Thank you so much greg. Im the owner of the pc and can't thank you enough. The pc has been working fine ever since and also thank you for the small upgrade. You won't imagine how big of a headache it was. Edit: Greg I was joking, but keep it pinned it's funny 🤣. I was trying to see if people would believe it. I'm also not the owner.
also could be bios related as he thought he had a 5600 and not the 5600g. asrock i just found out (by trying to upgrade cpu) has different bios for the apu's, compared to regular cpu's.
@@craiglortie8483 5600G and 5700G use the same bios, but yes you're right in that the APU architecture is actually a different (previous) generation. A 5600 might need a different bios. For that matter switching from 5600 to 5600X might need an updated bios as it was released later. I had this exact issue.
@@Lukiel666 i went from 1600 to the 5600, until the cpu fell and bent the pins... smh. 3-5 builds and never dropped a cpu, this time it flipped out after the holder caught on the box when i pulled it out. going from the 1600 to the 5600 series required serveral bios updates, had to look up each series by it's "arch name" and realized the apu's were different from regular cpu's. at least on asrock's website.
happens to the best of us, pretty much why i got sacked from Teleplan back in 2002/03, Laptop kept coming back and id throw parts at it for some reason it just didnt wanna stay fixed, i got shirty on the feedback form and got booted for it.
Great comment. One with which all introspective and authentic people can relate. I have had my share of curious and seemingly unexplainable events with computers. Quite frankly, when something works some of the time but not others, I simply say that it is almost a miracle that computers work at all. Minerals and metals mined and assembled from all over the world into chips and parts from all over the world, only to run software from all over the world. It is nothing short of a miracle.
This ep is exactly why I like this series. While it would be impossible for you to encounter every issue, these videos are just a valuable tool for PC builders to help in the troubleshooting process.
Greg..I have been a PC/MAC tech over 29 years. I have had these exact scenarios happen. It came to the point of MAJOR self doubt, incompetence. Had fellow tech check my findings with identical results. Boils down to this. Would you rather be right or happy? Sometimes you just need to smile...nod your head and let go.
I LOVE that you post the nitty gritty and the unexplainable. It happens, and it happens pretty often, and it sucks every time. I'm in a helpdesk IT position and sometimes when people get frustrated when I can't succinctly explain why something is fixed, I get frustrated myself, and **at** myself. Videos like this where people like you who are much more experienced than myself still have these issues pop up is pretty helpful to see.
It would explain it suddenly working on a different MB and RAM if the ram was a lower base spec as it would be less demanding of the memory controller. Even the MB can make a small difference based on how the traces are laid out, from physical length to topology, not to mention 1 vs 2 dims depends on how the memory controller works it's likely having to use more current to support two dims vs one.
Honestly I am beginning to suspect that the Ryzen 3000 and 5000 series CPUs have a serious issue with the memory controller, I have had four that developed memory issues, a 3100, a 3600X and a 5800X. At least they aren't likely to give me issues like those Intel CPUs.
@@kaseyboles30 it probably has more to do with voltage control of the RAM resulting in an unstable system. Also new AMD bios dont play that well older AMD APUs. I guess this is where AMD fails by keeping their AM4 socket for a long time. Also the old APU probably cant handle the timings of new DDR4 RAM. Keepin mind AM4 APUs are actually one generation behind.
@@cecilb7927 its actually known that am4 has some bizarre memory controller issues and even hardware memory controller failures. but again as i said its what you get when you support a socket for so long and have to support so many different RAM kits. you are going to run into issues eventually.
Poor Greg, always trying to explain and show his sincerely thoughts to viewers but as you said, it's ok to say "I don't know". Huge huge fan of your community work and a huge respect for your transparency.
Being vulnerable shows you are a strong person to let people see that side of you. So continue being yourself. But that definitely was funny and frustrating at the same time.. Love the series.❤
Negative tolerances accumulate. Components that work separately but won't play nicely as a group are a thing. Each is in spec, but those combine in a way that breaks the machine. It can occur with new parts, or as components age. And, it can be maddening to sort out.
I love that you are always honest. It helps the rest of us see that what we experience is completely normal. When I run into issues like this, it reminds me, that even the great Greg Salazar has this kind of day, haha. keep up the great work.
I suspect that there's probably a dry joint on the motherboard somewhere and just by twisting or moving a wire might make contact, I would gently tap the rig a few times and then turn it on, not just once but quite a few times and see if the problem comes back.
Good call. I had this happen to my daughter's computer. Every once and a while it wouldn't power up. I'd reinsert everything some number of times, and it would start working for a few months. Rinse and repeat 5 or 6 times, and then it finally gave up the ghost.
It's actually quite nice to see your candid reactions with how frustrated you were at the inconsistencies. I work in IT myself (specifically the programming side of things) and I can't count the number of times I've tested/written software where bugs would intermittently appear seemingly without reason. Makes one feel like they're going crazy sometimes.
A big part of why I enjoy your videos so much is the unscripted and honest way you work on PCs. We all know it can be a frustrating process at times but it's always educational which is what makes the fix or flop series one I never miss an episode of.
5:35 these definitely give the most confusion in electronic, i work on cellphones and phones will just randomly not work then randomly start working... Even worse when the customer is like "whats wrong with it"
Don’t know if you did this, but an BIOS flash can do wonders to RAM related issues. I got the exact same problems with an AM4 rig and dual channel didn’t even work on a certain CPU. A BIOS flash/update fixed it. PCs are so weirds nowadays…
@@GregSalazar Differents SKUs can behave differently even with the exact hardware. Though it’s is indeed pretty strange as it’s the same architecture. But in such cases, it’s best to try everything at this point even if you assume is it not even remotely possible to be the culprit
Professional PC repairman here, you're possibly looking at a corrupt bios. I would recommend reflashing it next time. If this was a physical problem, the heat cycling of the motherboard would have yielded a failure by now.
@@AdmrlSpeedy BIOSes can be corrupt/wacky in some ways and still boot. I’ve been recently trying to overclock my RAM and after some failures to train or even shorting the RESETCMOS jumpers while the motherboard is still being powered by my PSU corrupted my integrated TPM. It could not turn on TPM again as if there were never one. I flashed the BIOS and it came back
this was my exact thought. from other unexplainable issues solved with bios flashing seen on other youtube tech guy channels. But Graham Lord on Adamant IT has a bios flashing tool where you take the bios chip off the motherboard, so he has advantages.
@@AdmrlSpeedy yeah and firmware TPM is software, so prone to bugs since it’s handled by BIOS which is software too btw, and most Gaming PCs are using fTPM
Greg, don't stress to hard. My son's system crashed last night, when playing BO6. He tried to reboot and it sat at the loading screen, then he came and got me. After checking everything out, and all tested/looked good, I decided to do a fresh win11 install. Got it installed, but, shortly after, it crashed again. He was installing games on the 4TB HDD. So we go through the whole process again with same outcome. Disconnected the HDD and everything worked perfectly. Swapped his 4TB for one that I had in my computer. His system has been working great since. I put his HDD in my system and I'm not having any problems what so ever.
Was the HDD connected when installing Windows? Sometimes Windows decides to put important OS files onto the HDD which causes unusual problems and sluggish performance.
@@gcmanuel85x When some AAA games are taking 500gb nowadays its not cost effective to put everything onto an SSD. I only put my most played games on 2TB SSD and the rest on HDD.
I think it is refreshing that you are so candid with us and don't hide the mistakes you may make it has helped me learn a lot and also helped me to fix a problem I had with my system so I thank you very much and look forward to the next video and what I may learn from you.
Your candor is a very good reason to follow this series. I've been playing (and working) with computers for near 50 years. I've seen some really strange problems both with HW and software. Both can drive you bananas.
From my time as a PC repair technician and pc builder problems like this can be caused by physical damage. Usually a bit of damage you cant see. So a broken trace inside of the PCB or a cracked solder joint. Heck there was a pc that we swapped 4 CPUs into because it would act weird but only when it was REEEEALY hot. (not the parts... the ambient temperature) We finally swapped the mobo and it was fine. So I went over the MOBO with a magnifying glass and as far as I could see the MOBO was perfect. Cept for OOOONE LIIIIITTLE thing. There was 1 pin in the MOBO socket that was 2 hairs width lower than the other pins. So basically it was bent but not enough to be seen with the naked eye... What I'm guessing is that it wouldn't make proper contact. But only when it was hot... Meaning the problem was on a MICROSCOPIC level. To where a 10 c difference could affect the metal jussst enough to not make contact. Switching cpus would fic the issue for a day or 2... then it would start back up with the powering off in the middle of intense workloads. So what would happen is that mounting pressure could make it cut off during a stress test... Bumping into the case.... Hell bumping the table leg with my foot set it off once. And this is what led me down the "it could be physical damage" route.
Back in 2000 when I was studying. I got "broken" MB. When you put it in case it stopped booting and without case it booted OK. Problem was one broken solder join in bios chip socket. Basically when you screw that on case it bend that socket little it and that pin didn't have contact anymore. Just solder that pin and I got free working MB for my rig :)
Broken traces or a CPU socket with destroyed pins. I've had experience with both. I currently have an Asus ROG Z390 Maximus XI Extreme that I had gotten a larger case for (this board is EATX). In the process of moving the board I somehow bent multiple pins causing memory issues. I can use some slots but not all and for a single stick I have to put memory in slot that is not recommended!
Keep the vulnerability and honesty man. Imo that is a huge part in your videos that sets you apart from other PC ytubers, and it feels more authentic to watch your videos for me. Been watching you for almost 3 years at this point.
I hate when these issues happen, it makes you feel like there is no answer. BIOS updates I've had fixed this issue for me before after doing all the same tests, but that still doesnt fully explain it (unless the implementation of the BIOS revision for the brand of the motherboard had a strange bug for a certain CPU?) but i have no clue. Ive really enjoyed these videos, keep it up!! Ive been watching these since the beginning, i just don't normally comment.
watching this series has made me so passionate about building/fixing systems so finally bit the bullet and started to look for business so appreciate you man great series
Hey Greg, don't let it get you down. I've been building or upgrading computers since 1990 and I still run across things I absolutely cannot explain. None of this is 100% perfect 100% of the time or we wouldn't talk about 'the silicon lottery' when it comes to hardware inconsistencies. It's also okay to be transparent about your feelings, because those of us who have been there can relate, and those of us who don't have your skill absolutely know what it feels like to be in that situation. Thanks for these, Fixes and Flops alike, and please keep doing what you do knowing we love to be right there with you on the journey.
Seen that issue before and for me the b550 board for me had bios issue which was fix with an bios update the board you put the CPU in might have an updated bios which made it work
@soccerguy2433 yea It would there was a time on some b550 board and I don't know what but it something to do with bios revision couldn't tell what probably a bug where it just freak out
@@michaelpearce8371 A 5700G (same family chip, same architecture) also worked just fine in the same board with the same BIOS revision. I wouldn't be confident giving a PC back to its owner if a chip has already once randomly decided not to play ball with a compatible BIOS revision/AGESA update.
Greg you being real or authentic shows how it really is sometimes working on PC's. And that sometimes there isn't a reasonable explanation for some things.
Since AMD came out with Ryzen cpus ive had the Same problems... just bizarre problems. From a 2600 through to 9700x . b550/x570/b650 and 3-4 types of memory. Bios updates did nothing, cpu would work with specific combos of gear, regardless of them being on the QVL or not. I had th throw the parts cannon on systems to figure out the Right combo to get them to boot and boot Consistently. Happy to see its not Just me dealing with this amd bizarro world issues. I still dont get it. Even had issues with Radeon cards and Nvidia cards as well. One day hopefully we will get some answers? The only thing I can think of being the problem is the memory controller? I may e wrong but thats my guess. Thanks for sharing this with us mate. Cheers.
Another treat, and it's not even Christmas! Keep it going I always watch them 100%, it's always fun to follow along. I thought I knew it all, and learnt tons thanks to you Greg!
Hey! I just ran through a very similar situation when upgrading my Ryzen 5 2600 on an MSI B450 Tomahawk to a Ryzen 7 5700x3D. I went through all trouble shooting steps and turned out one of my RAM sticks had given out. Very frustrating but even more satisfying to discover and fix. Great video
Seeing your honest reactions to these issues is what makes this playlist awesome. Yes, even if you really can’t figure out why something happens. We’re all here for the ride; bumps and all.
Hi Greg, believe me when I say that hardware diagnostics can be a real pain, but when you keep going with the diagnosis, finding the root cause is so rewarding. I have had a couple trouble systems over the years since 2001 that took me days to resolve. I had a Dell server that just keeps on randomly freezing up, I ended up tearing it all the way down to individual components, and found a speck of dirt on a cpu pin. Cleaned it up, and it has been rock solid since. I had a desktop that would randomly just shut off, went through everything, and could not explain it after a few days. Finally the motherboard took a dump and quit powering up, replaced the motherboard, and it would randomly freeze up, found that one of the memory sticks was bad. After memory swap, the system ran rock solid afterwards. I built a new custom built pc in 2004, one day there was smoke in the room where it was at, looked it over and the front panel cables melted down while the computer was still running, nothing went bad besides the cables and the cables were not shorted against anything. It is still unexplainable today, never had a system melt down the front panel cables since then; completely baffled as to what happened.
Many of the cheaper SATA/NVMe heatsinks come with rubber bands to hold them on. They're usually black, though, and I THINK the ones I've gotten were silicone instead of rubber. I've also gotten kits with clear bands that were obviously silicone. Silicone will definitely hold up better over time. The ones on their drive look like they're rubber, because the one nearest the connector pins is already dry-rotting. They should probably replace it or at least swap it to the middle position.
Just two days ago I was building my iTX am5 system and while doing this I messed up front panel connectors on a mobo that was on outlet. I purposely didn't do a bench test so my imidiate thoughts were "crap, I bought a dud" and then a revelation "dude, you've been watching fix or flop for many months, get your sh1t together and rethink what was done". My mistake, no doubt, but I wish ASRock does not put usb and front panel prongs side by side ever again. EDIT: As for the 5600G it seems that at 7:44 there is a ever so slight smudge of thermal paste near two pins at the top of the shot. A bit far fetched but plausible.
Could be a wonky bios, a dying memory controller, faulty/out of spec motherboard vrm... or a combination of the three. I wouldn't get too flustered about it.
Im honestly just as stumped My initial thought was either a dead/dying Memory Channel, then perhaps a Mobo on the way out, but since neither of those panned out......I honestly got nothing Stil a great video, just....very puzzling
It still can be dying memory channel but other MB has better tolerances on memory channel so it will work on that MB some time until it's in same state again.
I didn't read all the comments, but my guess is that a tiny bit of gunk (thermal paste, oil from manufacturing, a hair, etc) was stuck on one of the CPU pins or socket preventing contact of that pin(s). As the CPU was removed and placed back in, the gunk was moved from friction, or fell off with motion, and it started working again. Great work Greg! I always appreciate how you show how frustrating it can be to work on computers sometimes
This exact thing happened to me recently, removed CPU cooler to replace paste. CPU came out of socket along with the cooler (apparently it happens to am4 CPUs somewhat regularly) and the system wouldnt boot after reinstalling the CPU. I thought there must be bent pins from the CPU coming out of the socket in an unintended way but visually I couldn't see any damage. System would boot with a single stick of ram in A2 slot. I removed and reinstalled the CPU several times trying to get it to boot and then noticed that some thermal paste had gotten on to the socket when the CPU came out attached to the cooler. I brushed it off, reinstalled the CPU and now the computer behaves as it did before the entire fiasco.
One thing that's been bugging me about these videos - what denotes a "season?" Greg does these videos throughout the year; he doesn't stop doing them over the summer or whatever. It's like when Rhett and Link say they're on season 18 or whatever but they have videos consistently (even if they might do a week here or there with less videos or with compilations). Or how WWE will randomly call an episode of Raw the "season premiere" when they've been on every week for like... 30 years.
I believe he's said in the past that he separates every 20 episodes into a season. I don't remember why specifically, maybe easier to categorize or keep track of that way, maybe just his own preference.
Seasons is just a way to organize episodes and break them down into more manageable chunk. It is arbitrary and mostly depends on the producers or creators on how they want to organize their show. In this case, if you look at the fix or flop playlist, you'll see that each season is broken down to 20 episodes per season. Ppl aren't gonna remember exact episode numbers all the time, especially when you can have a long running series like Simpson or in your example WWE, even in anime ppl will refer to arcs instead of individual episodes. Short answer. To make things more manageable.
So great to see new content. I learned how to deal with a "no picture out" problem due to your videos. Not only are your videos entertaining but they are educational as well. Bless you.
Greg I have fixed... computers for over 30 years, trust me, you will always find problems even strange ones, What you do is amazing helping others, i will be getting back in to computers in 2025 as I need to keep my mind busy.
