Last try before I throw this whole rig into the bin and buy a prebuild

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
V

VioDes

Prominent
Jun 19, 2017
38
0
530
Hallo people,
as you can see in the title, this is my very last attempt to get my new self built system to run.
I've been working on it for several months already, but as soon as one issue is resolved, a new one comes up.

I listed this as a "Windows 10 Subject" but really it is pretty much everything.

So lets start with the system:

CPU: Ryzen 5 1600
GPU: Zotac Gtx 1080 Amp Edition
MB: B350 Tomahawk MSI
PSU: 550 W Seasonic Semimodular
Ram: 4x 8GB Corsair Vengeance 3000 MHz LPX CMK16GX4M2B300C15B
SSD: ~500GB Sandisc Ultra II (OS is installed here)
HDD: 3TB-I don't even know the brand-Datagrave

The problems occur even though no other program is installed on the system. The OS is Windows 10 pro.

The issue(s):
1. When I try to run a heavily taxing program, like Furmark or games like Witcher 3 and/or GtA 5, it crashes to Desktop. Sometimes it simply freezes without even giving a crashreport. Sometimes it crashes and gives some sort of report (I don't remember right now what kind of reports I get in those cases). Sometimes it even works for an hour or so, but usually only for a couple of moments.
The windows event logger (I don't know the exact name in english, since I'm translating on the fly) tells me, that either xyz.exe has had some sort of error, yet it doesn't tell me what kind of error (these error messages appear when the game crashes with a notification onscreen) or it tells me that the nvlddmkm program was not found or stopped working or is not installed or corrupted (these error messages appear when the game freezes and simply goes to Desktop without further notice).
The eventlogger additionally gives me a metric F*ckton of other error messages, but after looong research in the web those seem to be non critical, so let's ignore them for now. Just wanted to have it mentioned.

Talking about research in the web, the nvlddmdkmkmkmkmkm error seems to be quite old already. I found threads dating back to 2010 where people complained about the issue. However not a single one could offer a solution. All those threads simply ended with people giving up and waiting for patches.
I found one article on the nvidia page, that linked the issue to a whole butt load of possibilities. Most of those didn't apply to me, however one thing proposed were problematic ramsticks.
[In this very moment my pc froze, not to be recovered. Had to force choke it off. Eventlogger said it was nvldmschmeckldnmdvbdsam again. It hadn't happened to me when only browsing until now. However Firefox somehow saved what I am texting here, so at least that didn't hurt.]

2. When building the system the first time, I already checked with memtest for faulty ramsticks. All came out working.
I tried it again, today. However this time I only tested one stick at a time. All passed, after that I took one stick and reseated it into each slot, to find out if anyone of those are broken. Nope, all working.
After that I put them all back in, but suddenly the system refused to post. It did that "try to boot 5 times, if it doesn't work by then, just stop trying"-cycle. After experimenting for a while, I found out, that while all slots are fine and all ramsticks are fine, I cant boot the system with all four installed. It doesn't matter which of the sticks I don't install, as soon as three or less sticks are in there, it boots. It also doesn't matter which slot remains free, tried 'em all. As long as one is free, it works.
Right now I have three installed and one stick is sitting on the table laughing at me. I literally changed nothing about the system, no damned idea why it suddenly refuses to eat all four sticks. They worked yesterday.

3. On a similar matter, that means ram again; I can't oc it. Or XMF/R or whatever it is called. They are rated for 3000 MHz, but as soon as I increase the frequency from the 2133 default, the system refuses to boot and goes cycling again until it reverts the change and boots again with 2133 Mhz. That however hasn't changed from the beginning. I thought I matched up a noncompatible stick-mobo combo and left it that way. However I doublechecked it and the vengeance lpx is supported. Again, I have no damned clue where the problem is.


what I already did:
install all drivers, reinstall them, downgrade them, even tried it without drivers at all (lel)
updated bios
cleared cmos
ram check up as mentioned above
i gave up
"repaired" windows with that repair function
formatted C:
reinstalled windows from scratch (same errors)
found a little hope again
all sorts of stresstests (memtest, furmark, some weird-ass cpu test)
gave up hope for good

tl;dr

-nvdlddlflklmsdlknsldkasdmdlksmdlsllknjdsdbsldblsbdd has some sort of issues and there is apparently no solution for it.
-my ramsticks had a sexchange last night and now only operate when in the right mood (#nosexism)
-can't oc my ram

I'd like to think the non-working nvldldldldlixlasidllxlixliaxjdliasdjasdapxoclkasdxalxkashdlamadbskdujasxölxkahsdöliasdasdax program might have a connection to the newly introduced femininity of my rig, however the nvldldldlasldlxxlasdilaisdxlxalsdahliwdhlakxldhalisdhwlmasbdkajwdgkausdhkjxkhuxlasidxlkaxlixhlmgg has had problems even when all four rams were working. So its just as likely, that there is no connection between the two at all.

I hope I made the matter clear-ish. Maybe some of you can help me out, even if only with one of the issues. If there is anything I forgot to mention, I will edit the post. The "what I already did"-section seems awfully short for such a long struggle so I must have forgotten halv. I'll update as soon as i remember.
Thanks for reading guys.

Cheers

Edit the 1st: All temps are always in normal range. During Furmark the Gpu hits 83 °C.
Edit the 2nd: Neither CPU nor GPU have been OC'ed. The gpu has a factory oc though. Used Afterburner to reduce the frequency by 500 Mhz, didn't change a thing.
 
Sort by date Sort by votes
for what its worth, I had a lot of bluescreens with win 10 (assuming pc boots up offcourse). There is a win 10 group that send you a analysis program, you run it, it checks why specifically crashed, and gives you a very specific reason. Might be worth a try as could point you in the right direction. Here is the link : https://www.tenforums.com/bsod-crashes-debugging/

I did use it and they were great.

Hope this helps a bit
 
Upvote 0 Downvote
From what I've read, it's most likely the MB, or PSU. You could get a PSU @ Best Buy, and return it if necessary (Make sure though). Also, don't throw away your PC. I'll take it.
 
Upvote 1 Downvote
I suspect that your memory is not compatible with the B350 Tomahawk MB. The MSI site says it supports:

• 4 x DDR4 memory slots, support up to 64GB
- Supports DDR4 1866/ 2133/ 2400/ 2667(OC)/ 2933(OC)/ 3200(OC)+ Mhz *
• Dual channel memory architecture

But your listed RAM is 3000mhz RAM. A search listing your Corsair RAM and the MSI B350 Tomahawk MB returns a number of folks reporting bad behavior with the combination.

It just a suspicion, but if you have any other RAM around you could swap in its place to test with, that might be a god place to start.
 
Upvote 1 Downvote
hey guys, thank you all for your input. I'll respond to each of you here. Ill mark your part with an @"yourname".

@grumpy275, well okay then all i can do is hope, that it is the mb. Can it be users error, that caused this static? Or is it a matter of production? I bought two of the four sticks on amazon. But its already a couple of months ago. Don't know if they are willing to take them back.
Germany has its own version of tomshardware. So it really is quite unlikely😛

@amc2002 | @robert_294 |@tom72468: I use a Seasonic 550 W psu. It delivers stable power. I know that, because the first gpu i bought was dead on arrival. I bought it a week before that mining craze went completely bonkers, so i was quite desperate to eliminate to possibility of the gpu being dead. Back then i took a multimeter and checked every single output of the psu. They were all where they are supposed to be and didn't fluctuate at all. Its rather unlikely, that that changed.
And what do you mean with PAR repair?
Also it was only a figure of speech, when i said, ill throw the system 😉

@32starview that might end up being helpful. Is there a specific group or thread i should go to? the link brings me to a more general "crash and bug" forum.

@signal5 do people really have issues with the combination? i checked the compatibility list of the b350 list(https://www.msi.com/Motherboard/support/B350-TOMAHAWK#support-mem-2) and there are a literal metric shitton of vengeance sticks compatible. i really cant find the exact one i have on the list (mine are the cmk16gx4m2b3000c15b (the b in the end stands for blue)). however since all others are compatible, why shouldn't mine be? i already noticed, that in the bios, i cant oc the memory to 3000. The 2933 version of my sticks is compatible, and 2933 is selectable in the bios. Does the incompatibility go beyond the fact, that i cant oc the last 67 Mhz? They don't even run stable at 2133 MHz anymore (the 2133 Version is supported, and a couple of months ago, the system was stable at 2133).
another thing that makes me doubt the incompatibility is that, paulshardware (the youtube guy) built a system with the same ramsticks i use (also 3000Mhz) in a b350 tomahawk.

thank you all guys for your aid. right now im still waiting for corsairs answer to my request. msi didnt bother answering another time, after they told me to update the bios.
as soon as i have verified, that neither of the 2 companies has motivation to help me nor an actual clue, i will try to rma the mb. It is also already a couple of months old, so i'm in doubt, that amazon will take it back, but we will see.
i cant rma the ram, since i only have the receipt of one pair (i bought 2 stacks of 2 8Gb sticks), so if it ends up beeing the ram, im kinda screwed over a couple of hundred bucks :/
also just fyi: while i bought one batch on amazon and the other batch in a local store somewhere, the serialnumbers are incredibly close to each other. they end with 62 and 63 from the first batch and 66 and 67 on the second. what are the chances? 😀
 
Upvote 0 Downvote
Not going to lie. I quit reading after about half way through. As Signal5 mention above. Make sure the motherboard supports the ram. If it does. And you have no way to test the ram with another pc or friend. (if you their is a pc repair near by. So if they will put the ram into one their computers if possble.) But if you can't do either of those things. I would send in the the motherboard for warranty. I recently put in 7th gen processor i7 in the motherboard. Had i5 6th gen before. And even after updating the bios. The motherboard didn't want go past the post screen. And didn't like my graphics to be in. I had to send my motherboard for warranty. They had to flash the bios. And re put in the bios. Basically the bios got bad someone. Mine was an ASUS. So you might have something similar. I would call or email your motherboard manufacturer. They will probably fix the issue. like they did mine. Or if it's unfixable they will probably send you an new one.
 
Upvote 0 Downvote
First, Corsair has lifetime warranty on Vengeance LPX ram and they have awesome service. I don't think you'll need the receipt. If it's still blue screening at 2133 (no OC) then it must be the MB.
 
Upvote 0 Downvote
I have built numerous computers over the years. I'd guess about 25% of the time they don't work. Here's what I would do in your place.

1. Try to make it fail easily. The only way to debug a hardware problem like yours is board swapping. If it takes 8 hours to make it fail, it's going to take a very long time to find the problem. Run memtest and anything else you got until you can make it fail in a short time, say 10 minutes at most.

2. If you don't have the parts to swap in, then take it to Frys and pay them to debug it. They have enough spare parts to swap in, but you need to be able to make it fail quickly, as mentioned above.

I like to find bargains, but when building computers I buy from a local store (usually Frys) where I can bring things back with no penalty. My last build was with an MSI motherboard. After a few hours of testing I discovered that running full out was creating enough rf noise to interfere with my OTA tv reception. I took the board back to frys, got a gigabyte board and the problem went away. Without the ability to swap all my parts, I might have given up. But I could repeat the problem easily and quickly, and could use frys to swap parts at no cost to me.
 
Upvote 0 Downvote


Have you tried to update your Motherboard BIOS & GPU drivers?
 
Upvote 0 Downvote
@ciloit | @xpjay got in touch with both msi and corsair, msi provided me the most ridiculous copy and past answer i ever received and corsair didnt answer at all :/ sooo yea...customersupport ftw

@rocket88 the issue is, that the bsods come incredibly unpredictable :/
however i usually buy all my parts of amazon and they have quite a good return policy, its just this time with 2 of the 4 ramsticks, that it is kinda messy :/

 
Upvote 0 Downvote


 
Upvote 0 Downvote
Hi

I have had previous similar problems which lasted months. I tried every fix that had been suggested on the forums which were mostly stupid. My instinct was it was the ram. I had tried Memtest 86 and windows own mem check. I ran tests overnight and no problems were shown. I tried the two sticks individually (Gskill Trident z DDR4 PC 3200) in every permutation of slots. It was shutting down unexpectedly when gaming, initially, then just randomly. The error checks indicated it was a PCH problem and I was considering a new motherboard or CPU as I had tested the ram. I persisted with the ram and I found one of the sticks was causing the problem. I bought Corsair 3000 memory and installed it. I never had the problem recur and returned the Gskill which had worked for 11 months. Persist in trying or swap the ram. I had a similar problem previously.

Good luck. Cee
 
Upvote 0 Downvote
I read it through, but may have missed it. Have you tried reinstalling the Motherboard BIOS? Checked for updates? Many times these updates happen well after the boards hit a storage room shelf.
 
Upvote 0 Downvote
First, I had pretty much the same exact issue at one point but only with a different setup. After many hours of trouble shooting I discovered that I had bent pins on the MB CPU socket. They're not easy to see and you will need plenty of light and possibly may need a decent magnifying glass. I was able to bend the pins back into place. This has happened twice to me now with this same MB, I have no idea what happened the first time but the second time I accidentally dropped the CPU onto the board... This sounds like what your issue could be.

One other thing to note is you did not purchase your RAM in the same set, you purchased two different sets. Which can be a bad thing. This can cause some serious issues with stability like you have mentioned and with using the auto RAM settings (it's called XMP on my MB). When you tested each stick of RAM by itself you mentioned that only one completely passed. But when testing did you use the same slot on the MB for each stick of RAM?
 
Upvote 0 Downvote
update: Got the MB Rma'd and got a new one, since there were no 100 % certified compatible mbs for my ram, i went with the b350 pc mate from msi since i am already familiar with the bios. All other b350 mbs didnt even support the slower/faster versions of my ram. The pc mate at least supports the slower and faster versions, so i kinda had tied hands on this one.
However, while the nvnvnncmvncmvnncmvncvmbcvmcnvmcvncmv error seems to be gone, programs still crash and i still get bluescreens. so it helped, but not enough.
What are the chances that the mb and the rams are faulty?

@ceefeesh im afraid this will be the next step i have to resort to. afraid because i dont have the receipt for 2 of the 4 ramsticks.
@yea, did that
@madmaxneo i highly doubt i am that stupid, but might actually be worth checking. note that i have an amd cpu, that means the pins are on the cpu and not on the mb. but that doesnt stop them from bending.
Also yea, while i did purchase them separately, they are the same. even the serial numbers are almost ongoing (only 2 other sticks were sold in between)
well yea, the testing was quite complicated. some passed, some didnt and in the end all caused bluescreens. that means when i tested the first one, it worked. the later ones did get errors and bluescreens. so i went back to the first one in order to boot, however now, that one caused bluescreens as well. I checked all sticks in all combinations in all dimmslots. Its really weird.
 
Upvote 0 Downvote
I'm not trying to break your balls man, but that's exactly the kind of thing that makes me more and more convinced that prebuilt is always the way to go.
Pick the parts yourself but let a tech put them together and then test the system in front of the tech and see if it all works fine. Also don't forget the warranty on all the parts of course.
 
Upvote 0 Downvote
@luminez, i think i actually might have solved it. i think at least one of the ramsticks is faulty as well, so i simply removed it and the other one from the same batch and didnt crash any more. also when only running with those faulty sticks i got a bluescreen when memtesting.

@rafael mestdag well, its a trade off. first of all, you get to pay ~30 % more for the same product.
2nd: you usually dont get exactly what you want. eg your favorite case etc.
3rd: you cant mod it without voiding the warranty, that usually means no led strips or anything since you cant open the case.
4th: with an error like this, a tec wouldnt be able to help you. youd send the pc to them and tell them there is something wrong, but you dont know what. So theyll end up doing their 5 min routine check which might just as well not crash it. Then they send it back to you and tell you, its working perfectly fine. been there, done that.

those put-it-together-for-me-companies are also kinda shitty. my brother had his system built by some company like that. however after a while it broke and even a lawyer couldnt get in contact with that company.
 
Upvote 0 Downvote


By no means if there are bent pins does that in any way indicate you are stupid, maybe a little clumsy (we all are in a way) but not stupid..
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom