This is the only channel that breaks down component level repair so people can ACTUALLY learn. Edit: Your content is slower and clearer and easier to follow. I send a lot of people to your channel if they are interested. Over the years I gained $10,000 worth of knowledge from your free content. That's why I am proud to pay for extras!
I cant remember when i subbed to this channel but i'm glad i did , taking the time to respond to comments to share views and knowledge in a humble way , tops in my book.
I really enjoy your channel as it has given me a lot of helpful guidance in my computer builds and repairs. I have only built a few for the family but they do require repair for time to time it is really good to watch just how you go about fixing things. Hope you have a great day and get better soon.
...Really appreciate your channel, so good to see no B.S. procedures and info. Lots of little details and thought-trains that help bring me up to speed rather than struggling with modern gear. Well Done!
Have a Merry Christmas Graham, and thank you for your channel content. I have been tinkering with computers since the early 90's (Pre Win 95) and it's now a side hustle to earn me a few extra bucks using my self-taught knowledge. I have been subscribed for about a year now and thoroughly enjoy watching your new and old posts. I do like this format following-up from a previous fix, as I don't usually get too far down the list of comments before moving to the next video.
I had success with sfc scannow and dism but of course not for bluescreens. It’s mostly for non working windows updates. It’s probably a good idea to run those AFTER the problem has been fixed to make sure the windows Installation is not in a corrupted state after having crashed a gazillion times.
Love this as as format for new videos. Thank you for the content this year Graham and may there be plenty more in 2025. Merry Christmas and Happy New Year.
I appreciate the discussion. One thing worth noting about bottlenecking: You want your GPU to be the bottleneck. If you're CPU-bound, you're going to have a worse experience. If you're GPU-bound, congratulations, you're getting the most out of the most expensive part in your computer. Using 100% of your GPU just means you're at your limit of FPS/graphics settings/resolution that your computer can handle. Being CPU-bound is a much worse scenario because the GPU will be waiting on the CPU to draw frames, causing stuttering, etc.
It is always bittersweet when a smaller, more cozy UA-camr is starting to take off. Your channel is growing, you are on the road to success but with that come the irate randos that have nothing better to do than to shit their half-knowledge all over the comment section. Whenever this moment comes for a channel I like, I cannot help but feeling torn about it. I wish you all the success you deserve of course, but it will most likely destroy the positive vibe the channel has going on in its community. Edit: That is not to say valid criticism isn‘t welcome. It‘s not what you say, it‘s how you say it.
Totally agree. Graham needs to just skip over all the silly comments and not ever get disheartened. I immensely enjoy watching the videos on this channel and my knowledge had increased greatly from the knowledge he passes on.
sfc /scannow worked once for me. Samsung data migration util would not complete an image, ran sfc/scannow and it was able to complete the image. Only time it ever made a difference. Great video!!
That's lucky with it working once. I've tried it many times but hasn't worked at all. It's funny casue it's like any time you end up on the Microsoft forums/help pages, someone always replies with this command.!
A couple years ago I stumbled onto a setting that might save a lot of CPUs from an early grave. I had a cheap laptop that was locking up a lot, at idle it was at 90C. This was a 10th gen i5. cleaning and new paste did nothing. Had no BIOS clock settings at all. So all I did was in Windows Advanced Power Settings, change maximum processor power from 100% to 99%. Processor dropped to low 70C's at idle and never locked up again.
Nice to see you spotted this comment too. Certainly worth looking into in the future. I have a Surface Book 2 that runs hot and I intent to look into the possibility of a heat pipe failure in the coming weeks amongst other things.
the degradation is a myth, the cpu itseñf is not degradating itseñf, it was motherboard vendors pushing too high voltages on the cpu for some reason the most affected is r5 3600, and it was worse like 3 years ago it is the same reason intel tried to push in the begining of the 14th gen fiasco, but it was not that it would bw nice that motherboard manufacturers fought for parts duratiin and not for who runs faster until it explodes
you always make good points looking at trouble shooting of issues, I started with Packard Bell before windows came to life and talking with customers over the phone the only why to find fault with like a no post issue was to take out parts out of the computer until get to the motherboard and relay on beep codes. I aways use let me check what I think is causing the issues and work from there
Very cool.. Lots of people have idea but as you said experience and time in the field 9 times out of 10 will get you the answer or at least in the area of the fault. Love your vids and watching you repair and fix computer and latops and other things..merry christmas to you and happy new year.
I agree with your comments regarding "fragile fixes".The best thing for the customer, and probably most economic given the price of your time, is to fix it correctly and get it out of the door. A massive risk is a fragile fix, even at the request of the owner, can fail again and the customer starts telling others you fixed it but it failed again ... not worth your reputation.
