong in a wall receptacle must not be connected to an earth ground rod.

This is quite common.

Neutral and safety ground wires must always remain separate except where they meet inside the breaker box.

This is never the correct way to do it, and it could cause a problem, but it's not banned specifcally by code, at least not that I can find.
I happen to have the NEC book on my desk.. perhaps I missed it.

If you want some sort of protection on that line you can install a GFIC outlet, it won't fix the floating ground, but will protect the user.

I warned to do you own home work, of course having a qualified, licensed electrician is allways the best course.

edited: For clarity and for more information..

[Trellot]

Thanks everyone for your comments,

Here is where I'm at with my friend's computer: I picked it up and brought it to my home and it won't boot into the OS. It freezes after the POST just after getting to the Windows screen. It freezes trying to boot into Safe Mode also. Here's what I've done so far:

- I've tested the power supply with a little power supply tester and it doesn't report anythying unusual.
- Plugged in a known good power supply I have and still getting the freeze ups.
- Ran Samsung specific tests on the hard drive (since that's the brand); also ran Spin Rite (after backing up all his data of course by plugging the drive into my PC) - no issues.
- Ran Chkdsk with no issues, also while attached in my PC
- Pulled all 4 memory modules and test booted with each one individually and I get same freeze up upon Windows load screen
- System has integrated video with a DSUB and DVI port; I've tried the monitor on either port thinking maybe a video issue but no difference. (I should mention that when I'm plugged into the integrated video ports after the system goes to the windows load screen everything turns black but my monitor still shows a blue light as if there is some video connection still).
- Tried a known good video card in the system with same freeze up results (however, the difference is that when the system freezes up on the Windows load screen the monitor doesn't go black, but instead shows the frozen Windows logo).
- Scanned for further viruses but all is clean

One weird thing the HP computer does when I turn on the power is that the power supply resets 3 or 4 times turning off and on before finally booting up; almost like I've pulled CMOS or something and it's doing a couple resets.

At any rate, at this point I'm wondering if the OS is corrupted or something? Any advice would be very much appreciated.

Thanks,

Trellot

[videobits]

This is where using a Ubuntu or simlar live CD is a nice thing. If the computer will run that OS, then you can start to assume the Windows install is bad rather than bad hardware.

[ludi]

Maybe so, but the instant that plug goes into a wall receptacle, ground and neutral are the same thing.

Actually, since the Neutral has return current flowing on it, and since the sum path from the outlet to the Neutral ground at the service panel has at least an ohm or two of impedance, this method puts a floating potential on the case. It will usually be in the range of a few volts, which is generally harmless, but it's a bad practice, especially since there may be other outlets on the circuit adding to the neutral current load downstream, and thus raising the net voltage rise at the outlet where the computer is plugged in.

A more serious risk is that old outlets are sometimes not polarized, and even when they are, three-wire plugs usually rely on the ground prong to enforce polarity. Cut that off instead of using a grounding adapter, as many users tend to do, and it can be inserted either way. Also, houses which are old enough to lack earth grounds often tend to have a lot of novice wiring in them as well, so even a polarized outlet may be wired in reverse polarity.

[ludi]

One weird thing the HP computer does when I turn on the power is that the power supply resets 3 or 4 times turning off and on before finally booting up; almost like I've pulled CMOS or something and it's doing a couple resets.

That sounds like a power supply problem, a motherboard problem, or an intermittent short somewhere in the system. One thing to check for would be broken or damaged pins inside a USB connector. The USB controller is supposed to self-protect in that situation, but they don't always handle it gracefully.

[Trellot]

One weird thing the HP computer does when I turn on the power is that the power supply resets 3 or 4 times turning off and on before finally booting up; almost like I've pulled CMOS or something and it's doing a couple resets.

That sounds like a power supply problem, a motherboard problem, or an intermittent short somewhere in the system. One thing to check for would be broken or damaged pins inside a USB connector. The USB controller is supposed to self-protect in that situation, but they don't always handle it gracefully.

That was my thought, but I tried a good PSU I have and I was getting the same 3 to 4 resets and then, of course, the impending freeze up. I also was thinking about an itermittent short, so I will pull the mobo and run it outside the case also. Another thought is that these resets happen every time I shut down and restart, so It's not that intermittent...3 or 4 resets, then computer starts to boot, then it freezes...every time.

Thanks,

Trellot

[westom]

That was my thought, but I tried a good PSU I have and I was getting the same 3 to 4 resets and then, of course, the impending freeze up.

With a multimeter and one minute of labor, my next post will say without speculation what is and is not working. Especially useless is a power supply tester. Defective supplies can be reported good by that PSU tester. You don't need any answers that say 'maybe'. You need answers that say 'yes' or 'no' - without any speculation.

Best solution starts by measuring without disconnecting or swapping anything.

The computer is also an HP. All computer manufacturers have comprehensive hardware diagnostics so that problems can be seen before changing anything. But only the better computer manufacturers provide them for free. HP is one of those companies. Always execute the diagnostics to see a defect before even disconnecting one wire. Those diagnostics are on the hard drive, on a provided CD-rom, and on the web site.

Do you know about the power controller? A power controller determines if a system can even power on. Its many functions are also reported by the one minute labor with a multimeter. So many automatically blame a PSU without even knowing about the power controller. Even the OS has no control over that controller. Many power controller inputs determine if a power supply or a CPU is even permitted to operate. Only numbers from a meter can even see the intermittent even when the intermittent is not causing failures.

Many suggestions so that the better informed need not stay silent. Your actions at home mean those with better knowledge can provide no useful suggestions or define the actual failure.

[Trellot]

Many suggestions so that the better informed need not stay silent. Your actions at home mean those with better knowledge can provide no useful suggestions or define the actual failure.

While I respect your opinion and thank you for the time you spent commenting on this thread, I do not appreciate your criticism here at the end of your post. First, I don't own a multi-meter because I can't afford one (expensive brand or cheap) so I am getting by with the troubleshooting techniques that I have learned, that I have known, and that I have used for years to my success. Perhaps this is the one case I've run into thus far where ONLY a multi-meter can tell me what is going on...if that is the case, so be it. But, on the whole, I've not had many issues determining the right faulty part. Again, thank you for your comments but a smidgeon of humility might serve you better.

~ Trellot

[P5-133XL]

If you have booting problems during POST -- The problem is NOT anything involved in Windows or malware. Windows does not get involved in the booting process till after the POST. POST is an internal testing process of the hardware to see if there is a hardware problem. If your computer locks up during POST you have hardware issues that prevent bootup.

If you are lucky your MB has an internal diagnostic LED. Perhaps there are some speaker pins on your motherboard that you can use for diagnosis (beep speakers are cheap) for POST will issue beeps to the speaker when it finds a problem. There are also diagnostic boards one can buy that you can plug into a free slot but those are expensive. A good physical examination of your motherboard may identify bulging capacitors; a burnt plug or a burnt smell may tell you something; even broken traces. Short of that, you end up swapping out components till you find the problem(s)

[Trellot]

If your computer locks up during POST you have hardware issues that prevent bootup.

This isn't always the case though is it? What about a boot sector virus or some such anomoly? Or, for me, I've experienced lockups because the drive requires IDE mode and is set to AHCI in BIOS...stuff like that. Some strange BIOS setting can cause weird results.

~ Trellot

[westom]

Perhaps this is the one case I've run into thus far where ONLY a multi-meter can tell me what is going on...if that is the case, so be it. But, on the whole, I've not had many issues determining the right faulty part. Again, thank you for your comments but a smidgeon of humility might serve you better.

Humility is accused when one fears a 100% technical post. One full minute of labor with a meter means a next post says what is and is not. Without speculation or so much emotion. Unfortunately even jobs are lost when so many fear to learn a concept widely taught in Japan. Work smarter; not harder.

Only two options exist. 1) Keep fixing good parts until something works. Also called shotgunning. Or 2) find a problem before fixing anything. In your case, one minute with a meter answers everything.

I do appreciate what you are really saying. You prefer shotgunning to learning better and faster techniques. So this post is not for your benefit. This post is for the 30+ other readers who read this stuff to learn.

All should appreciate what PS-133XL accurately stated (and because a virus obviously was not a suspect):

The problem is NOT anything involved in Windows or malware. Windows does not get involved in the booting process till after the POST. POST is an internal testing process of the hardware to see if there is a hardware problem. If your computer locks up during POST you have hardware issues that prevent bootup.

That suspect sometimes does not even let POST execute. Causes other failures that act like a boot sector virus. And creates those other symptoms. POST and other booting problems are routine when one of many power system components is completely defective. Therefore act intermittent. A tool that can ‘see’ that defect is a multimeter. Even good PSU can be intermittent in an otherwise good computer – another reason why shotgunning so often causes confusion.

Only numbers define or exonerate all suspects without doubt. Only numbers can obtain useful assistance from those who better know this stuff. We learn this stuff from posters who have been doing this stuff for generations. Who even know why a multimeter is so useful. PS-133XL accurately defined a suspect. Quickly eliminated other suspects that exist only due to shotgunning. Unfortunately many will do anything to avoid learning what is only feared. This case requires only a tool so simple as to be marketed to 13 year olds in K-mart. Unfortunately many would rather work harder – use shotgunning. Because a meter for $5 in Harbor Freight is scary. Humility is irrelevant. Shotgunning is why this problem remains unsolved for so long. Provided for the many others who can learn from that problem and these posts. Work smarter; not harder.

Grounding was also blamed due to specualtion and no numbers. By not first learning how computers really work. A myth that also would have been eliminated quickly by learning/using the same one minute of labor using a meter.

[Ryu Connor]

Be nice.

[P5-133XL]

If your computer locks up during POST you have hardware issues that prevent bootup.

This isn't always the case though is it? What about a boot sector virus or some such anomoly? Or, for me, I've experienced lockups because the drive requires IDE mode and is set to AHCI in BIOS...stuff like that. Some strange BIOS setting can cause weird results.

~ Trellot

You're wrong in one respect. The POST (Power On Self Test) finishes before it even checks for a boot sector's existence. A boot sector virus will only affect your machine after POST when the boot process transfers control from the BIOS to the software/program stored in boot sector. The actual POST software is stored in your BIOS and is not normally subject to any external software influences like malware, an application, drivers, or even an operating system such as Linux or Windows.

The only way software could even affect the POST would be if the malware rewrote your system BIOS (i.e. a BIOS update) or directly modified your CMOS settings and I have yet to hear of one that has gone to that level. If there were such a malware, the only real solution would to have pre-purchased a dual-BIOS motherboard and replace the modified BIOS with the spare original and/or reset CMOS if that had been modified. For some few motherboards, the BIOS is socketed so you could theoretically order a replacement from the manufacturer or burn a new one externally. But again, I've never even heard of malware targeting the BIOS. Those dual-bios motherboards and socketed BIOS chips were designed to fix failed BIOS updates rather than deal with malware but they would work.

A failed BIOS is not a possibility for your symptoms because it would not be intermittent.

Now a conflict between a CMOS BIOS setting and the actual hardware can cause a POST failure like excessively over-clocking your processor in the CMOS setup (a much better example than the one you gave, because it will actually be caught by the POST while your IDE vs. AHCI issue gets caught by the Windows HAL (Hardware Abstraction Layer)). The BIOS determines how the basic hardware operates. If you give the BIOS an incorrect or invalid setting then it is possible that the BIOS will try to operate the hardware incorrectly thereby causing the POST to fail. I point out that this really isn't hardware, firmware, or a software flaw and definitely not a Windows or Malware problem but rather just pure user-error. It just happens this user error can be caught at the POST level.

The simple solution to a CMOS setting conflict with the hardware is to use the reset CMOS jumper on your motherboard, or use the "reset all settings to" default/optimized option from inside CMOS setup if you have access. A third way would be to remove the power cable, and remove the motherboard's battery and wait an hour or two before replacing.

There is one other error that POST can detect that isn't an actual hardware failure and that is if you physically hold down a key on your keyboard while it is POSTing -- POST will detect that as a stuck key and fail your keyboard. Again, another user-error.

I really can't think of a POST error other than those that are not true hardware failures.

[PrecambrianRabbit]

One full minute of labor with a meter

You'd be more helpful if you suggested what should be done with said multimeter to diagnose this problem.

