[SOLVED] What is killing my ram?

[oldgen]

Recently I built a new pc my specs are as follows:
Ryzen 5 3600
Sapphire rx590 nitro+
Adata d60g 3200 mhz ram (2x8)
Asrock b450m pro4
Cougar stx 700 something psu( kept this from my old rig, which was working perfectly fine I just wanted to upgrade my system,along with my new ssd and old hdd)
Never OC'ed any component

I set up my new rig with no boot problems everything was working fine for 5 days than I got my first bsod and I thought it was caused by storemi with 2gb ram cache. With further inspection I realised that the bsod had different errors at the beginning but then it always said memory management. When I tried the win10 memory diagnostic tool it showed my ram had problems I narrowed it down to one stick of ram since only one of them showed problems in the test. Didn't give much thought to it and RMA'ed the faulty ram, the new stick came two weeks later and until that I used my pc with no problems with the working stick. I tested the new stick and it had no problems, overjoyed with my now complete system I put in the new stick and used my system with no problems (I didn't use storemi again btw thinking it was the culprit) for 5 DAYS AGAIN and the new stick got corrupted as well. The working stick was in the 2nd DIMM slot and the other sticks were always installed on the 4th DIMM slot.
So I either got so lucky to receive 2 faulty ram sticks and have a story to tell my kids or my mobo definitely has a problem (some people suggested a psu check but I'm rather certain it has no problems since it is about 1 year old and I had absolutely no problems in my old rig)
Right now I can use my pc no problem BUT I am so scared something is going to happen to my cpu and gpu.
I bought the mobo and the ram sticks from a distributor of asrock in my country but that firm is such a problematic firm and I am quite certain they will either not rma the mobo or try to pull a trick on the deal (they've had 240+ complaints in a year from customers).
I'm considering buying a new b450 mobo because I really don't want to be left without a pc in a possible time of lockdown what do you guys think should I bite the bullet and buy a mobo or try my chances with a possibly faulty mobo that worked for 2-3 weeks

Is there any way I can check my mobo without risking the working stick?

 

[kanewolf]

Recently I built a new pc my specs are as follows:
Ryzen 5 3600
Sapphire rx590 nitro+
Adata d60g 3200 mhz ram (2x8)
Asrock b450m pro4
Cougar stx 700 something psu( kept this from my old rig, which was working perfectly fine I just wanted to upgrade my system,along with my new ssd and old hdd)
Never OC'ed any component

I set up my new rig with no boot problems everything was working fine for 5 days than I got my first bsod and I thought it was caused by storemi with 2gb ram cache. With further inspection I realised that the bsod had different errors at the beginning but then it always said memory management. When I tried the win10 memory diagnostic tool it showed my ram had problems I narrowed it down to one stick of ram since only one of them showed problems in the test. Didn't give much thought to it and RMA'ed the faulty ram, the new stick came two weeks later and until that I used my pc with no problems with the working stick. I tested the new stick and it had no problems, overjoyed with my now complete system I put in the new stick and used my system with no problems (I didn't use storemi again btw thinking it was the culprit) for 5 DAYS AGAIN and the new stick got corrupted as well. The working stick was in the 2nd DIMM slot and the other sticks were always installed on the 4th DIMM slot.
So I either got so lucky to receive 2 faulty ram sticks and have a story to tell my kids or my mobo definitely has a problem (some people suggested a psu check but I'm rather certain it has no problems since it is about 1 year old and I had absolutely no problems in my old rig)
Right now I can use my pc no problem BUT I am so scared something is going to happen to my cpu and gpu.
I bought the mobo and the ram sticks from a distributor of asrock in my country but that firm is such a problematic firm and I am quite certain they will either not rma the mobo or try to pull a trick on the deal (they've had 240+ complaints in a year from customers).
I'm considering buying a new b450 mobo because I really don't want to be left without a pc in a possible time of lockdown what do you guys think should I bite the bullet and buy a mobo or try my chances with a possibly faulty mobo that worked for 2-3 weeks

Is there any way I can check my mobo without risking the working stick?

What is "killing" it may be that you bought two individual DIMMs rather than a set. The fact that you can RMA a single DIMM says you didn't buy a set. There is no guarantee that two random DIMMs, even "the same" will work together.