10:31 , Greg , owner's motherboard had P1.30 bios which predates 5600G support . 5000G series support starts with version P1.50 , you can check it out for youself at B550M-C support page. This is a rootcause of inconsistent behavior. Some boards can act weird and initialize newer CPU's without flasing BIOS to newer version . And you said that B550 boards come with native 5000 series CPU support but it's not true , they were released alongside Ryzen 3000 series. Sorry to disappoint you , but you didn't fix the problem , and owner will likely experience same issue he had after a while . Anyway , thank you for great content ! Keep it up !
"And you said that B550 boards come with native 5000 series CPU support but it's not true" Greg said "the boards support it natively, at least with the BIOS revisions that are on these boards". Greg was correct.
@@sergiobisonte Bios version P1.50 in it's discription literally states following "To support Ryzen 5000 G-series processors, it requires to update the BIOS with Matisse, Renoir or Vermeer CPU" in yellow text . In this particulsr case CPU support list is total BS .
One thing I found with AMD boards is that the pins in the AM3 socket can sometimes make an intermittent contact if the pins under the plastics has been damaged even though the plastics look clean. If the CPU was taken out previously improperly it may have damaged the pins underneath. Or it could have been a corrupted BIOS
As someone who used to repair PCs between 2002 and 2014, what you're seeing happens. Some parts and combinations have issues while others don't. Changing one piece of hardware is all it takes. There have been times I've replaced IDE cables and that solved boot and stability issues. Or I remember a issue where I swapped a CPU for an identical model and the fixed system worked fine and the old CPU worked perfectly fine in another system. Sometimes you encounter the right combination of hardware that causes an issue.
@@GregSalazar also could be bios related as he thought he had a 5600 and not the 5600g. asrock i just found out (by trying to upgrade cpu) has different bios for the apu's, compared to regular cpu's.
I've owned a r7 3700x that I couldn't get working with 2 motherbords and memory kits. The last one was a MSI bord where you could try a downclocked memory profile. It now is stable, been working fine for 5 years! They are voodoo!
Faced a similar issue on asus b550f wifi. I was trying to add 2 more of identical memory sticks. The board refused to retrain memory on its own. Only a bios clear retrained the memory. I am guessing it was due to an old bios version. I tend to defer updates if everything is working.
they often come with the very cheap m.2 heatsinks you find on ebay/aliexpress/amazon. for an extra £1 you can get a much better quality heatsink with a proper mounting mechanism like one from Jeyi...
When an IT specialist in the Netherlands doesn't understand what is happening in a PC, they pull the "Fuzzy Logic" card. As if this helps.... It is just someone being not able to give a normal reaction to a customer and try to get away with it. You being onest of being confuzed is a breath of fresh air. I'm still learning with you, troubleshooting and solving problems.
This comment is critical, but it is intended with all possible kindness. Not being able to explain or at least narrow down a failure often comes down to having too many variables still part of the system, and too many unchecked assumptions. You could say there's two fundamentally opposite approaches to fault analysis: top-down and bottom-up. Top-down being, you take the complete thing you're analyzing, and using your intuition and knowledge of common faults, you change a variable here or a variable there, and you try to discover the fault. That will find common and intuitive faults the fastest. But since every variable of the system is still in play, rare and unintuitive faults will be nigh impossible to track down that way, the permutations are too numerous, and it ends up being blind guesswork until you get tired and frustrated that you can't guess the right one. The bottom-up approach is the opposite: tear a system down to its smallest possible component, validate that one thing works, re-introduce one other component, validate it still works, and so on, until you add back in the component where the fault reveals itself to you. This takes longer for common cases, but with faults that defy intuition, there is simply no alternative to it. It replaces the guesswork with pure experimentation. The biggest mistake I see people trying to troubleshoot a problematic PC is assumptions. Don't assume the power supply isn't the problem just because it passes a test. Don't assume that 3 seemingly perfectly working parts can't result in a fault once combined. Don't assume that something as mundane as a case fan or a SATA cable or a tweaked CPU fan wire or the damn case itself couldn't be a contributor to a system-halting fault, even if it seems there's no logical reason for it to be so. Watching this series, you sometimes take baby steps towards a bottom-up approach (like testing the CPU on the "known good" platform in this video), but you have often voiced your resistance to doing more substantial tear-down. And (again, these comments are meant with kindness, not malice) I believe this fundamentally limits your chances of successfully identifying more difficult-to-find faults. I understand of course that time is one of the biggest constraining factors here, and that tearing every system down to parts may not be feasible (though I do wonder how many times a troubleshooting attempt has taken significantly longer than a tear-down and incremental rebuild would have). But the simple act of removing the platform from the case drastically reduces the number of variables in play. And I guess that would be the recommendation that I want to voice: once normal troubleshooting fails, assume that the fault is something you're NOT going to be able to intuit, and start slashing the number of variables in play. Some faults are simply beyond intuition, and there's no way to track them down except for shrinking the problem space until there's nowhere left for them to hide.
What I've found in my journey in repair is cleaning is king when you have an intermittent fault with a PC. I always clean each edge connector twice with isopropyl and a fresh cloth and wash down the pins of the CPU with PCB cleaning solvent. I don't often need to clean the sockets but you will need plastic safe solvent if you do. It reduces the amount of situations where a part will work now and then or in different systems significantly, it was quite a dramatic night and day change for me. I've had good luck just cleaning dust off mainboards solving similar issues as this. Aside from the repairs I often get scrap systems given to me that other shops or organisations have given up on, I'd say over 80% work fine with original parts and some cleaning. Swapping the power supply can also have dramatic and unexpected changes like gains in 100fps even though it had great ripple values and delivered the power requested. Said PC was 100% stable at full load with minimal 1% lows, the only symptom of it failing was a slightly lower framerate.
hey greg if u are reading this, thank u so much for teaching me bout PC trouble shooting and in-depth stuffs about PCs. this playlist helped me learn a lot bout PC stuffs.
I recently bought a 800w power supply second hand, it claimed to be brand new and was sealed when I received it. But when under heavy load it makes a coil whining noise. Did I get scammed or can this just happen out of the box
A possibility is "tolerance" or "tolerance stacking". Tolerance as in spec +/- percentage . Theorizing an example for this case the 5600G needs just a little bit more voltage on a certain pin. The test bed MB provides it and the system MB does not. Any of the three can be in or out of spec, test MB slightly high voltage, system MB slightly low voltage or the CPU needing more voltage than spec as an example. It doesn't have to be voltage, could be timings or frequencies, latencies, etc.
Typically, base tolerances for these chips are _vastly_ lower than what the motherboard provides at stock conditions. This is, of course, by design. So if, in fact, the 5600G demanded a minimum influx of [insert variable here] that exceeded the board's often very liberal supply of [insert variable here], then a CPU replacement was warranted. Our 5700G (a more powerful version of the 5600G and relying on the exact same architecture) worked perfectly fine under identical conditions with the same board. In essence, this was why I came to the conclusion I did by replacing the CPU.
@@GregSalazar After a lifetime of fixing computers I've come up with a saying of "there's a logical reason for everything a computer does and sometimes you just can't find what that is". :) I would done the same thing as you in this case. You do good work, keep it up.
@@GregSalazar I think silicon aging is playing part of this. So other MB have more favorable tolerances so it may work few more months. Almost reminds me like dying memcontroller like my 9900k case. Stopped working with XMP 3200 then and when run with JEDEC worked 3 more months until total dead. (Basically XMP overclocked memcontroller (max 2900 MHz) and speed up aging process) Got new CPU from RMA and set ram to 2900 and it's still in use. Donated platform to my niece when I upgraded to i9-12900k.
I haven't watched this yet.. but the main thing that I keep seeing is the 4pin (or more) CPU power plug that is commonly an issue. Especially if you're overclocking.. I've seen those plastic connectors get brittle and break and have bad connections.. On mine, the pins were actually burnt or charred looking from arcing over time.. and then I realized that one of the pins in the board was loose and moving around.. After pulling the board and resoldering the pins and also soldering on a new connector to the wires.. It has been flawless ever since. This issue was a common find online too. Often times it will even melt the connectors.
Another nice FoF video. Your services really helps the pc community as a while and i personally believe there should be some people who should take example from you. (Bring back the gt710!)