I cooked a motherboard years ago when I discharged static... bumped this webcam I had hooked up, it had a steel enclosure, Bolt cam off my hand and literally blew a chip on the board. It looked like someone put out a cigar on it. Since then I make dam sure to discharge before I touch my system and have for 20 yrs now. I lucked out and board was replaced by MSI with no q's asked. It can happen. I too was like you, started Long time before you did and yes, built on the carpet. My son also did the same, he preferred to build pcs instead of lego when he was a toddler. No issues... was just Dad who fouled up. Cheers and Merry xmas to you and family mate.
it's like system restore, or startup repair, it most likely won't work but sometimes you just need to try it just to rule it out, cross that off the list of things to check just in case it is a rare situation that it works.
Thanks for the video. I watched the repair video aswell. Looking forward to you experimenting with that 5700G. Maybe disabling certain cores and stuff. See if anything changes anything. I find this very interesting. Also built computers since about mid 90s. Been into retro computers the last 10 years. But the last 7 or 8 months my interest for modern PC stuff has blossomed again. Built myself 3 new gaming desktop computers. Two them based on 5700x3d and one based on the i5 13400f.
In my experience it was the stock Wraith Prsim cooler mounted down too tight that destroyed my 2600 and 2700x. For my 5900x I removed the plate on the Wraith Prism that sits in between the heatsink and mounting clips. CPU has been fine for 2 years. Obviously this only worked for me but it is worth investigating.
An impressive beard is a sure sign of brilliance. Big Clive....beard Graham....beard Me....shaved my beard off and now I am unable to tell the difference between a PCI slot and a USB port.
Static damage isn’t an issue in the UK because the humidity level is always high enough to avoid significant static buildup (unless we really try). Even in the middle of summer the relative humidity is normally 40% or higher. In other countries with hotter drier climates it may well be more of an issue.
Have you ever had a following issue (combo is 5800X/AsRock B550) - PC came in for intermittently crashing. Changed to a new Nvme with fresh Windows,, installed all te drivers, put the said NVMe in to another slot, RAM (several combos), updated BIOS to the latest stable, changed graphic cards, etc., the computer freezes or restarts. After it became stable for short while I ran an OCCT & Cinenbench R23 & Furmark at the same time for several runs and all is fine (hours). After letting it cool it continued to freeze. Had to change the mobo. What I find strange that it'll crash/freeze/restart under idle, and under stress it works fine.
Great video. I am a novice for sure.....but....could there be a bad setting in the BIOS? Load BIOS with standard / basic / default settings? Could this be possible?
I’ve experienced “chinese” usb wifi receivers that have prevented systems from booting with a bluescreen error. Happened on three different systems with the same wifi dongle (actually all of them). Ended up narrowing it down to the driver, but just stopped using them.
One thing (imho) worth considering is the similarity to many recent issues that I've seen. While I believe it's related to DMA channel or controller faults, it may not be the actual cause, however similar I believe it to be. Since I have no way of testing this theory, it does appear to be "thematic" of these types of faults. Not sure that this is very helpful information now. If you do decide to further investigate the fault, wonder if crystal disk or similar might trigger it? I don't have one to test currently. I would be interested in your thoughts. thanks for sharing your process. very cool :)
It was quite good you managed to troubleshoot that CPU . Sometimes we can be stuck with a strange problem , that goes rabbit hole . I have one of my older gaming rigs on a Asus MB that haves the NVIDIA chipset , running a early intel quadcore Q6600 before the i 3/5/7 era . Never had a problem at Windows 7 (64bits) era. upgraded to win 10 , but ASUS never released drivers for 10 , one of the missing drivers : the asus ATK ACPI . I'w running the NVidia raid with SSD drives in Raid 5 . suddenly , after a windows 10 update, the PC goes beserk : sometimes taking ages to power up (black screen), often getting stuck , rebooting will get it stuck on the BS ... needed to cut the power from the MB to manage to wake it up again (reset didn't work) . when looking to windows events , manage to see that windows didn't managed my C levels from the CPU correctly , and surprise , the ASUS ATK has been replaced by a MS "approved" . I changed the BIOS C Level configurations , memory size/emplacements, still no stable system if sleep/reboot/shutdown . Changed the power supply ... same problem... starting to think the MB was maybe getting a old cond or something , until I installed linux Mint in dual boot and ... Surprise : the PC works flawless from Linux , power down/sleep/reboot ... so even the OS can generate strange rabbit holes . Merry Xmas Graham
I'm still a firm believer that in this case it was degradation of that particular CPU in hand, It could have been a weak sample from the factory to begin with, Even the golden samples of the highest end line could eventually degrade so much that a crucial link between cores or memory could just give out...sure its unlikely but with silicon quality, Heat cycles etc you never know. :) Merry Christmas Adam :)
In the end it was either a defective or degraded CPU. I guess out of warranty. Seems very likely, being an AIO gaming PC, that the user had probably been running it highly overclocked. When it started acting up, the default clocks were probably reset in an attempt to fix the issue. There's a chance the CPU could be stable with underclocking and light use but gaming days are done.