[bthylafh]

One full minute of labor with a meter

You'd be more helpful if you suggested what should be done with said multimeter to diagnose this problem.

I can offer some suggestions!

[ludi]

While I respect your opinion and thank you for the time you spent commenting on this thread, I do not appreciate your criticism here at the end of your post. First, I don't own a multi-meter because I can't afford one (expensive brand or cheap) so I am getting by with the troubleshooting techniques that I have learned, that I have known, and that I have used for years to my success. Perhaps this is the one case I've run into thus far where ONLY a multi-meter can tell me what is going on...if that is the case, so be it. But, on the whole, I've not had many issues determining the right faulty part. Again, thank you for your comments but a smidgeon of humility might serve you better.

You're evidently new to Westom, so here's the Cliff's Notes: Westom is the Internet's self-appointed Sir Lancelot, bravely rescuing all threads related to surge protectors and attendant power system hardware by slashing his sword at anything that moves. Sometimes he hits the target with a precise blow, sometimes he just makes a mess of things. As long as you don't mind culling through his mill wash to find the gold flakes, you can get something useful; otherwise, suggest just letting it pass. But you probably won't improve his attitude in any case: he's got practice

[westom]

You'd be more helpful if you suggested what should be done with said multimeter to diagnose this problem.

Set a multimeter to 20 VDC. Attach the black probe to the chassis. Touch the red problem to a purple wire from the power supply where it connects to the motherboard. That will read about 5 volts even when the computer is off. Record all three digits.

Repeat same for the green and gray wires both before and as the power switch it pressed. Note those numbers and the behavior when the switch is pressed.

And finally measure any one red, orange, and yellow wires as the power switch is pressed. Note the behavior and final number. Then report all numbers here. You would not believe the amount of information contained in those few numbers. Too much to post here. But those numbers make possible definitive replies from the fewer who actually know this stuff. No more 'it could be that' or 'might be this' replies.

The previous post was already too long. Long posts make some people angry. So I often include this information only if someone asks.

Problems such as Trellot's often exist months or even a year before it starts causing failures. Meter numbers might have identified it that long ago.

[Coran Fixx]

Mr Hand frowns on your shenanigans.
Mr Hand requires you to test the power supply you already know is not the problem.
Mr Hand requires you to report for further instructions after school.

I suggest you catch some tasty waves instead.

[bthylafh]

What about Mr. Hands?

[bryanl]

re: "Dig deep into any current US-standard breaker panel and you'll find a bus bar that ties the ground terminal block to the neutral terminal block."

You shouldn't have to dig deep but ground and neutral should only be bonded at the service entrance where the meter is, not the distribution panels. This is to prevent circulating currents and voltage differentials between circuit branch chassis grounds.

This topic gets RV types going when they use portable gensets because their little 'power OK' indicators do not show a hot to ground voltage. Portable, plug-in gensets up to 5kW are allowed to have a 'floating' neutral by code. (more than 5kW makes the risk of induced currents start to get real). There have been some very interesting and enlightening 'discussions' of this topic on some RV forums.

As for the OP problem, my first guess would be a flaky or inadequate power supply and, after that, capacitor failure. I would try to make sure that all of the equipment was plugged into a common power strip and that the ground adapting only be done at the wall outlet. That will assure a common chassis ground for the equipment and provide a first level EMP protection.

As for the pre-sixties 2 wire outlets, use the adapters that provide a ground lead that is put under the outlet cover screw. It is never a good idea to castrate a modern 3 pin plug.

If you find that the computer problems seem to correlate with someone turning on the vacuum or the A/C or furnace coming on or some other motorized load, then surge protection and some additional grounding may be needed - see the amateur radio websites for how they ground their stations for ideas. Static discharges are also good to avoid so you should have some 'safe' piece of equipment you can touch first if needed, one that is plugged into the power strip serving as a common ground.

[Trellot]

Everyone,

Thanks again for all of your comments on this thread. Sorry I haven't posted this sooner, but I've been busy with work and all of that. I was able to discover and subsequently fix what was causing this computer freeze/no boot issue. It turns out there were three bad capacitors on the motherboard, one of which was bulging out on the bottom; was able to remove the old caps and solder on three new ones. Computer is back in business.

Issue Resolved,

Trellot