Yup that’s a good one. I wonder if the bios version between the two motherboards has something to do with it. Newer revisions tend to get more memory stability fixes. Just a thought
Having serviced computers for way to long I understand you. All I can say is that there are a few basic steps you always try first. Clear the CMOS and test. Flash the BIOS with the latest version available and test again after another CMOPS clear. Minimize the system. Connect a known working PSU and try it. Swap out the GPU to a known working one, make it as simple as possible. it's a bonus if it doesn't need a power cable. Test again. Remove everything not needed for it to post and test. Swap in some known good working memory. Now we get to the interesting parts. I would suggest testing the processor on the cheapest and simplest known good MB you have. Now if that works then the motherboard is most likely the culprit. Having a lot of memory, processors, graphics cards and motherboards is a real help, and unfortunately it's not something everyone has access to. Trying to fix something like this when all you have is the one computer that has failed can be close to impossible. But some of the early steps anyone can test. Clear CMOS doesn't cost anything and if the motherboard has the option for programming the BIOS without a CPU or memory installed then that's also an option. If the processor has a integrated GPU and the motherboard has ports for it you can ditch the graphics card and see if that changes anything. You can also try with only one memory stick and also try swapping the sockets. Don't miss testing both memory sticks, as any computer really should have two of them. One bad memory can be enough for the computer to fail POST. Oh, and POST is short for Power On Self Test. It's the first procedure that the BIOS has and will be run before it passes control to the boot code for the OS. Just forgot that not everyone is speaking basic computer lingo.
When I tell you me and my friend have experienced these issues constantly, whether it's on builds we were flipping or our own personal builds. We've seen this happen between CPUs and RAM kits on different motherboards. I could rant about this in a 10 minute video if you like (a viewers experience POV). Additionally I've found AMD CPU compatibility to be so ass between different motherboards, to the point where you have to configure it in bios a specific way to make it work...
Oh ya brother of mine. Memory compatibility in the modern age. Even main board vendors don't have time to vet a fraction of DDR4 , 5 & CUDIMM. Memory remains expensive for all the wrong reasons. Read my post sidgySOHO_1960
Going waaaaaaay out on a limb with this one: maybe there was something interfering with the pins in the socket like oxidation or just a piece of schmutz? Wouldn't be easily visible, and would be corrected by unseating and reseating a new CPU. I've seen the same kind of thing with vacuum tubes. Like I said, out on a limb but it seems plausible if nothing else.
Being a G model chip, initial thoughts would be that it wasn't initialized correctly to run a dedicated GPU vs the on-die version, but that's very curious that it works on another board, and a new chip works on the owner board. When you get down to these seemingly granular issues between hardware ghosts and programming, it turns into a maddening roadblock that goes from one of a few things to one conflict behind a wall that you need access to in windows to sort out and said conflict prevents you from going further to get to the solution.
Mounting pressure, or a cracked trace. Happened to me with a B350, stock cooler couldn't post the PC unless you untight the screws slightly. Other coolers worked fine but with a different spring mechanism. 2nd cooler uses clip mechanism, 1st cooler uses 4 springs screws
The only thing I can think of is that it might be to do with the original owners RAMs timings, because it is the same speed not sure about the timings. Or a bios update? idk truly something you can't explain.
I would say that either the ram controller in the cpu is dying OR, there is a microfracture in the chip (both things can cause very similar conditions and are VERY hard to resolve/figure out) on a microfracture just a little moving in the wrong way, or heat expansion can cause the internal circuitry lines to disconnect.
I've had similar situations. One I can remember was a friend's GPU stopped working in his PC and his PC only. It worked perfectly fine in the other 5 or 6 PCs/test benches I tested it in, and all other GPUs worked in his PC. Any time I put his original GPU back into his PC, there would just be no picture at all. The PC was actually POSTing but there was no display. He ended up trading someone for the same exact model and both worked exactly as they should in each other's PCs lol
at 7:43 there is a smidge at the top of the CPU pins, it's barely noticeable but visible nevertheless. Could be the cause of the issue, especially if the motherboard also contained a smidge at the same place as well.
Check the output from the onboard hdmi port. I have a 5700G and it will sometimes dump bios/uefi errors to the onboard display. I had to use 3 monitors to figure this out. One on onboard hdmi, one on video card hdmi and one on video card display port.
Hi, please do not apologise I work in IT and I get but it was working yesterday4 never feel the need to shy away i have been watching you for 2 years and i keep coming back because you are you!
PCs can be very frustrating, especially when you have a problem you just can't pinpoint. I was having random freezes and went through so much stuff even reinstalling windows... only to find out the board was having issues. Ended up replacing it with the same GIGABYTE one I've seen you use in a few of these episodes, seeing as none of those PCs reappeared I figured it was a safe bet.
Wooo... I've seen that setup before. I can say with confidence this is a Skytech build as the single fan 3060, the Geil ram sticks and the use of a 5600g. I'd be surprised if it wasnt one of theirs. That being said, good work Greg. Sometimes, it's just something that is finicky but I would say it's likely that CPU.
My guess is maybe bad contact in the socket. If it were brought to me the first thing I would have tried was reseating the CPU and making sure to give it a good jiggle in the socket before pulling down the tension arm, I might have even tried contact cleaner. My other guesses for what the issue could be are a marginal memory controller which makes the problem intermittent.
My theory is a very small amount of thermal paste or something got on one of the pins of the CPU or on one of the ram chips and this contamination was then transferred into one of holes of the socket and when you changed the CPU that moved things around in that hole and this pin or pad was able to make contact again. When you get strange behaviors like these, it can be debris in the motherboard socket or in the ram sockets. After you have built like 1000 or so PC's and work in computer repair for over 20 years, you start to see every scenario possible, but I like your honesty man and I have defiantly had a few over the years that defied explanation like this. Here's one, you have a system that wont post, you check the ram and it's seated properly, you take out the ram sticks run the computer with no ram and get the expected ram error for the missing ram, reinstall the ram and the computer then works?? The only explanation I have is that debris was in the ram socket or on one of the sticks, and by taking them out and moving things around it either came off the stick, out of of the socket, or just moved out of the way enough to make contact again..
Greg, in this instance, I would have done an update to the BIOS. I have seen this before where a board has native support for a CPU, but doesn't work, but putting in another CPU works. Maybe the Microcode in the BIOS got corrupted and it just didn't like that CPU. My ASUS ROG Strix X570-E Gaming had native support for Ryzen 3000 at release, but yet my 3800x didn't work. My Ryzen 5 2600 did work though. I upgraded the BIOS to the latest version and put the 3800x back in and it worked. I know the 3800x worked because I tested it in my other PC which was a Gigabyte Aorus X570 Pro, and the CPU worked just fine. Maybe something to try if you see this issue pop up on any future Fix or Flop episodes.
Thank you so much greg. Im the owner of the pc and can't thank you enough. The pc has been working fine ever since and also thank you for the small upgrade. You won't imagine how big of a headache it was.
Edit: Greg I was joking, but keep it pinned it's funny 🤣. I was trying to see if people would believe it. I'm also not the owner.
nice pfp
also could be bios related as he thought he had a 5600 and not the 5600g. asrock i just found out (by trying to upgrade cpu) has different bios for the apu's, compared to regular cpu's.
@@craiglortie8483 5600G and 5700G use the same bios, but yes you're right in that the APU architecture is actually a different (previous) generation.
A 5600 might need a different bios. For that matter switching from 5600 to 5600X might need an updated bios as it was released later. I had this exact issue.
@@Lukiel666 i went from 1600 to the 5600, until the cpu fell and bent the pins... smh. 3-5 builds and never dropped a cpu, this time it flipped out after the holder caught on the box when i pulled it out. going from the 1600 to the 5600 series required serveral bios updates, had to look up each series by it's "arch name" and realized the apu's were different from regular cpu's. at least on asrock's website.
Lying is funny?
Had to fix one of my PCs the other day. Thought to myself.. WWGD. So I looked at my cat and said "This here is a viewer's broken gaming PC.."
What did the cat do? . . Be Honest.
cat : “what are you talking about now…” 😂
Stop diddling my cat
😂
And then all you had to do was turn on the power strip.
Greg, being vulnerable and transparent shows authenticity, which is a good thing for viewers. Thanks For being you!!
happens to the best of us, pretty much why i got sacked from Teleplan back in 2002/03, Laptop kept coming back and id throw parts at it for some reason it just didnt wanna stay fixed, i got shirty on the feedback form and got booted for it.
it always gives inauthentic to me not sure why i fear love his videos though
Great comment. One with which all introspective and authentic people can relate. I have had my share of curious and seemingly unexplainable events with computers. Quite frankly, when something works some of the time but not others, I simply say that it is almost a miracle that computers work at all. Minerals and metals mined and assembled from all over the world into chips and parts from all over the world, only to run software from all over the world. It is nothing short of a miracle.
This ep is exactly why I like this series. While it would be impossible for you to encounter every issue, these videos are just a valuable tool for PC builders to help in the troubleshooting process.
Greg..I have been a PC/MAC tech over 29 years.
I have had these exact scenarios happen.
It came to the point of MAJOR self doubt, incompetence. Had fellow tech check my findings with identical results.