Most of the timings aren’t motherboard dependent. Like 99% of the timings are internal to the mem ic, though tightening the timings do increase the strain on all of the cpu by the fact of it being more data.
With spare hardware desktop pc are quite easy to diagnose as you stated. Im currently struggling with a defective 2tb m.2 ssd that freezes the file explorer (not boot drive) in windows, drops out and crashes while trying to copying off files. Smart shows it as health yet it has file integrity errors that CMD also confirms. Currently starting a RMA.
Thank you for all the education. And finally, somebody with some gravitas telling the unwashed that the whole "bottlenecking" thing is basically a non-factor. If the parts don't actually damage each other, who cares?
In a 5700g there is still the usb and stuff on the chip which is why the x300 “chipset” exists but apparently the desktops chips do have some onboard Ethernet and usb.
People actually thought it was a corrupt file system from these symptoms? LOL. I hate the internet. "Needs defragged" lol. EDI: bottlenecking too, lol. That popped up when, about 10 years ago? That all depends on what load you're running and what resolution. 4k vs 1080p run massively different on the same hardware. I agree, static has been almost a non-issue for a quarter CENTURY. I've seen maybe 2 dead RAM cards. Nothing else. Not a single dead CPU and I never wore a static strap. I just touch the (plugged in, grounded) case when I start. I do use them on the $240,000 computers I work with though, but only when they're open, not just when plugging.
Graham, on the CPU degradation and Overclock/BIOS reset topics I think you missed something else that I saw no one told you on the past video either. First, I would like for you to see the video that Wendell did on the Intel degradation on 13th and 14th gen testing. The reason I point that out is because I think the culprit here is the freaking Gigabyte motherboard. I've seen a lot of Gigabyte and ASUS motherboards 'causing degradation on the field on common CPUs with normal people using them, CPUs that don't get touched EVER for overclocking. ASUS and Gigabyte give a lot of voltage out of the box with their motherboards, even on Server motherboards for CPUs that should NOT get higher Wattage and voltages than normal for 24/7 use. See the motherboard that Wendell uses on that video for a CPU that should not run at such a high wattage/voltage as the 14th gen 14900k, because he uses a W tier/spec motherboard and the not overclocking CPU goes sky high with voltage/wattage, poor thing. Most of the boards from Gigabyte and ASUS since 2020 up until now have such an aggressive curve for voltage, AMD have real good automation for PBO but, and is a big BUT, if the motherboard sends the signal for more watts/voltage, AMD automatic tuning goes out of spec, just as with the Intel CPUs and THEN the degradation starts. You were TOTALLY RIGHT that a G processor should not degrade or go that high on voltage. But Asus and Gigabyte allow that and push that. ASRock boards are now the best ones for safe voltage boosting and more reasonable power delivery and AMD/Intel safe specs out of the box. I think that trying disabling some cores would be good for your own use just for testing, but it wouldn't be a CPU for someone for everyday use, let alone a customer. Greetings from Chiapas, Mexico. What a lovely dissection video. Happy Christmas.
I have to bring into discussion the fact that 99.9% of motherboards integrators have been overshooting voltages even from AM4 era. I've seen this so many times with multiple integrators that I manually set Vcore on all systems that go thru my hands. BTW the Intel fiasco is a consequnce of the same overshooting done by mobo integrators, like going 1.45 vcore on a 13700kf when the cpu is perfectly stable at 1.16v...
Ryzen…….earlier Ryzen’s have memory controller issues. I’ve seen quite a few Ryzen 5-3600’s with memory controller failures. Increasing SOC voltage seems to help.
Hi there, I have watched a lot of your videos over the years and enjoy watching. I have built many PC's over the years from P4/ AMD3000+ days. I do believe in this case it is CPU degradation. The AMD motherboards do push high voltage no matter the rated power. Having said this I did have something similar to this and it was the RAM needed more voltage. In this case I would push another 0.1V Vcore. I am sure this cpu would then run ok, showing degradation. If this didnt work I would then push another 0.1v on the RAM. I would then reduce voltages until it crashes again. Hope this helps.
It's not so much arguing with the comments, but talking about the ones who might miss the mark, but still raise good points that can be relevant at other times.
Your video are great, quite in-depth, I am quite computer savvy, but your knowledge of the circuits and voltages is way beyond my comprehension in most cases.
22:30 A weak(er) card with a good CPU is not a bottleneck. 1% and .1% are very important in any game in regard to how smooth it feels. The only "complaint" i have of you, is that you refer often to a single person as "they" and that is confusing for non native English speakers like me. You content is great, keep doing what you are doing and i hope in 2025 you will go above 200K subscribers !
I have learned a lot from you. I have a computer I built 4 years ago with the second best Items. Water Cooled NZXT, AMD 9 3800x, Gigabyte x570 Master Motherboard. Gskill Ram, Gigabyte 2090. I have been having Crazy crashes some at startup some at processor high use. Replaced power supply, graphics card, disconnected case switch,, did your mem86 test, windows from a thumb drive, just had Asics symbols on the screen, mounted the processor. I have been messing with this for 3 months now now so I pulled it out and built another computer to replace it. I laid the old computer on it's side and it works fine. Every time I set it up it messes up and crashes at the usual odd times. I have 27 years experience with computers and I have never seen anything like this.
the one thing you did not do in the video was to have cleaned and replaced the thermal paste on the cpu that is usually the first thing i do when i suspect overheating because of computer freezing, because in over 30 years of pc repair i have almost never seen freezing caused by software, crashing yes but freezing hardly ever, so with your experience i wonder if you went that route because you were making video content?
I too do not get overly concerned about ESD as well, but that being said; In the mid-90s when I was first getting into computers as a profession, I had a job with the local computer shop and we had built 26 new computers for a local business. Latest gen everything. Even had the new high speed CD-ROM drives. Each system had been loaded using the Windows 95 CD disc, so we knew all the CD-ROM drives worked when built. Upon delivery day, I went with the boss to help and was unboxing systems on the floor. I opened the box, rolled it over open side down and lifted the box from it. I laid the box down and pulled the styrofoam and bag from the system and shoved them into the box. I turned to pick up the system and as I placed my hand on top of the mid-tower chassis and felt a static discharge as I touched it. Never gave it a thought and continued working. Once the systems were powered up and operating, that one I'd been shocked on had a failed CD-ROM drive. We swapped the drive and it was fine after that, but that was my first and only experience ever having such an issue. The only time I get concerned about ESD is when the hardware I'm working with is irreplaceable.
Too late for comment, i though customer did a very bad OC, leaving the CPU almost dead. A little chance to keep using it for a while lowering Mhz or giving more volt.
Bottlenecking is so overblown by most Tech UA-camrs. Some component is always bottlenecking something. That’s why real world Nvme speed is considerably slower than that rated speed on the box. A PC is choke full of bottleneks………..everywhere. QVL’s were important 40 years ago……..not so much anymore.
This is the only channel that breaks down component level repair so people can ACTUALLY learn.
Edit: Your content is slower and clearer and easier to follow. I send a lot of people to your channel if they are interested. Over the years I gained $10,000 worth of knowledge from your free content. That's why I am proud to pay for extras!
Did you learn how to repair from him alone ?
My only feedback is I would like more of these type of fix videos. Keep them coming :)
Have a Very Merry Christmas, Graham!!
I cant remember when i subbed to this channel but i'm glad i did , taking the time to respond to comments to share views and knowledge in a humble way , tops in my book.
I really enjoy your channel as it has given me a lot of helpful guidance in my computer builds and repairs. I have only built a few for the family but they do require repair for time to time it is really good to watch just how you go about fixing things. Hope you have a great day and get better soon.
Really enjoyed this, hope you post more reading/responding to comments.
SFC and DISM saved tonnes of systems for me, but like you said was defo hardware in this case.
Be sure to DISM before you SFC :)
Enjoyed the vid...Thx Graham!
Merry Christmas to you and yours!
Merry Christmas Adam. Love your channel and thanks for sharing.
...Really appreciate your channel, so good to see no B.S. procedures and info. Lots of little details and thought-trains that help bring me up to speed rather than struggling with modern gear. Well Done!
Have a Merry Christmas Graham, and thank you for your channel content. I have been tinkering with computers since the early 90's (Pre Win 95) and it's now a side hustle to earn me a few extra bucks using my self-taught knowledge. I have been subscribed for about a year now and thoroughly enjoy watching your new and old posts. I do like this format following-up from a previous fix, as I don't usually get too far down the list of comments before moving to the next video.
Enjoyed that ...merry xmas Graham!
I had success with sfc scannow and dism but of course not for bluescreens. It’s mostly for non working windows updates.
It’s probably a good idea to run those AFTER the problem has been fixed to make sure the windows Installation is not in a corrupted state after having crashed a gazillion times.
"on the carpet with a wooly jumper on" 🤣
and now with The Life of Brian in Mind
+1 to beard power
As a fellow techie, I had to subscribe to you and I enjoy your content.
Keep it up brother
Nice to know what you are thinking!
Love this as as format for new videos. Thank you for the content this year Graham and may there be plenty more in 2025. Merry Christmas and Happy New Year.
I appreciate the discussion. One thing worth noting about bottlenecking: You want your GPU to be the bottleneck. If you're CPU-bound, you're going to have a worse experience. If you're GPU-bound, congratulations, you're getting the most out of the most expensive part in your computer. Using 100% of your GPU just means you're at your limit of FPS/graphics settings/resolution that your computer can handle. Being CPU-bound is a much worse scenario because the GPU will be waiting on the CPU to draw frames, causing stuttering, etc.
Great vid Graham, would like to see this more often. And Merry Xmas! :)
Merry Merry Christmas, Adam!!
Thank You And Have A Merry Christmas
It is always bittersweet when a smaller, more cozy UA-camr is starting to take off. Your channel is growing, you are on the road to success but with that come the irate randos that have nothing better to do than to shit their half-knowledge all over the comment section.
Whenever this moment comes for a channel I like, I cannot help but feeling torn about it. I wish you all the success you deserve of course, but it will most likely destroy the positive vibe the channel has going on in its community.
Edit: That is not to say valid criticism isn‘t welcome. It‘s not what you say, it‘s how you say it.
Totally agree. Graham needs to just skip over all the silly comments and not ever get disheartened. I immensely enjoy watching the videos on this channel and my knowledge had increased greatly from the knowledge he passes on.
sfc /scannow worked once for me. Samsung data migration util would not complete an image, ran sfc/scannow and it was able to complete the image. Only time it ever made a difference. Great video!!
It worked for me 2 or 3 times too.
That's lucky with it working once. I've tried it many times but hasn't worked at all. It's funny casue it's like any time you end up on the Microsoft forums/help pages, someone always replies with this command.!
never for me
Good to see you reading the comments , great content you always do on your channel too , wishing you a Merry Christmas Graham 👍
"It's still a broken CPU". I agree. Once we know that we can't fix it, we're done!
A couple years ago I stumbled onto a setting that might save a lot of CPUs from an early grave. I had a cheap laptop that was locking up a lot, at idle it was at 90C. This was a 10th gen i5. cleaning and new paste did nothing. Had no BIOS clock settings at all. So all I did was in Windows Advanced Power Settings, change maximum processor power from 100% to 99%. Processor dropped to low 70C's at idle and never locked up again.
Interesting method, could be good for reducing noise as well. You likely had a bad heatpipe though, check video #423!
Nice to see you spotted this comment too. Certainly worth looking into in the future. I have a Surface Book 2 that runs hot and I intent to look into the possibility of a heat pipe failure in the coming weeks amongst other things.
the degradation is a myth, the cpu itseñf is not degradating itseñf, it was motherboard vendors pushing too high voltages on the cpu
for some reason the most affected is r5 3600, and it was worse like 3 years ago
it is the same reason intel tried to push in the begining of the 14th gen fiasco, but it was not that
it would bw nice that motherboard manufacturers fought for parts duratiin and not for who runs faster until it explodes
Thanks for the great content over the past year, Merry Christmas and a Happy New Year to you, I'm looking forward to seeing your videos next year
It is interesting to hear your response to the less crazy UA-cam comments. I fully support you ignoring the trolls. Don't feed the trolls.
Merry Christmas to you. Good video.
Yep, enjoyed thIs one and more like it would be welcome. Merry Christmas.
you always make good points looking at trouble shooting of issues, I started with Packard Bell before windows came to life and talking with customers over the phone the only why to find fault with like a no post issue was to take out parts out of the computer until get to the motherboard and relay on beep codes. I aways use let me check what I think is causing the issues and work from there
Very cool.. Lots of people have idea but as you said experience and time in the field 9 times out of 10 will get you the answer or at least in the area of the fault. Love your vids and watching you repair and fix computer and latops and other things..merry christmas to you and happy new year.
I agree with your comments regarding "fragile fixes".The best thing for the customer, and probably most economic given the price of your time, is to fix it correctly and get it out of the door. A massive risk is a fragile fix, even at the request of the owner, can fail again and the customer starts telling others you fixed it but it failed again ... not worth your reputation.
I cooked a motherboard years ago when I discharged static... bumped this webcam I had hooked up, it had a steel enclosure, Bolt cam off my hand and literally blew a chip on the board. It looked like someone put out a cigar on it. Since then I make dam sure to discharge before I touch my system and have for 20 yrs now. I lucked out and board was replaced by MSI with no q's asked. It can happen. I too was like you, started Long time before you did and yes, built on the carpet. My son also did the same, he preferred to build pcs instead of lego when he was a toddler. No issues... was just Dad who fouled up. Cheers and Merry xmas to you and family mate.
I've been in this game for about as long as you, Adamant IT, and everything you have said is spot on! :)
i do enjoy watching you fault find and repair the computers and laptops you do really good
Great informations. Thank you
Thank you for your replies! Very informative!
Edit: Almost forgot... Merry Christmas and happy new year! :D
I have had luck with DISM and SFC *once* on a damaged windows install. Significantly less than 1% success rate.
it's like system restore, or startup repair, it most likely won't work but sometimes you just need to try it just to rule it out, cross that off the list of things to check just in case it is a rare situation that it works.
Thanks for the video. I watched the repair video aswell. Looking forward to you experimenting with that 5700G. Maybe disabling certain cores and stuff. See if anything changes anything. I find this very interesting. Also built computers since about mid 90s. Been into retro computers the last 10 years. But the last 7 or 8 months my interest for modern PC stuff has blossomed again. Built myself 3 new gaming desktop computers. Two them based on 5700x3d and one based on the i5 13400f.
In my experience it was the stock Wraith Prsim cooler mounted down too tight that destroyed my 2600 and 2700x. For my 5900x I removed the plate on the Wraith Prism that sits in between the heatsink and mounting clips. CPU has been fine for 2 years. Obviously this only worked for me but it is worth investigating.
In the case of power supplies, they can test as “fine” with a tester, yet fail when they come off of load. Had one of those.
Merry Christmas sir! 🤓
An impressive beard is a sure sign of brilliance.
Big Clive....beard
Graham....beard
Me....shaved my beard off and now I am unable to tell the difference between a PCI slot and a USB port.
Love your videos, specially the Desktop ones
This info is coming from puget systems in the states, there data shows that ryzen CPU’s have a high degradation rate.
Static damage isn’t an issue in the UK because the humidity level is always high enough to avoid significant static buildup (unless we really try). Even in the middle of summer the relative humidity is normally 40% or higher. In other countries with hotter drier climates it may well be more of an issue.
in short: It's wet.
Have you ever had a following issue (combo is 5800X/AsRock B550) - PC came in for intermittently crashing. Changed to a new Nvme with fresh Windows,, installed all te drivers, put the said NVMe in to another slot, RAM (several combos), updated BIOS to the latest stable, changed graphic cards, etc., the computer freezes or restarts. After it became stable for short while I ran an OCCT & Cinenbench R23 & Furmark at the same time for several runs and all is fine (hours). After letting it cool it continued to freeze. Had to change the mobo. What I find strange that it'll crash/freeze/restart under idle, and under stress it works fine.
Happy Holidays to all.
Great video. I am a novice for sure.....but....could there be a bad setting in the BIOS? Load BIOS with standard / basic / default settings? Could this be possible?
REALLY like these kind of talking head videos, great for a listen at work
I’ve experienced “chinese” usb wifi receivers that have prevented systems from booting with a bluescreen error. Happened on three different systems with the same wifi dongle (actually all of them). Ended up narrowing it down to the driver, but just stopped using them.
One thing (imho) worth considering is the similarity to many recent issues that I've seen. While I believe it's related to DMA channel or controller faults, it may not be the actual cause, however similar I believe it to be. Since I have no way of testing this theory, it does appear to be "thematic" of these types of faults.
Not sure that this is very helpful information now. If you do decide to further investigate the fault, wonder if crystal disk or similar might trigger it? I don't have one to test currently. I would be interested in your thoughts.
thanks for sharing your process. very cool :)
It was quite good you managed to troubleshoot that CPU . Sometimes we can be stuck with a strange problem , that goes rabbit hole . I have one of my older gaming rigs on a Asus MB that haves the NVIDIA chipset , running a early intel quadcore Q6600 before the i 3/5/7 era . Never had a problem at Windows 7 (64bits) era. upgraded to win 10 , but ASUS never released drivers for 10 , one of the missing drivers : the asus ATK ACPI . I'w running the NVidia raid with SSD drives in Raid 5 . suddenly , after a windows 10 update, the PC goes beserk : sometimes taking ages to power up (black screen), often getting stuck , rebooting will get it stuck on the BS ... needed to cut the power from the MB to manage to wake it up again (reset didn't work) . when looking to windows events , manage to see that windows didn't managed my C levels from the CPU correctly , and surprise , the ASUS ATK has been replaced by a MS "approved" . I changed the BIOS C Level configurations , memory size/emplacements, still no stable system if sleep/reboot/shutdown . Changed the power supply ... same problem... starting to think the MB was maybe getting a old cond or something , until I installed linux Mint in dual boot and ... Surprise : the PC works flawless from Linux , power down/sleep/reboot ... so even the OS can generate strange rabbit holes . Merry Xmas Graham
I'm still a firm believer that in this case it was degradation of that particular CPU in hand, It could have been a weak sample from the factory to begin with, Even the golden samples of the highest end line could eventually degrade so much that a crucial link between cores or memory could just give out...sure its unlikely but with silicon quality, Heat cycles etc you never know. :) Merry Christmas Adam :)
In the end it was either a defective or degraded CPU. I guess out of warranty. Seems very likely, being an AIO gaming PC, that the user had probably been running it highly overclocked. When it started acting up, the default clocks were probably reset in an attempt to fix the issue. There's a chance the CPU could be stable with underclocking and light use but gaming days are done.
Most of the timings aren’t motherboard dependent. Like 99% of the timings are internal to the mem ic, though tightening the timings do increase the strain on all of the cpu by the fact of it being more data.
Merry Christmas all
I still suggest to spray CPU with contact cleaner and place it 4 or 5 times in the MB. I tend to agree with everything you have state here.
I would be fascinated to know what system you use the most in personal use, what it is and why?
I have a Z80 that has been running since 1981, still running fine, replaced RAM 2 times though ;) (regarding older CPU's)
With spare hardware desktop pc are quite easy to diagnose as you stated. Im currently struggling with a defective 2tb m.2 ssd that freezes the file explorer (not boot drive) in windows, drops out and crashes while trying to copying off files. Smart shows it as health yet it has file integrity errors that CMD also confirms. Currently starting a RMA.
As a builder, I’ve experienced many failed SSD’s and Nvme drives. And when they fail, they go fast……really fast. Ain’t no data recovery happening.
Thank you for all the education. And finally, somebody with some gravitas telling the unwashed that the whole "bottlenecking" thing is basically a non-factor. If the parts don't actually damage each other, who cares?
There's literally always going to be a bottleneck, if there wasn't you'd be getting infinite performance.
In a 5700g there is still the usb and stuff on the chip which is why the x300 “chipset” exists but apparently the desktops chips do have some onboard Ethernet and usb.
The 5700g is just a laptop chip to onto the desktop
People actually thought it was a corrupt file system from these symptoms? LOL. I hate the internet. "Needs defragged" lol. EDI: bottlenecking too, lol. That popped up when, about 10 years ago? That all depends on what load you're running and what resolution. 4k vs 1080p run massively different on the same hardware. I agree, static has been almost a non-issue for a quarter CENTURY. I've seen maybe 2 dead RAM cards. Nothing else. Not a single dead CPU and I never wore a static strap. I just touch the (plugged in, grounded) case when I start. I do use them on the $240,000 computers I work with though, but only when they're open, not just when plugging.
"You need to change your thermal paste" is another favorite meme fix
@rossbittner6831 usually. Sometimes it's true with aftermarket paste due to thermal cycle pumpout. Not true for OEM and the secret stuff I use.
when it comes to static electricity.. linus from ltt and electroboom.. proved that it does not kill your hardware
Yes, this was a good video and it's preferable on this channel. Thanks for sharing your wealth of knowledge ...
🎉 grand video
Graham, on the CPU degradation and Overclock/BIOS reset topics I think you missed something else that I saw no one told you on the past video either.
First, I would like for you to see the video that Wendell did on the Intel degradation on 13th and 14th gen testing.
The reason I point that out is because I think the culprit here is the freaking Gigabyte motherboard. I've seen a lot of Gigabyte and ASUS motherboards 'causing degradation on the field on common CPUs with normal people using them, CPUs that don't get touched EVER for overclocking.
ASUS and Gigabyte give a lot of voltage out of the box with their motherboards, even on Server motherboards for CPUs that should NOT get higher Wattage and voltages than normal for 24/7 use.
See the motherboard that Wendell uses on that video for a CPU that should not run at such a high wattage/voltage as the 14th gen 14900k, because he uses a W tier/spec motherboard and the not overclocking CPU goes sky high with voltage/wattage, poor thing.
Most of the boards from Gigabyte and ASUS since 2020 up until now have such an aggressive curve for voltage, AMD have real good automation for PBO but, and is a big BUT, if the motherboard sends the signal for more watts/voltage, AMD automatic tuning goes out of spec, just as with the Intel CPUs and THEN the degradation starts.
You were TOTALLY RIGHT that a G processor should not degrade or go that high on voltage. But Asus and Gigabyte allow that and push that.
ASRock boards are now the best ones for safe voltage boosting and more reasonable power delivery and AMD/Intel safe specs out of the box.
I think that trying disabling some cores would be good for your own use just for testing, but it wouldn't be a CPU for someone for everyday use, let alone a customer.
Greetings from Chiapas, Mexico. What a lovely dissection video. Happy Christmas.
I have to bring into discussion the fact that 99.9% of motherboards integrators have been overshooting voltages even from AM4 era. I've seen this so many times with multiple integrators that I manually set Vcore on all systems that go thru my hands. BTW the Intel fiasco is a consequnce of the same overshooting done by mobo integrators, like going 1.45 vcore on a 13700kf when the cpu is perfectly stable at 1.16v...
with ryzen the bios needs to be up to date also the mb chips . i recently went through this with my own pc just a couple days ago
Like this stuff, So i can learn from a honest dude like you are
I bought a Ryzen 5 3600 some time ago for 5 USD that was faulty, and it turned out that the pins were corroded. That is, liquid damaged.
