17

I'm afraid I have made a grave mistake. I was just finishing up my new PC build with some pretty pricey components when I had made a simple, yet gut-wrenching mistake. I was under the presumption that the PC was off, but alas; it was doing a reboot. I know I should've unplugged everything but I learned a great lesson that patience goes a very long way when dealing with hardware...

Anyway, while the PC was booting (no OS, just BIOS) I had reseated the RAM and inserted right as everything turned on (lights, fans etc.). I panicked and quickly unplugged the power a second after I had realized my mistake.

I'm a bit paranoid in nature so I took out the RAM, replugged it to see if it worked (as in, I had booted into the BIOS again to see if both of my ram sticks were detected in the same slots); it did indeed.

I then proceeded to sniff the entire computer to see if I had fried anything; it doesn't seem to appear so visually at a glance and I didn't see or smell any smoke. It just smells like electronics.

I thought the worst and thought I fried my new CPU that I didn't even get to use, I took out the cooler, rubbed off the thermal paste and inspected the top and the pins. I didn't smell or see anything out of the ordinary.

I inspected the GPU next, but that seemed fine too. Booting the PC seems normal, but my pessimistic mind keeps thinking I did irreversible invisible damage and the only reason I'm not seeing it now is because the hardware is new (except for the GPU).

My (relevant) specs are:

  • Ryzen 9 5900X (new)
  • MSI GTX 1070 Ti (5 years old, but only used for a little bit after 5 years)
  • 600W EVGA PSU (5 years old but completely unused and practically new usage wise but it is technically old)
  • (8x2)GB SP RAM (new)
  • MSI B550 Tomahawk Motherboard (new)

How do I know if something messed up? Would I have to wait until I do a real OS install before I can really know?


UPDATE: 10/6/2022

Ran (memtest86) on my system (twice just for reassurance), it came out with a pass on all the tests and my system is working as expected :). Again many thanks. I figured I would put this note here to let others know in the future what was the outcome and figure how it can apply to their situation(s).

Also ran the PassMark PerformanceTest that showed the CPU running on par with the average score, actually scoring higher; but that's probably because I went overkill with my cooling setup and I live in New England so it's nice and chilly.

UPDATE: 10/11/2022

Again, future paranoid viewers: The only real issue I saw before and after was that one of my HDDs starting throwing errors that it hadn't even done before. It was already old but after this entire hassle S.M.A.R.T had noted it was writing more read and write errors. Most likely bad timing but it may be possible that could've taken some sort of damaged. This is a heavily used HDD from 2014 however. And it was plugged into the SATA port when this had occured of course. All 3 other HDDs are completely fine.

8
  • 20
    Regardless of whether you made such a mistake, an assembly of a new system should be subjected to a burn-in for at least overnight or 24 hours. Running Memtest86+ would help reassure you.
    – sawdust
    Sep 30 at 21:18
  • 1
    Prime95 is a good stress-test for flakiness in the memory bus interface (by transferring a lot of data while running the CPU hot), vs. memtest86+ being good for catching problems inside the RAM with specific cells. Oct 1 at 10:40
  • 2
    I once did the same thing on a trash computer that was going to be thrown away (it was DDR2 ram) just to see what would happen. The ram sort of plugged in at an angle which shorted some power lines and cused part of the socket to burn red. Safe to say the socket did not work after that. Oct 1 at 11:12
  • 3
    @sawdust: Why do you think any kind of “burn in” is required? Just wastes a lot of electricity.
    – Michael
    Oct 2 at 15:55
  • 2
    @Michael -- The purpose of a burn-in is to expose early failures before you spend hours installing software and copying over your files. This is not just what "I think", but standard practice for some equipment. It's a step typically skipped by manufacturers of consumer-grade stuff because of cost. It's up to the end-user to deal with out-of-the-box and early failures. Confidence in new equipment is IMO worth a few Euros for electricity.
    – sawdust
    Oct 2 at 22:36

6 Answers 6

65

I'd have just taken a deep breath… and booted it again to see if it worked. You kind of skipped the simplest test & ran for the full paranoia.

If something's fried you find out right away. If it's weakened… you find out eventually. No test will indicate a shortened life that wasn't immediately terminal.

2
  • 3
    There's just no way to know what kind of damage, if any, would happen. There's just no way to capture exactly which pins you powered in what order, and what its effects are. Nobody's going to waste time testing RAM being inserted while powered on, so nobody's going to know.
    – Nelson
    Oct 1 at 12:16
  • 2
    He could have blown a memory range. I'd run a memory check tool; memtestx86's first test of each address, own address really quickly finds blowouts.
    – Joshua
    Oct 3 at 23:47
20

You're fine. And lucky.

If something had gotten damaged, it would not get past the BIOS. If things would've broken, it would either be the RAM chip itself or the motherboard.

Both will cause error beeps or no proper booting past the BIOS screen, so the fact that you already got there means nothing serious happened.

5
  • 2
    Modern desktop processors, including OP's Ryzen 5th generation, have the RAM controller(s) inside the CPU. So if something did go wrong, it's highly likely that the damage would have been there, inside the CPU die.
    – Ben Voigt
    Sep 30 at 21:24
  • 4
    I'm curious, what's exactly the issue with inserting RAM while it's booting?
    – justhalf
    Oct 1 at 7:56
  • 25
    @justhalf, connectors usually have to be designed to be hot-pluggable. Often this includes making sure that ground and power supply connect first, so they're in a defined state before data lines connect. Otherwise sensitive electronics might be damaged by unexpected voltages on data lines. There will probably be other considerations for specific connectors as well. But RAM simply isn't designed for this.
    – Kritzefitz
    Oct 1 at 8:22
  • 3
    @Kritzefitz: Consumer-grade RAM modules aren't designed for this. But there exist industrial backplane connectors for hot-plugging RAM (which is naturally on matching hot-pluggable RAM cards). Google "Hot-plug Memory Expansion Board" and you'll find plenty.
    – Ben Voigt
    Oct 3 at 18:19
  • (luck) well, it's nice their PC will still work Oct 19 at 14:22
9

There isn't any guarantee that your RAM is fine at this point simply because it's booting.

Burn a CD or USB of memtest, boot your computer off of it and run it overnight. If there aren't any errors, you're probably fine (for the moment). Long-term damage that isn't visible in the first instance can sometimes occur with this scenario. If you check it again in six months and it's still fine, you're good to go.

Besides that, I hope you grounded yourself by keeping one hand on the metal part of the case while touching the RAM, or that you used an anti-static wristband attached to the metal part of the case.

7
  • 2
    Why do you think running memtest over night is necessary? One full run should be plenty enough.
    – Michael
    Oct 2 at 15:58
  • 5
    If the RAM has been damaged, the damage can be quite subtle. A particular bit may work most of the time, but fail occasionally. Oct 2 at 20:43
  • 1
    It's really hard to actually break RAM with static... Worrying about grounding really isn't worth the headspace anymore Oct 3 at 12:27
  • 2
    Electrical components have only gotten smaller and therefore more prone to damage from electrostatic discharge. Please do worry about grounding. Oct 3 at 12:32
  • 1
    I've had mixed results with memtest as technology advances. If it causes issues, there's also memtest.org which is a different thing now.
    – Criggie
    Oct 4 at 0:13
3

Stop worrying.

If you'd fried something, you'd know about it. The damage from plugging something in whilst the system is powered generally comes from a short when the wrong pins make contact, this isn't something that is going to "weaken" components. It'll totally fry them if it happens.

If your components are good quality, it's unlikely they'd be damaged from being plugged in while the system was powered anyway. This is especially true for RAM because it's pretty hard to short something as it's going into a keyed slot, so everything physically has to align.

5
  • 2
    Damage can occur if data pins are powered up before the power pins. Hot-swappable things like SATA have different length 'pins' for this reason. It's also very easy to misalign a RAM module in the process of plugging it in and create shorts!
    – abb
    Oct 4 at 0:05
  • 1
    @abb, yes, but the damage in this case is "all or nothing": either it works properly, or it's completely non-functional. Hot-plugging something not designed for hot-plugging doesn't "weaken" it.
    – Mark
    Oct 4 at 3:39
  • @abb - Technically, yes... In reality, it's unlikely with anything even partially well designed. I also don't think it's that easy to misalign a stick of ram, unless you're just cack-handedly trying to force it in. The key slot makes sure it goes in at least vaguely straight unless you're just full on jamming one side down Oct 4 at 8:50
  • 1
    @Mark - partial breakage is definitely possible. I've seen broken RAM that still allowed the OS to boot. That's why we have memtest :)
    – abb
    Oct 5 at 5:28
  • 2
    @abb - Sure, partial RAM breakage is possible, but from shorting power pins, probably not... That'd make the blue smoke come out Oct 5 at 9:27
1

The worst case scenario for you isn't that the RAM fails completely, but that it has become unreliable. Failed RAM is easy to see and fix. Unreliable RAM will seem to work, possibly for months, and then give you random inexplicable errors. This has sometimes been known to happen if you don't properly ground yourself before touching chips.

Also, there is the possibility that the other side of the RAM connector (i.e., the motherboard) could have been damaged.

I'm not saying anything about how likely either scenario is.

I second the recommendation from others: run a burn-in test using memtest86+. Given the scenario, I would run this test for longer than normal, possibly as much as a week or longer. I had one system that had obvious, but rare, symptoms of memory failure, and memtest86+ basically never saw the error even if run for a week. I might have needed it to run even longer.

Another aspect to consider, although that is unlikely in your scenario because you never accessed the hard disk. I had one system with memory failure. Even after replacing the memory, the system would crash during operating system install; the faulty memory had caused the SSD to get corrupted. Wiping the SSD from the BIOS solved that problem.

Finally, one more thing I noticed. You said that your PSU has been sitting on the shelf, unused, for five years. Five years is not all that old, but do keep in mind that some components, capacitors in particular, do age with or without usage.

2
  • Do you have any source I can read about this? All my searches the last couple of days that backs up that memory can degrade by an event like this leads to nowhere.
    – LPChip
    Oct 3 at 10:46
  • No memory tester can test for all possible usage patterns. Overnight or a little longer memtest86+ is to be trusted. If memtest86+ ran for 48 hours with 0 errors, it would likely run successfully for a year with 0 errors. Which still doesn't completely rule out an occasional hardware failure under very specific conditions, including some of those that recur during ordinary use - or shortened life. Oct 3 at 20:22
-2

You would have needed to SERIOUSLY fry something, for you to actually be able to smell it. Trust me, I fried many boards in my younger days, and never did they actually smell any different. But I stopped building boards; I just had no good need that was worth the risk, and unlike some people, I got no particular joy out of it. If it still boots fine, and everything else seems good, including the new RAM, you're probably ok. You can still run some tests as others have mentioned; that's probably a good idea on any new build. Once the basics checkout, you can find a stress/bench test to really see how well you did.

2
  • 1
    Your answer could be improved with additional supporting information. Please edit to add further details, such as citations or documentation, so that others can confirm that your answer is correct. You can find more information on how to write good answers in the help center.
    – Community Bot
    Oct 4 at 18:19
  • "Worth the risk"? If you're even somewhat familiar with PC hardware, building is less risky than buying a prebuilt. In both cases you can return faulty hardware.
    – Navin
    Oct 22 at 5:56