Boils down to this. Would you rather be right or happy? Sometimes you just need to smile...nod your head and let go.
I LOVE that you post the nitty gritty and the unexplainable. It happens, and it happens pretty often, and it sucks every time. I'm in a helpdesk IT position and sometimes when people get frustrated when I can't succinctly explain why something is fixed, I get frustrated myself, and **at** myself. Videos like this where people like you who are much more experienced than myself still have these issues pop up is pretty helpful to see.
Just wanted to say, this is by far my favorite RUclips series! I do the same work, and find this highly entertaining, and informative. 😁
Thank you! I appreciate that.
My guess is the 5600g has a dying memory controller
It would explain it suddenly working on a different MB and RAM if the ram was a lower base spec as it would be less demanding of the memory controller. Even the MB can make a small difference based on how the traces are laid out, from physical length to topology, not to mention 1 vs 2 dims depends on how the memory controller works it's likely having to use more current to support two dims vs one.
Honestly I am beginning to suspect that the Ryzen 3000 and 5000 series CPUs have a serious issue with the memory controller, I have had four that developed memory issues, a 3100, a 3600X and a 5800X. At least they aren't likely to give me issues like those Intel CPUs.
@@kaseyboles30 it probably has more to do with voltage control of the RAM resulting in an unstable system. Also new AMD bios dont play that well older AMD APUs. I guess this is where AMD fails by keeping their AM4 socket for a long time. Also the old APU probably cant handle the timings of new DDR4 RAM. Keepin mind AM4 APUs are actually one generation behind.
@@cecilb7927 its actually known that am4 has some bizarre memory controller issues and even hardware memory controller failures. but again as i said its what you get when you support a socket for so long and have to support so many different RAM kits. you are going to run into issues eventually.
@@turtlefrog369 Valid suspicion. Hard to say without a full dx, which isn't practical.
Poor Greg, always trying to explain and show his sincerely thoughts to viewers but as you said, it's ok to say "I don't know".
Huge huge fan of your community work and a huge respect for your transparency.
Being vulnerable shows you are a strong person to let people see that side of you. So continue being yourself. But that definitely was funny and frustrating at the same time.. Love the series.❤
Negative tolerances accumulate. Components that work separately but won't play nicely as a group are a thing. Each is in spec, but those combine in a way that breaks the machine. It can occur with new parts, or as components age. And, it can be maddening to sort out.
I love that you are always honest. It helps the rest of us see that what we experience is completely normal. When I run into issues like this, it reminds me, that even the great Greg Salazar has this kind of day, haha. keep up the great work.
I suspect that there's probably a dry joint on the motherboard somewhere and just by twisting or moving a wire might make contact, I would gently tap the rig a few times and then turn it on, not just once but quite a few times and see if the problem comes back.
Good call. I had this happen to my daughter's computer. Every once and a while it wouldn't power up. I'd reinsert everything some number of times, and it would start working for a few months. Rinse and repeat 5 or 6 times, and then it finally gave up the ghost.
It's actually quite nice to see your candid reactions with how frustrated you were at the inconsistencies. I work in IT myself (specifically the programming side of things) and I can't count the number of times I've tested/written software where bugs would intermittently appear seemingly without reason. Makes one feel like they're going crazy sometimes.
A big part of why I enjoy your videos so much is the unscripted and honest way you work on PCs. We all know it can be a frustrating process at times but it's always educational which is what makes the fix or flop series one I never miss an episode of.
5:35 these definitely give the most confusion in electronic, i work on cellphones and phones will just randomly not work then randomly start working... Even worse when the customer is like "whats wrong with it"
Could it be a bios issue?
"Did you try rebooting it ?"
Don’t know if you did this, but an BIOS flash can do wonders to RAM related issues.
I got the exact same problems with an AM4 rig and dual channel didn’t even work on a certain CPU. A BIOS flash/update fixed it.
PCs are so weirds nowadays…
I just would wonder then why a 5700G worked in the same rig with the same board, same BIOS, and same RAM.
@@GregSalazar Differents SKUs can behave differently even with the exact hardware. Though it’s is indeed pretty strange as it’s the same architecture.
But in such cases, it’s best to try everything at this point even if you assume is it not even remotely possible to be the culprit
Professional PC repairman here, you're possibly looking at a corrupt bios. I would recommend reflashing it next time. If this was a physical problem, the heat cycling of the motherboard would have yielded a failure by now.
Lol you're not very professional if you think that's even a remote possibility.
If the BIOS was "corrupt" it would simply never work.
@@AdmrlSpeedy BIOSes can be corrupt/wacky in some ways and still boot.
I’ve been recently trying to overclock my RAM and after some failures to train or even shorting the RESETCMOS jumpers while the motherboard is still being powered by my PSU corrupted my integrated TPM.
It could not turn on TPM again as if there were never one. I flashed the BIOS and it came back
this was my exact thought. from other unexplainable issues solved with bios flashing seen on other youtube tech guy channels. But Graham Lord on Adamant IT has a bios flashing tool where you take the bios chip off the motherboard, so he has advantages.
@@linvyr that doesn't even make sense lol. Unless you have a hardware TPM module it's entirely handled by software.
@@AdmrlSpeedy yeah and firmware TPM is software, so prone to bugs since it’s handled by BIOS which is software too btw, and most Gaming PCs are using fTPM
Greg, don't stress to hard. My son's system crashed last night, when playing BO6. He tried to reboot and it sat at the loading screen, then he came and got me. After checking everything out, and all tested/looked good, I decided to do a fresh win11 install. Got it installed, but, shortly after, it crashed again. He was installing games on the 4TB HDD. So we go through the whole process again with same outcome. Disconnected the HDD and everything worked perfectly. Swapped his 4TB for one that I had in my computer. His system has been working great since.
I put his HDD in my system and I'm not having any problems what so ever.
Was the HDD connected when installing Windows? Sometimes Windows decides to put important OS files onto the HDD which causes unusual problems and sluggish performance.
@@unlap9377 it was. I didnt think about that.
@@DennisBuys Yikes, you are running games from HDD ? I only do video files from HDD nowdays.
@@gcmanuel85x When some AAA games are taking 500gb nowadays its not cost effective to put everything onto an SSD. I only put my most played games on 2TB SSD and the rest on HDD.
It's called money, it doesn't grow on trees.
"Is that really all it's going to be"
*Video still has 14 more minutes left*
Uh, I assume not.
I think it is refreshing that you are so candid with us and don't hide the mistakes you may make it has helped me learn a lot and also helped me to fix a problem I had with my system so I thank you very much and look forward to the next video and what I may learn from you.
Love the owner's desktop background, that elden ring picture is perfect.
Your candor is a very good reason to follow this series. I've been playing (and working) with computers for near 50 years. I've seen some really strange problems both with HW and software. Both can drive you bananas.
From my time as a PC repair technician and pc builder problems like this can be caused by physical damage.
Usually a bit of damage you cant see. So a broken trace inside of the PCB or a cracked solder joint. Heck there was a pc that we swapped 4 CPUs into because it would act weird but only when it was REEEEALY hot. (not the parts... the ambient temperature)
We finally swapped the mobo and it was fine. So I went over the MOBO with a magnifying glass and as far as I could see the MOBO was perfect. Cept for OOOONE LIIIIITTLE thing.
There was 1 pin in the MOBO socket that was 2 hairs width lower than the other pins.
So basically it was bent but not enough to be seen with the naked eye... What I'm guessing is that it wouldn't make proper contact. But only when it was hot... Meaning the problem was on a MICROSCOPIC level. To where a 10 c difference could affect the metal jussst enough to not make contact. Switching cpus would fic the issue for a day or 2... then it would start back up with the powering off in the middle of intense workloads.
So what would happen is that mounting pressure could make it cut off during a stress test... Bumping into the case.... Hell bumping the table leg with my foot set it off once. And this is what led me down the "it could be physical damage" route.
Back in 2000 when I was studying. I got "broken" MB. When you put it in case it stopped booting and without case it booted OK. Problem was one broken solder join in bios chip socket. Basically when you screw that on case it bend that socket little it and that pin didn't have contact anymore. Just solder that pin and I got free working MB for my rig :)
Broken traces or a CPU socket with destroyed pins. I've had experience with both. I currently have an Asus ROG Z390 Maximus XI Extreme that I had gotten a larger case for (this board is EATX). In the process of moving the board I somehow bent multiple pins causing memory issues. I can use some slots but not all and for a single stick I have to put memory in slot that is not recommended!
Keep the vulnerability and honesty man. Imo that is a huge part in your videos that sets you apart from other PC ytubers, and it feels more authentic to watch your videos for me. Been watching you for almost 3 years at this point.
I hate when these issues happen, it makes you feel like there is no answer. BIOS updates I've had fixed this issue for me before after doing all the same tests, but that still doesnt fully explain it (unless the implementation of the BIOS revision for the brand of the motherboard had a strange bug for a certain CPU?) but i have no clue. Ive really enjoyed these videos, keep it up!! Ive been watching these since the beginning, i just don't normally comment.
I hadn't read your post till after I posted and ya, so much hardware to support these days you have to research 9 ways to Sunday to find a bug.
I think this is the greatest troubleshooting series out there, keep the good work Greg, I enjoy your videos a lot.
12:19 This right here is every PC guy trying to troubleshoot a PC
watching this series has made me so passionate about building/fixing systems so finally bit the bullet and started to look for business so appreciate you man great series
Bios Issue maybe.
the 5600G POST'd in both motherboards
And he also reset cmos
Nah failing memory controller
Hey Greg, don't let it get you down. I've been building or upgrading computers since 1990 and I still run across things I absolutely cannot explain. None of this is 100% perfect 100% of the time or we wouldn't talk about 'the silicon lottery' when it comes to hardware inconsistencies.
It's also okay to be transparent about your feelings, because those of us who have been there can relate, and those of us who don't have your skill absolutely know what it feels like to be in that situation.
Thanks for these, Fixes and Flops alike, and please keep doing what you do knowing we love to be right there with you on the journey.
Seen that issue before and for me the b550 board for me had bios issue which was fix with an bios update the board you put the CPU in might have an updated bios which made it work
it worked in the owners rig once as well though
@soccerguy2433 yea It would there was a time on some b550 board and I don't know what but it something to do with bios revision couldn't tell what probably a bug where it just freak out
@@michaelpearce8371 A 5700G (same family chip, same architecture) also worked just fine in the same board with the same BIOS revision. I wouldn't be confident giving a PC back to its owner if a chip has already once randomly decided not to play ball with a compatible BIOS revision/AGESA update.
@@GregSalazar my bad never seen the bios revision most likely a dodgy CPU. Keep up the good work Greg watching from the UK 😆
Greg you being real or authentic shows how it really is sometimes working on PC's. And that sometimes there isn't a reasonable explanation for some things.
Adventures in pc repair!!!! Always a fun ride!
10:30 Those nvme heatsinks come with those rubber bands and a thermal pad. I've gotten a few of them off Amazon.
must be some special rubber then ^^ the regular rubber bands will dry out like nothing - then you have brittle stuff that sticks to the surface
hello Greg, love your videos man
Very much appreciated!
Since AMD came out with Ryzen cpus ive had the Same problems... just bizarre problems. From a 2600 through to 9700x . b550/x570/b650 and 3-4 types of memory. Bios updates did nothing, cpu would work with specific combos of gear, regardless of them being on the QVL or not. I had th throw the parts cannon on systems to figure out the Right combo to get them to boot and boot Consistently. Happy to see its not Just me dealing with this amd bizarro world issues. I still dont get it. Even had issues with Radeon cards and Nvidia cards as well. One day hopefully we will get some answers? The only thing I can think of being the problem is the memory controller? I may e wrong but thats my guess. Thanks for sharing this with us mate. Cheers.
Hows it going Greg, uploaded just as im about to clean my pc out👍
Time to write the Dean back at school
@ confusing as im from Northern Ireland and our language is pretty much the polar opposite
@@Plasma-Panda425 his college degree is in electrical engineering
@@mqcapps ahh
Another treat, and it's not even Christmas! Keep it going I always watch them 100%, it's always fun to follow along. I thought I knew it all, and learnt tons thanks to you Greg!
Always love these vids
Hey! I just ran through a very similar situation when upgrading my Ryzen 5 2600 on an MSI B450 Tomahawk to a Ryzen 7 5700x3D. I went through all trouble shooting steps and turned out one of my RAM sticks had given out. Very frustrating but even more satisfying to discover and fix. Great video
Fluffy stuff under the gpu o.o and rubber bands on the SSD lol. And Gigabyte PSU ohhh boiiiiii
Seeing your honest reactions to these issues is what makes this playlist awesome. Yes, even if you really can’t figure out why something happens. We’re all here for the ride; bumps and all.
Hi Greg, believe me when I say that hardware diagnostics can be a real pain, but when you keep going with the diagnosis, finding the root cause is so rewarding. I have had a couple trouble systems over the years since 2001 that took me days to resolve. I had a Dell server that just keeps on randomly freezing up, I ended up tearing it all the way down to individual components, and found a speck of dirt on a cpu pin. Cleaned it up, and it has been rock solid since. I had a desktop that would randomly just shut off, went through everything, and could not explain it after a few days. Finally the motherboard took a dump and quit powering up, replaced the motherboard, and it would randomly freeze up, found that one of the memory sticks was bad. After memory swap, the system ran rock solid afterwards. I built a new custom built pc in 2004, one day there was smoke in the room where it was at, looked it over and the front panel cables melted down while the computer was still running, nothing went bad besides the cables and the cables were not shorted against anything. It is still unexplainable today, never had a system melt down the front panel cables since then; completely baffled as to what happened.
This is most likely the answer, pin was blocked by dusk or debree, it doesnt have to be a bent pin
Many of the cheaper SATA/NVMe heatsinks come with rubber bands to hold them on. They're usually black, though, and I THINK the ones I've gotten were silicone instead of rubber. I've also gotten kits with clear bands that were obviously silicone. Silicone will definitely hold up better over time.
The ones on their drive look like they're rubber, because the one nearest the connector pins is already dry-rotting. They should probably replace it or at least swap it to the middle position.
Just two days ago I was building my iTX am5 system and while doing this I messed up front panel connectors on a mobo that was on outlet. I purposely didn't do a bench test so my imidiate thoughts were "crap, I bought a dud" and then a revelation "dude, you've been watching fix or flop for many months, get your sh1t together and rethink what was done". My mistake, no doubt, but I wish ASRock does not put usb and front panel prongs side by side ever again.
EDIT: As for the 5600G it seems that at 7:44 there is a ever so slight smudge of thermal paste near two pins at the top of the shot. A bit far fetched but plausible.
Could be a wonky bios, a dying memory controller, faulty/out of spec motherboard vrm... or a combination of the three. I wouldn't get too flustered about it.
Im honestly just as stumped
My initial thought was either a dead/dying Memory Channel, then perhaps a Mobo on the way out, but since neither of those panned out......I honestly got nothing
Stil a great video, just....very puzzling
It still can be dying memory channel but other MB has better tolerances on memory channel so it will work on that MB some time until it's in same state again.
I didn't read all the comments, but my guess is that a tiny bit of gunk (thermal paste, oil from manufacturing, a hair, etc) was stuck on one of the CPU pins or socket preventing contact of that pin(s). As the CPU was removed and placed back in, the gunk was moved from friction, or fell off with motion, and it started working again.
Great work Greg! I always appreciate how you show how frustrating it can be to work on computers sometimes
This exact thing happened to me recently, removed CPU cooler to replace paste. CPU came out of socket along with the cooler (apparently it happens to am4 CPUs somewhat regularly) and the system wouldnt boot after reinstalling the CPU. I thought there must be bent pins from the CPU coming out of the socket in an unintended way but visually I couldn't see any damage. System would boot with a single stick of ram in A2 slot. I removed and reinstalled the CPU several times trying to get it to boot and then noticed that some thermal paste had gotten on to the socket when the CPU came out attached to the cooler. I brushed it off, reinstalled the CPU and now the computer behaves as it did before the entire fiasco.
(Broken gaming PC)
I'm mad how smug he is that he knows we did it lol
Am I missing a joke 🤨
don’t be too hard on yourself greg. i have learned so much from this series and enjoy every episode
You know its a good day when Greg posts a video.
It's been a solid minute since I've seen one of these vids. I forgot how much they helped me to troubleshoot mine and a friends PC. Keep it up Greg :D
One thing that's been bugging me about these videos - what denotes a "season?" Greg does these videos throughout the year; he doesn't stop doing them over the summer or whatever. It's like when Rhett and Link say they're on season 18 or whatever but they have videos consistently (even if they might do a week here or there with less videos or with compilations). Or how WWE will randomly call an episode of Raw the "season premiere" when they've been on every week for like... 30 years.
I believe he's said in the past that he separates every 20 episodes into a season. I don't remember why specifically, maybe easier to categorize or keep track of that way, maybe just his own preference.
Seasons is just a way to organize episodes and break them down into more manageable chunk. It is arbitrary and mostly depends on the producers or creators on how they want to organize their show. In this case, if you look at the fix or flop playlist, you'll see that each season is broken down to 20 episodes per season. Ppl aren't gonna remember exact episode numbers all the time, especially when you can have a long running series like Simpson or in your example WWE, even in anime ppl will refer to arcs instead of individual episodes. Short answer. To make things more manageable.
@@Ynot563 Basically this ^
So great to see new content. I learned how to deal with a "no picture out" problem due to your videos. Not only are your videos entertaining but they are educational as well. Bless you.
Teacher: "All together, class! This here is a viewer's..."
Great video. Getting frustrated with finnicky tech proves you're human. Stay honest and consistent, and you'll continue to be successful.
Yes another video let's go!!!!
Greg I have fixed... computers for over 30 years, trust me, you will always find problems even strange ones,
What you do is amazing helping others, i will be getting back in to computers in 2025 as I need to keep my mind busy.
10:31 , Greg , owner's motherboard had P1.30 bios which predates 5600G support . 5000G series support starts with version P1.50 , you can check it out for youself at B550M-C support page.
This is a rootcause of inconsistent behavior.
Some boards can act weird and initialize newer CPU's without flasing BIOS to newer version . And you said that B550 boards come with native 5000 series CPU support but it's not true , they were released alongside Ryzen 3000 series.
Sorry to disappoint you , but you didn't fix the problem , and owner will likely experience same issue he had after a while .
Anyway , thank you for great content ! Keep it up !
go to the 'cpu support list'. 5600g/5700g is supported since version 1.00
I thought he replaced it with a normal 5700 but it was the 5700G. If that works then the 5600G should have worked too.
"And you said that B550 boards come with native 5000 series CPU support but it's not true"
Greg said "the boards support it natively, at least with the BIOS revisions that are on these boards". Greg was correct.
@@sergiobisonte Bios version P1.50 in it's discription literally states following "To support Ryzen 5000 G-series processors, it requires to update the BIOS with Matisse, Renoir or Vermeer CPU" in yellow text . In this particulsr case CPU support list is total BS .
@@ihorbychkov8742 I know. google that to see what it means... many people got confused too... you must see the cpu support list
One thing I found with AMD boards is that the pins in the AM3 socket can sometimes make an intermittent contact if the pins under the plastics has been damaged even though the plastics look clean. If the CPU was taken out previously improperly it may have damaged the pins underneath. Or it could have been a corrupted BIOS
As someone who used to repair PCs between 2002 and 2014, what you're seeing happens.
Some parts and combinations have issues while others don't. Changing one piece of hardware is all it takes.
There have been times I've replaced IDE cables and that solved boot and stability issues. Or I remember a issue where I swapped a CPU for an identical model and the fixed system worked fine and the old CPU worked perfectly fine in another system.
Sometimes you encounter the right combination of hardware that causes an issue.
I hate that this is probably the best explanation for what we saw, but I think you're right 😅
@@GregSalazar also could be bios related as he thought he had a 5600 and not the 5600g. asrock i just found out (by trying to upgrade cpu) has different bios for the apu's, compared to regular cpu's.
I've owned a r7 3700x that I couldn't get working with 2 motherbords and memory kits.
The last one was a MSI bord where you could try a downclocked memory profile. It now is stable, been working fine for 5 years! They are voodoo!
Maybe motherboard going out....
Faced a similar issue on asus b550f wifi. I was trying to add 2 more of identical memory sticks. The board refused to retrain memory on its own. Only a bios clear retrained the memory. I am guessing it was due to an old bios version. I tend to defer updates if everything is working.
Greg gets frustrated, we get entertained. I love these videos. Very informative regardless of the outcome.
7:04 Elastic bands? That well known application process.
It works for many different situations 😅
Not to mention the fact they're already degrading and cracking from the heat and about to break
they often come with the very cheap m.2 heatsinks you find on ebay/aliexpress/amazon. for an extra £1 you can get a much better quality heatsink with a proper mounting mechanism like one from Jeyi...
When an IT specialist in the Netherlands doesn't understand what is happening in a PC, they pull the "Fuzzy Logic" card. As if this helps.... It is just someone being not able to give a normal reaction to a customer and try to get away with it. You being onest of being confuzed is a breath of fresh air. I'm still learning with you, troubleshooting and solving problems.
This comment is critical, but it is intended with all possible kindness.
Not being able to explain or at least narrow down a failure often comes down to having too many variables still part of the system, and too many unchecked assumptions.
You could say there's two fundamentally opposite approaches to fault analysis: top-down and bottom-up. Top-down being, you take the complete thing you're analyzing, and using your intuition and knowledge of common faults, you change a variable here or a variable there, and you try to discover the fault. That will find common and intuitive faults the fastest. But since every variable of the system is still in play, rare and unintuitive faults will be nigh impossible to track down that way, the permutations are too numerous, and it ends up being blind guesswork until you get tired and frustrated that you can't guess the right one.
The bottom-up approach is the opposite: tear a system down to its smallest possible component, validate that one thing works, re-introduce one other component, validate it still works, and so on, until you add back in the component where the fault reveals itself to you. This takes longer for common cases, but with faults that defy intuition, there is simply no alternative to it. It replaces the guesswork with pure experimentation.
The biggest mistake I see people trying to troubleshoot a problematic PC is assumptions. Don't assume the power supply isn't the problem just because it passes a test. Don't assume that 3 seemingly perfectly working parts can't result in a fault once combined. Don't assume that something as mundane as a case fan or a SATA cable or a tweaked CPU fan wire or the damn case itself couldn't be a contributor to a system-halting fault, even if it seems there's no logical reason for it to be so.
Watching this series, you sometimes take baby steps towards a bottom-up approach (like testing the CPU on the "known good" platform in this video), but you have often voiced your resistance to doing more substantial tear-down. And (again, these comments are meant with kindness, not malice) I believe this fundamentally limits your chances of successfully identifying more difficult-to-find faults.
I understand of course that time is one of the biggest constraining factors here, and that tearing every system down to parts may not be feasible (though I do wonder how many times a troubleshooting attempt has taken significantly longer than a tear-down and incremental rebuild would have). But the simple act of removing the platform from the case drastically reduces the number of variables in play.
And I guess that would be the recommendation that I want to voice: once normal troubleshooting fails, assume that the fault is something you're NOT going to be able to intuit, and start slashing the number of variables in play. Some faults are simply beyond intuition, and there's no way to track them down except for shrinking the problem space until there's nowhere left for them to hide.
Appreciate your honesty. I feel your frustration. Been there.
I've learnt a lot from your videos.
What I've found in my journey in repair is cleaning is king when you have an intermittent fault with a PC. I always clean each edge connector twice with isopropyl and a fresh cloth and wash down the pins of the CPU with PCB cleaning solvent. I don't often need to clean the sockets but you will need plastic safe solvent if you do. It reduces the amount of situations where a part will work now and then or in different systems significantly, it was quite a dramatic night and day change for me. I've had good luck just cleaning dust off mainboards solving similar issues as this. Aside from the repairs I often get scrap systems given to me that other shops or organisations have given up on, I'd say over 80% work fine with original parts and some cleaning. Swapping the power supply can also have dramatic and unexpected changes like gains in 100fps even though it had great ripple values and delivered the power requested. Said PC was 100% stable at full load with minimal 1% lows, the only symptom of it failing was a slightly lower framerate.
hey greg if u are reading this, thank u so much for teaching me bout PC trouble shooting and in-depth stuffs about PCs. this playlist helped me learn a lot bout PC stuffs.
I recently bought a 800w power supply second hand, it claimed to be brand new and was sealed when I received it. But when under heavy load it makes a coil whining noise. Did I get scammed or can this just happen out of the box
A possibility is "tolerance" or "tolerance stacking". Tolerance as in spec +/- percentage . Theorizing an example for this case the 5600G needs just a little bit more voltage on a certain pin. The test bed MB provides it and the system MB does not. Any of the three can be in or out of spec, test MB slightly high voltage, system MB slightly low voltage or the CPU needing more voltage than spec as an example. It doesn't have to be voltage, could be timings or frequencies, latencies, etc.
Typically, base tolerances for these chips are _vastly_ lower than what the motherboard provides at stock conditions. This is, of course, by design. So if, in fact, the 5600G demanded a minimum influx of [insert variable here] that exceeded the board's often very liberal supply of [insert variable here], then a CPU replacement was warranted. Our 5700G (a more powerful version of the 5600G and relying on the exact same architecture) worked perfectly fine under identical conditions with the same board. In essence, this was why I came to the conclusion I did by replacing the CPU.
@@GregSalazar After a lifetime of fixing computers I've come up with a saying of "there's a logical reason for everything a computer does and sometimes you just can't find what that is". :)
I would done the same thing as you in this case.
You do good work, keep it up.
@@GregSalazar I think silicon aging is playing part of this. So other MB have more favorable tolerances so it may work few more months.
Almost reminds me like dying memcontroller like my 9900k case. Stopped working with XMP 3200 then and when run with JEDEC worked 3 more months until total dead. (Basically XMP overclocked memcontroller (max 2900 MHz) and speed up aging process) Got new CPU from RMA and set ram to 2900 and it's still in use. Donated platform to my niece when I upgraded to i9-12900k.
Great video as usual. the candid reaction is also another reason why I watch this channel :-).
I have to say this series helped me feel comfortable enough to build my own rig, thanks Greg 😊
I haven't watched this yet.. but the main thing that I keep seeing is the 4pin (or more) CPU power plug that is commonly an issue. Especially if you're overclocking.. I've seen those plastic connectors get brittle and break and have bad connections.. On mine, the pins were actually burnt or charred looking from arcing over time.. and then I realized that one of the pins in the board was loose and moving around.. After pulling the board and resoldering the pins and also soldering on a new connector to the wires.. It has been flawless ever since. This issue was a common find online too. Often times it will even melt the connectors.
There is a phrase I learned some time ago which fits this situation. "Do not call it a computer call it a confuser".
Another nice FoF video. Your services really helps the pc community as a while and i personally believe there should be some people who should take example from you. (Bring back the gt710!)
I just love this series so much! BTW nice lock screen wallpaper
Yup that’s a good one. I wonder if the bios version between the two motherboards has something to do with it. Newer revisions tend to get more memory stability fixes. Just a thought
Having serviced computers for way to long I understand you. All I can say is that there are a few basic steps you always try first. Clear the CMOS and test. Flash the BIOS with the latest version available and test again after another CMOPS clear. Minimize the system. Connect a known working PSU and try it. Swap out the GPU to a known working one, make it as simple as possible. it's a bonus if it doesn't need a power cable. Test again. Remove everything not needed for it to post and test. Swap in some known good working memory. Now we get to the interesting parts. I would suggest testing the processor on the cheapest and simplest known good MB you have. Now if that works then the motherboard is most likely the culprit.
Having a lot of memory, processors, graphics cards and motherboards is a real help, and unfortunately it's not something everyone has access to. Trying to fix something like this when all you have is the one computer that has failed can be close to impossible. But some of the early steps anyone can test. Clear CMOS doesn't cost anything and if the motherboard has the option for programming the BIOS without a CPU or memory installed then that's also an option. If the processor has a integrated GPU and the motherboard has ports for it you can ditch the graphics card and see if that changes anything. You can also try with only one memory stick and also try swapping the sockets. Don't miss testing both memory sticks, as any computer really should have two of them. One bad memory can be enough for the computer to fail POST.
Oh, and POST is short for Power On Self Test. It's the first procedure that the BIOS has and will be run before it passes control to the boot code for the OS. Just forgot that not everyone is speaking basic computer lingo.
When I tell you me and my friend have experienced these issues constantly, whether it's on builds we were flipping or our own personal builds. We've seen this happen between CPUs and RAM kits on different motherboards. I could rant about this in a 10 minute video if you like (a viewers experience POV). Additionally I've found AMD CPU compatibility to be so ass between different motherboards, to the point where you have to configure it in bios a specific way to make it work...
Oh ya brother of mine. Memory compatibility in the modern age. Even main board vendors don't have time to vet a fraction of DDR4 , 5 & CUDIMM. Memory remains expensive for all the wrong reasons. Read my post sidgySOHO_1960
Going waaaaaaay out on a limb with this one: maybe there was something interfering with the pins in the socket like oxidation or just a piece of schmutz?
Wouldn't be easily visible, and would be corrected by unseating and reseating a new CPU. I've seen the same kind of thing with vacuum tubes.
Like I said, out on a limb but it seems plausible if nothing else.
Being a G model chip, initial thoughts would be that it wasn't initialized correctly to run a dedicated GPU vs the on-die version, but that's very curious that it works on another board, and a new chip works on the owner board. When you get down to these seemingly granular issues between hardware ghosts and programming, it turns into a maddening roadblock that goes from one of a few things to one conflict behind a wall that you need access to in windows to sort out and said conflict prevents you from going further to get to the solution.
Mounting pressure, or a cracked trace. Happened to me with a B350, stock cooler couldn't post the PC unless you untight the screws slightly. Other coolers worked fine but with a different spring mechanism. 2nd cooler uses clip mechanism, 1st cooler uses 4 springs screws
In cases like this I suspect the CPU socket connections, it can be there is a slightly bend pin in the socket causing inconsistent connections.
The only thing I can think of is that it might be to do with the original owners RAMs timings, because it is the same speed not sure about the timings. Or a bios update? idk truly something you can't explain.
I would say that either the ram controller in the cpu is dying OR, there is a microfracture in the chip (both things can cause very similar conditions and are VERY hard to resolve/figure out) on a microfracture just a little moving in the wrong way, or heat expansion can cause the internal circuitry lines to disconnect.
You are a Good Guy Greg
I've had similar situations. One I can remember was a friend's GPU stopped working in his PC and his PC only. It worked perfectly fine in the other 5 or 6 PCs/test benches I tested it in, and all other GPUs worked in his PC. Any time I put his original GPU back into his PC, there would just be no picture at all. The PC was actually POSTing but there was no display. He ended up trading someone for the same exact model and both worked exactly as they should in each other's PCs lol
at 7:43 there is a smidge at the top of the CPU pins, it's barely noticeable but visible nevertheless. Could be the cause of the issue, especially if the motherboard also contained a smidge at the same place as well.
Check the output from the onboard hdmi port. I have a 5700G and it will sometimes dump bios/uefi errors to the onboard display. I had to use 3 monitors to figure this out. One on onboard hdmi, one on video card hdmi and one on video card display port.
Hi, please do not apologise I work in IT and I get but it was working yesterday4
never feel the need to shy away i have been watching you for 2 years and i keep coming back because you are you!
Could be how the board manufacture's bios is written, maybe do a bios update on the other board???
PCs can be very frustrating, especially when you have a problem you just can't pinpoint. I was having random freezes and went through so much stuff even reinstalling windows... only to find out the board was having issues. Ended up replacing it with the same GIGABYTE one I've seen you use in a few of these episodes, seeing as none of those PCs reappeared I figured it was a safe bet.
Wooo... I've seen that setup before. I can say with confidence this is a Skytech build as the single fan 3060, the Geil ram sticks and the use of a 5600g. I'd be surprised if it wasnt one of theirs. That being said, good work Greg. Sometimes, it's just something that is finicky but I would say it's likely that CPU.
My guess is maybe bad contact in the socket. If it were brought to me the first thing I would have tried was reseating the CPU and making sure to give it a good jiggle in the socket before pulling down the tension arm, I might have even tried contact cleaner. My other guesses for what the issue could be are a marginal memory controller which makes the problem intermittent.
My theory is a very small amount of thermal paste or something got on one of the pins of the CPU or on one of the ram chips and this contamination was then transferred into one of holes of the socket and when you changed the CPU that moved things around in that hole and this pin or pad was able to make contact again. When you get strange behaviors like these, it can be debris in the motherboard socket or in the ram sockets. After you have built like 1000 or so PC's and work in computer repair for over 20 years, you start to see every scenario possible, but I like your honesty man and I have defiantly had a few over the years that defied explanation like this. Here's one, you have a system that wont post, you check the ram and it's seated properly, you take out the ram sticks run the computer with no ram and get the expected ram error for the missing ram, reinstall the ram and the computer then works?? The only explanation I have is that debris was in the ram socket or on one of the sticks, and by taking them out and moving things around it either came off the stick, out of of the socket, or just moved out of the way enough to make contact again..
Greg, in this instance, I would have done an update to the BIOS. I have seen this before where a board has native support for a CPU, but doesn't work, but putting in another CPU works. Maybe the Microcode in the BIOS got corrupted and it just didn't like that CPU.
My ASUS ROG Strix X570-E Gaming had native support for Ryzen 3000 at release, but yet my 3800x didn't work. My Ryzen 5 2600 did work though. I upgraded the BIOS to the latest version and put the 3800x back in and it worked.
I know the 3800x worked because I tested it in my other PC which was a Gigabyte Aorus X570 Pro, and the CPU worked just fine.
Maybe something to try if you see this issue pop up on any future Fix or Flop episodes.