How can I? Diagnose PC hardware issues

sataconn

sataconn

sataconn

Associate
Joined
25 Feb 2021
Posts
3
I have a newish custom built computer, that I believe has been damaged after plugging into a faulty ungrounded mains socket, at my previous rented address (an old rural house in Portugal).

The PC was a big investment and its current unreliability is, very disappointing.

It freezes far too often during normal operation, and rarely resumes after Suspend. It also fails to boot on first attempt.

Using stock BIOS settings, haven't overclocked.

I'd love to salvage some of the parts for my next PC, or even better, to repair it.

These are the specs; just to give you an idea of what I'm unhappy over...
Corsair Vengeance 2x8GB DDR4-3200
AMD Ryzen 7 1700X
Samsung SSD 970 EVO 500GB NVMe
Radeon RX 550 640SP
High quality PSU (don't recall who made by) with more than adequate supply
High quality case
Asrock X470 Master SLI mobo
Top end CPU fan
Extra, good quality case fans

I've been several PCs prior to this one, I work in IT and know about and am cautious about static electricity; just to give you an idea of my skill level.

But, I don't have a clue whether I'm taking a risk, if I transplant a component which may be the root cause into another machine.

Nor do I know if there might be some way I could troubleshoot the hardware issues that have plagued this machine since I built it over a year ago (unfortunately, whilst living in a house with unhealthy levels of humidity and dodgy electrics).

Any help greatly appreciated - at present I don't know what to do other than buy a new machine with all new parts, and sell this one with a big caveat empor that it's damaged in some unknown way.
 
Newish three generations old CPU?;)
And 3200MHz would be stretching it for Zen(1)'s very mediocre memory controller.

As you mentioned previous address did you move recently?
If problems started after that, did you make sure everything is plugged firmly in their places?
Something could have moved little during moving your stuff to new place.

How do you know socket was faulty?
Lack of safety ground itself doesn't prevent PSU from operating.
 
Try clocking the memory down from 3200mhz.

You said it has been doing this since new or plugging it in in Portugal, your post contradicts itself.

But try clocking the memory down, and/or increasing the memory voltage (slightly!).
 
> Newish three generations old CPU?;)

In my mind this computer was about 2 years old, but now I think about it, I moved slightly longer ago, so it's just over 2.5 years old.
Realizing that makes me feel a bit better about this - knowing it's not brand new anymore. I spent a bit less on the CPU when I built this machine too - knew I didn't need the latest.

> And 3200MHz would be stretching it for Zen(1)'s very mediocre memory controller.

Good to know, I'll bump this down a notch.

> As you mentioned previous address did you move recently?

I built it whilst living at the previous address.

> If problems started after that, did you make sure everything is plugged firmly in their places?
> Something could have moved little during moving your stuff to new place.

Have had problems since machine was newly built, though they've been better or worse at different times.

> How do you know socket was faulty?
Touching wires connected to an Oculus Rift that was later setup in the same room would give a small shock.

> Lack of safety ground itself doesn't prevent PSU from operating.

Great to know :).

Thanks, really helpful.

> Try clocking the memory down from 3200mhz.

Will try this next.

> You said it has been doing this since new or plugging it in in Portugal, your post contradicts itself.

Built it while living there. Still living elsewhere in Portugal.

> But try clocking the memory down, and/or increasing the memory voltage (slightly!).

I've increased the "vcore offset" by 130mV, since I didn't seem to be able to increase the memory voltage.

After doing so, I was able to successfully boot on first try (in recent days, I've had to try several times).
More incredibly, I told my machine to go into Suspend, and then resumed it (I don't think I've been able to do that in over a year!).

Am now going to walk away from my computer and come back later, and see if it hasn't frozen like it normally would have.

Fingers crossed increasing the voltage was all I needed to do, that would be awesome
 
I had a Ryzen 1700 and had overclocked memory to 3000 successfully, it wouldnt run correctly at 3200 as i would have odd crashes and blue screens.

Strip your graphic drivers right back using DDU (display driver uninstaller), use the safe mode as suggested.

Ensure you have the latest AMD chipset drievrs installed for your board.
 
Thanks very much, I'm running at 2933 and +130mW (iirc, changed these settings days ago now) and it has resolved the problems that have plagued this computer since I put it together. Wish I'd asked here 2+ years ago. Was on the verge of buying a replacement, before deciding to go ask some people who might know more than me (wasn't hoping for much, figured diagnosing these types of problems would be difficult and/or costly). Who knew running my RAM at its stock speed, with the default voltage, would cause such instability...

I've even just brought my computer out of being suspended for >24 hours. Usually it can't put up with being suspended for any amount of time, upon resume nothing will happen.

Can't really thank you enough for taking the time to help me avoid spending quite a chunk replacing my machine, something I was particularly not looking forwards to since I could imagine sitting down to use the new machine and instead of being "excited" to be using a better machine, it feeling like the same old machine essentially in terms of performance, just without the crashing.
 
Glad you're up and running!

By the way, XMP is technically an overclock, not guaranteed to work. The 'stock' speed of DDR4 is 2133 or 2400MHz depending on your motherboard. XMP profiles are just preset overclocks, every now and again a system will have a problem with them.
 
sataconn said:
Thanks very much, I'm running at 2933 and +130mW (iirc, changed these settings days ago now) and it has resolved the problems that have plagued this computer since I put it together. Wish I'd asked here 2+ years ago. Was on the verge of buying a replacement, before deciding to go ask some people who might know more than me (wasn't hoping for much, figured diagnosing these types of problems would be difficult and/or costly). Who knew running my RAM at its stock speed, with the default voltage, would cause such instability...
Click to expand...
Like said technically any kind XMP profile is overclocking.
And while JEDEC eventually updated DDR4 standard up to 3200MHz, latencies are way worser.

Anyway highest officially supported memory bus speed for Zen(1) is 2667 Mhz:
https://www.amd.com/en/products/cpu/amd-ryzen-7-1700x
(with tweaked Zen+ Ryzen 2xxx upping that to 2933 MHz)
Anything over that is overclocking also memory controller built into CPU and InfinityFabric bus used to connect functional blocks inside CPU.
Thorough reviews would have told that.

sataconn said:
Touching wires connected to an Oculus Rift that was later setup in the same room would give a small shock.
Click to expand...
Without safety ground in wall outlet capacitors in PSU's EMI/line filter form voltage divider making PSU's casing and PC case float half way between live and neutral. (115V)
Those capacitors (special fail open type) simply limit current from shock to annoyance level.
 
The memory is most suspect here. Because the memory will run at 3200 does not mean your PC will. Folks don't seem to realise that using XMP IS overclocking your PC as well as your RAM so it may not work. Just try running it without XMP enabled for a month and see what happens.
 
You must log in or register to reply here.
Back
Top Bottom