 

[oldgen]

What is "killing" it may be that you bought two individual DIMMs rather than a set. The fact that you can RMA a single DIMM says you didn't buy a set. There is no guarantee that two random DIMMs, even "the same" will work together.

But would that be enough to cause such a problem and damage it that the single ram would not work on its own? Plus if that was the case why would it work for 5 days exactly it would make sense if didn't work right from the start but why would it stop working after a certain time?

 

[Darkbreeze]

The Cougar STX units are low quality, and that 700w model is likely VERY old. I would definitely start there. Lack of adequate clean power delivery can certainly be reflected in poorly working hardware.

Discussion - PSU recommendations and power supply discussion thread - Tom's hardware

Below you will find MY standard list of recommended power supplies and beyond that this thread is intended as a landing place for questions or discussions regarding specific units, platforms or related PSU tech, all of which are all welcome to be discussed here. If it's related to power supplies...

forums.tomshardware.com

Furthermore, since you are using mixed sticks, that too is a big problem. From one boot to the next, the system may re-train and settle on completely different settings if you are using a mixture of DIMMs that did not ALL come together in one kit. The bottom line is, there are never any guarantees that you will have trouble free memory operation unless every stick in use came together in a single kit. It is far more likely to be your memory configuration especially since you admit that you had no problems at all when using a single stick.

Get a new, reliable quality power supply and use ONLY memory modules that came together. Even multiple kits of the same exact memory kit model opens you up to the potential for incompatibility.

 

[oldgen]

The Cougar STX units are low quality, and that 700w model is likely VERY old. I would definitely start there. Lack of adequate clean power delivery can certainly be reflected in poorly working hardware.

Discussion - PSU recommendations and power supply discussion thread - Tom's hardware

Below you will find MY standard list of recommended power supplies and beyond that this thread is intended as a landing place for questions or discussions regarding specific units, platforms or related PSU tech, all of which are all welcome to be discussed here. If it's related to power supplies...

forums.tomshardware.com

Furthermore, since you are using mixed sticks, that too is a big problem. From one boot to the next, the system may re-train and settle on completely different settings if you are using a mixture of DIMMs that did not ALL come together in one kit. The bottom line is, there are never any guarantees that you will have trouble free memory operation unless every stick in use came together in a single kit. It is far more likely to be your memory configuration especially since you admit that you had no problems at all when using a single stick.

Get a new, reliable quality power supply and use ONLY memory modules that came together. Even multiple kits of the same exact memory kit model opens you up to the potential for incompatibility.

Oh no when I tested the sticks one by one one of the sticks showed hardware errors, the sticks are not a kit but they're the exact same modules I don't see why an incompatibility issue would fry a ram stick

 

[Darkbreeze]

Ok, well good luck. It can only be the fact that you have cheap memory, or a cheap power supply, or a cheap motherboard, leading to failure. Perhaps a combination of all three. Could also be something else such as a CPU with bent pins that is damaging your memory. It would be a good idea to remove the CPU and TRIPLE check for ANY abnormal or bent pins. Also, it's not uncommon for motherboard shorts to cause this, so removing the motherboard and checking to see that there are no visual signs of damage and that there are no standoffs in the wrong place or other potential issues for shorts present.

This would be recommended, regardless of the apparent issue or lack of indication of WHAT the issue is.

How To - Motherboard bench testing

Finding the problem through bench testing If you are here then it’s likely you have encountered a serious hardware issue and have been unable to resolve it using the standard no-POST troubleshooting procedures. If you have not yet attempted to resolve your issues using the no-POST...

forums.tomshardware.com

Also,

Voltage regulation measures how closely the PSU's DC output voltages adhere to the voltages defined by the ATX12V Power Supply Design Guide Specification across the full range of the PSU's output load. A PSU with perfect voltage regulation would have no variation in its output voltages no matter what load is placed on it.

Ripple voltage is an artifact of the AC to DC conversion process by the rectifier. It appears as a high frequency (i.e. choppers switching frequency or a harmonic of that frequency) AC voltage that is superimposed onto the DC output voltages. Capacitors are used to filter out the ripple but its not perfect so some ripple does still get through. Insufficient filtering in the PSU just means that the attached devices own filtering circuits, if they have a filtering circuit, will have to deal with the high levels of ripple. If the ripple is excessive it can cause damage to the capacitors by overworking them on the attached devices (e.g. motherboard, graphics card, HDD's controller board, etc.)

So that cheap PSU, and that Cougar STX 700 IS a cheap power supply, and one that is also likely old given that the 700w model hasn't been manufactured in quite some time, could definitely have caused some damage to your motherboard which could account for the memory problems you are having as well. Replacing hardware does no good if the source of the problem is still present. Replacing that unit with something that has substantially better quality SHOULD be the FIRST order of business, to ensure that what you replace later that IS having a direct result on your memory being damaged, stops happening. Even if that scenario is not exactly right, it CAN be, and that is enough to warrant replacement with a reliable unit. Otherwise, you are likely to end up simply chasing ghosts indefinitely.

 

[Calvin7]

If you detect faulty RAM, return ENTIRE kit under warranty for replacement, not half, otherwise you end up no longer having a kit of matched and tested memory.

 

[oldgen]

If you detect faulty RAM, return ENTIRE kit under warranty for replacement, not half, otherwise you end up no longer having a kit of matched and tested memory.

I have TWO SINGLE MODULES NOT A KIT and one of the modules just DIE when I put it in B2 slot after 5 days NOT LOSE SYNERGY/COMPATIBILITY/FRIENDSHIP/RELATIONSHIP with the other one it just straight up DIES/Loses the ability to run on its own.

 

[Calvin7]

Memory modules sold as Single Modules have no guarantee to work in any configuration other than as a Single Module.

Use of multiple modules of them have no guarantee to work together.

You are free to do your own compatibility testing of those combinations, with no guarantees.

 

[Darkbreeze]

If memory dies when you put it in the motherboard, then either you are putting in the wrong kind of memory, you are not installing it correctly OR the motherboard or CPU are killing it. As I ALREADY said, you NEED to pull the CPU and check for bent pins that may have occurred during installation. One bent pin could create a problem where a single DIMM is affected or damaged, while not affecting the other. The same is possible for a standoff or fastener under the motherboard between the motherboard tray and motherboard in the wrong place or in just the right place in the case of a fastener.

An aftermarket CPU cooler that is too tight on one side or one corner can sometimes cause similar problems, although I've never seen it actually damage memory that way, I have seen it cause it to not work or various other problems since it tends to cause the CPU to "cock" in the socket and either break contact or contact a pin against something that it shouldn't be contacting. Evenly tightened aftermarket coolers are what you want to see. One or two turns tighter in on spot than the others is probably enough to cause a problem.

If there are no bent pins, and you need to CHECK to be sure, and nothing between the case and motherboard such as a standoff intended for a different form factor motherboard that is in a spot where there is no standoff hole on the motherboard being used, and the cooler is not relevant, then either you have a faulty motherboard, bent connections in the DIMM slot, etc., or a power supply issue of some kind or you are simply getting bad memory. With Adata, it wouldn't surprise me at all. They are not a high quality manufacturer IMO. They are a budget manufacturer and as such you are a lot more likely to get an inferior product or even have been sent back the same memory module you sent them, so that you actually installed back in the same problem you had to begin with.

 

[MurdyPlay]

i am not sure if you can kill ram with software. if you always have different errors in bsod, most likely you have bad memory chip. That is not your fault, that is factory issue. Try to use linux distros or downgrade to windows 8.1. Some linux distros can work with bad memories (I had pc with ddr2 that had more that 320 errors in memtest+. I installed ubuntu there and used it for more than 2 years. Of course i had freezes and lags. But it didnot crash and worked good)

 

[Darkbreeze]

If you had even ONE error in Memtest, then you are accumulating micro errors every time you run something that writes data to the drive, or even just the OS itself. Eventually, and by eventually I mean usually within a few months if not sooner, depending on the severity of the instability, you WILL start having freezes, blue screens, crashes and unrecoverable data corruption from the cumulative errors. That is just a fact. There is no way around that. Memory errors are not a joke and you should never continue to use ANY system that has memory errors or instability at all. Ever.