So were you be able to clean the corrosion or not?
Ryzen…….earlier Ryzen’s have memory controller issues. I’ve seen quite a few Ryzen 5-3600’s with memory controller failures. Increasing SOC voltage seems to help.
Gut Feelings, based on experience, are not Gut Feelings, they be good Feelings
I enjoyed this , would like to see it again. Thank You...
I never knew you had a podcast! You must advertise for it when you do your videos
Hi there, I have watched a lot of your videos over the years and enjoy watching. I have built many PC's over the years from P4/ AMD3000+ days. I do believe in this case it is CPU degradation. The AMD motherboards do push high voltage no matter the rated power. Having said this I did have something similar to this and it was the RAM needed more voltage. In this case I would push another 0.1V Vcore. I am sure this cpu would then run ok, showing degradation. If this didnt work I would then push another 0.1v on the RAM. I would then reduce voltages until it crashes again. Hope this helps.
I'm sorry Graham but I can't see the point behind this extentded argument; you replaced the CPU and it is working end of discussion. Respect!
It's not so much arguing with the comments, but talking about the ones who might miss the mark, but still raise good points that can be relevant at other times.
Also more technically correct terminology: mm is "millimeters" not "mils" (a mil is a thousandth of an inch)
Where are you based Adam?
This dude has a high percent of crying... when angels die.
Your video are great, quite in-depth, I am quite computer savvy, but your knowledge of the circuits and voltages is way beyond my comprehension in most cases.
22:30 A weak(er) card with a good CPU is not a bottleneck. 1% and .1% are very important in any game in regard to how smooth it feels. The only "complaint" i have of you, is that you refer often to a single person as "they" and that is confusing for non native English speakers like me. You content is great, keep doing what you are doing and i hope in 2025 you will go above 200K subscribers !
18:53 Greg salazar had a few failed ryzen on his channel
I have learned a lot from you. I have a computer I built 4 years ago with the second best Items. Water Cooled NZXT, AMD 9 3800x, Gigabyte x570 Master Motherboard. Gskill Ram, Gigabyte 2090. I have been having Crazy crashes some at startup some at processor high use. Replaced power supply, graphics card, disconnected case switch,, did your mem86 test, windows from a thumb drive, just had Asics symbols on the screen, mounted the processor. I have been messing with this for 3 months now now so I pulled it out and built another computer to replace it. I laid the old computer on it's side and it works fine. Every time I set it up it messes up and crashes at the usual odd times. I have 27 years experience with computers and I have never seen anything like this.
Lord Graham, prince Loek here and i liked the video :)
the one thing you did not do in the video was to have cleaned and replaced the thermal paste on the cpu that is usually the first thing i do when i suspect overheating because of computer freezing, because in over 30 years of pc repair i have almost never seen freezing caused by software, crashing yes but freezing hardly ever, so with your experience i wonder if you went that route because you were making video content?
I too do not get overly concerned about ESD as well, but that being said; In the mid-90s when I was first getting into computers as a profession, I had a job with the local computer shop and we had built 26 new computers for a local business. Latest gen everything. Even had the new high speed CD-ROM drives. Each system had been loaded using the Windows 95 CD disc, so we knew all the CD-ROM drives worked when built. Upon delivery day, I went with the boss to help and was unboxing systems on the floor. I opened the box, rolled it over open side down and lifted the box from it. I laid the box down and pulled the styrofoam and bag from the system and shoved them into the box. I turned to pick up the system and as I placed my hand on top of the mid-tower chassis and felt a static discharge as I touched it. Never gave it a thought and continued working. Once the systems were powered up and operating, that one I'd been shocked on had a failed CD-ROM drive. We swapped the drive and it was fine after that, but that was my first and only experience ever having such an issue. The only time I get concerned about ESD is when the hardware I'm working with is irreplaceable.
Love the Goatee haters. It's lovely in my humble opinion, mine happens to be about the same length. Keep it growing!
I will run red or blue witch has the best value to performance, color is not important for me
Too late for comment, i though customer did a very bad OC, leaving the CPU almost dead. A little chance to keep using it for a while lowering Mhz or giving more volt.
Way too many cheap AIO’s out there and even the good ones are destined to fail. When an air cooler fails, the fan stops working. Easy to see.
Ignore the idiots!
You were doing this in the 90's? A lot harder then but a lot more money. Looking good for your age! 😀
I love your beard.
Bottlenecking is so overblown by most Tech UA-camrs. Some component is always bottlenecking something. That’s why real world Nvme speed is considerably slower than that rated speed on the box. A PC is choke full of bottleneks………..everywhere. QVL’s were important 40 years ago……..not so much anymore.