PC resetting after adding memory

Craig R

Craig R

Craig R

Associate
Joined
11 Oct 2005
Posts
2,358
My motherboard (MSI Gaming 470 plus) had 16gb ram in it (2 x 8gb 3200mhz PC4-25699 Patriot Viper stick with timings 16-18-18-36 CL16).

It's been running fine with those for a few years but I decided it was time to beef it up to 32gb and add another couple of 8gb sticks.

I ordered Patriot Viper again with what was advertised as the same timings as the above sticks on the product page.

However I was getting random PC resets after sticking them in. I've removed the original 2 sticks, put the 2 new ones in, and it seems to be stable with no reset so far.

However, using CPU-Z, it says the timings of these new sticks are actually 16-20-20-40. Not what was listed on the product page (16-18-18-36).

Is this likely the cause of the crashing when all 4 sticks are in? Should I return the new ones?
 
What's the CPU?

What happens if you set it to 3200 with board timings and the XMP DRAM voltage?

Does CPU-Z say they're the same memory manufacturer?
 
CPU is Ryzen 9 5900x.

Spoke too soon, just reset itself with only the new ram connected.

How do I set the board timings and voltage? I don't see these options in my bios.

20221219-165835.jpg
 
Craig R said:
How do I set the board timings and voltage? I don't see these options in my bios.
Click to expand...

It will likely just use default timings if XMP is off and you set a manual frequency. Not sure where DRAM voltage is in your BIOS, I looked at buildzoid's BIOS walkthrough of a MSI X470 board and it was in: overclocking: DRAM voltage, if that helps.
 
Tetras said:
It will likely just use default timings if XMP is off and you set a manual frequency. Not sure where DRAM voltage is in your BIOS, I looked at buildzoid's BIOS walkthrough of a MSI X470 board and it was in: overclocking: DRAM voltage, if that helps.
Click to expand...

Hmm, manually set it to 1.35v (it had set auto to 1.365) but it doesn't seem to have helped, still getting crashes.

Put the old ram back in and it shows all the correct timings in bios right away.
 
Craig R said:
Hmm, manually set it to 1.35v (it had set auto to 1.365) but it doesn't seem to have helped, still getting crashes.

Put the old ram back in and it shows all the correct timings in bios right away.
Click to expand...

You don't want a lower voltage than the default, if anything 4 sticks should have a slightly higher one.

The reason I suggested motherboard default timings is because they're usually looser than XMP and if you're using mismatched sticks (& 4 of them) looser timings might help with stability.
 
It looks like auto stuck it on 1.2v and 2400mhz by default.

I've manually set this to 1.38v and 3200. Fingers crossed that gets me back to stability on the old sticks. Held for almost 10 mins so far.
 
Craig R said:
Same slots they were in previously (2&4).

Just back on auto for frequency and dram voltage.
Click to expand...

With XMP enabled? Sometimes XMP increases other system voltages too, to ensure stability.

If your system is random rebooting with 2400 @ 1.2v then I'm afraid the problem might not be RAM related.

Did you touch anything else when you put the RAM in, like reseating the cooler, or pulling power plugs?

By 'crashed', do you literally mean it resets, with no blue screen, or nothing?
 
Tetras said:
With XMP enabled? Sometimes XMP increases other system voltages too, to ensure stability.

If your system is random rebooting with 2400 @ 1.2v then I'm afraid the problem might not be RAM related.

Did you touch anything else when you put the RAM in, like reseating the cooler, or pulling power plugs?

By 'crashed', do you literally mean it resets, with no blue screen, or nothing?
Click to expand...

Straight reset. No error. Xmp disabled.

Didn't pull any plugs (at least that Im aware, everything still looks to be connected ok) or reseat the cooler.
 
running all ranks populated on any config can often be fiddly as the memory controllers often struggle a bit more when trying to run all dimms at faster speeds, id up the voltage to 1.35v, vddcr/soc voltage up to around 1v (for 3000 series but id check online for recommended 5000 series voltages), the memory controller in ryzen is in its uncore so you need more voltage to that. Id also have some load line set as well. I had to mess around for a while to get 8x8gb sticks all working at my memory's rated 3466mhz on my 3960x. As above what is the system doing resetting itself, hard locking? what sort of crash are we talking here?
 
Last edited:
in event viewer filter system logs for critical, error and warning? Should be able to find the kenel power error and have a look at the timestamp and look what happened in the run up. But if its stable with just the two sticks now but not 4 sticks then you need to tune some voltages :)
 
There are kernel errors but they just randomly appear with nothing obvious happening in the run up.

It's not stable with 2 sticks any more either
 
Craig R said:
Could this be a dodgy psu?

There's a spare I might be able to pull out of my old work pc when I'm in the office tomorrow.
Click to expand...

Logically, if the system worked 100% fine before you installed the new memory and you touched nothing else, then the answer should be no.

My best guess would be that you haven't restored the settings to whatever they were, when it was previously stable.

But, if the new memory is a red herring (i.e. concealing an existing problem) then yes, a dodgy PSU can cause resets, but then, so can other things, like the RAM, the CPU & the board, or even the case if something is shorting (or the front panel is faulty). Another left-field example: malfunctioning usb device that shorts or trips the ports.

Does it always reset doing the same things, or can it reset when gaming, when idle, when browsing?
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom