Вставка
- Опубліковано 12 вер 2024
- It's another custom build no-POST diagnostics session, this one keeps turning off and on again. Spoilers down below...
CPU wasn't seated properly, and had trace amounts of thermal paste on the pads, and a bit of motion on the CPU cooler caused it to intermittently not POST.
_
Thanks to music bros down below!
Synthis - / samsynthis
"It's never the CPU, until it is"
Too-tightly screwed down CPU coolers can disable memory channels.
The memory is in the wrong slots to begin with as well.. it has Ram in slots 3, 2 and 4 and nothing in slot 1. never mind that this configuration also guarantees single channel ram operation and the customer would have been better off with just the two blue modules.
@@bdhale34 yep
Which is why I'm glad I go Noctua, their coolers are impossible to overtighten. ;)
really? didn't know that. can you explain it further please?
@@spacecy when the cooler is too tight it puts more pressure to the cpu. This pushes the cpu to the motherboard's sockets and can possible bend pins. Bent cpu socket pins can affect ram memory slots.
Nice work! Good to see proper trouble isolation being performed.
I had some weird issues with a z170 board once. The cooler on the cpu was a hyper 212 evo. After removing the cooler and the mounting bracket I noticed one of the stand offs for the cooler was missing the rubber washer that goes between it and the board. Screwing it down had exposed the traces on the board which caused the standoff to short the traces causing one of my memory slots to not recognize ram. I used one of the red cardboard type washers, reassembled and everything worked perfectly after that. I thought this video was going to end up the same way. Especially after seeing how tight the cooler was screwed down. Anyways good work as always!
Hey I know its been a while but I wanted to let you know that your comment worked with my z170 too! I have been working on this for over a month and this worked for me so thanks 👍
@@charlesbwilliams Awesome, glad it helped!
@@pas3pas365 You have bo idea how much this helped 😅
When I was on the test bench many moons ago the MOD and Post Office insisted on a 10g shake for 3 minutes before testing for acceptance started.
Same problem here with the original Hyper 212 Evo. I'm glad that they fixed it with later versions. Plus that mounting system was a pain in the arse.
You are the Big Clive of computer repairs. Excellent content!
Hmmm.. wasn't the 212 back plate on backwards, specifically with it's green isolation pad facing the wrong way round?
the other side has pads on the feet, if I'm not mistaken one side is intel the other side is amd.
My heart skipped a couple of beats when he started shaking that computer.
Shaking devices, even hitting them (without bending things of course), is part of standard QA testing. Only devices I won't do that on are used printers.
@@eustacehendrix9659 Odd, I've never seen that on a bench QA checklist for a PC 🤣
Like he says in the video, it's really only for when you suspect the issue is directly related to that. Typically it would neither be necessary nor recommended.
(And yes, my heart skipped a beat as well, even though I knew it was coming and I knew I'd do the same) :) only thing I might've done different is then test the case right side up; perhaps even shake it in that position or leave it right side up for a couple of days and then test. Bad CPU sockets are a huge pain to troubleshoot, one reason being because they rarely fail in the same way; I'd want to see if the lateral force from the weight of the cooler eventually shifted the CPU *just* enough for it to fail again.
25:08 😳😱🫣
how the hell do you rip off a pcie connector
The fucker literally ripped out the gpu together with the slot holy shit
Maybe the PC got shipped and got destroyed in shipment
@@DarkoPetreski I have seen this one or two times when i was working for a big pc vendor. The slots have a lock on the back and if you do not open it before pulling tha card out either the pcb on the card cracks or the slot gets damaged
@@svenhoff2653 yeah i know but how didn't he realize that the card isn't moving?
@@DarkoPetreski It is hard to understand! But imagine someone comes into your shop and is asking if you could help him because his VGA card is not working. You check the card and not only the hook on the end of the PCI express connector is missing NO the last few pinns are also brocken off. And i am not talking about a 100$ card. Expensive ones 600$+ cards. It was sometimes redicoulus. Completly burned out PCs because someone put a power connector the wrong way onto a sound card (1200W single rail PSU). Even the isolation on the cable was completly gone. Not melted. Completly vaporized. And they come in screaming and shouting because the new bought PSU allegedly damaged their PC. Sometimes it was hard to reconstruct what happend because it was not much left. Or completly corroded heatsinks because the custommer does not belive you when you say Do NOT USE Liquid Metal on Aluminium coolers. . . . But you also had nice custommers and they were very very happy if you were able to fix things like a pushed out pin on IDE HDD (because they pushed the connector in the wrong way). Solder the brocken pin back in place and use some glue to keep the loose pin where it should be. Or change a HDD PCB (we had the same model and revision at hand) to recover his doctoral thesis 3 days before it had to be finished. . .
I got my pci express slot replaced 2 years ago. Thing is i had second slot but the gpu was not working properly there. The guy did an amazing job and the mobo is still working like it is new. Best repair shop in Cyprus and only one that does it as far as i know.
Another excellent, real-world trouble-shooting video typical for a repair process! Excellent work, Sir!
They call it "the British B-Road simulation test" :P
Adam is shaking the pc
Hdd : wtf is going on xd
ssd is the future XD
Once you cleared cmos, you should let the PC do some power cicles. When the bios is reset, the CPU/BIOS has to train the memory again. Memory training can take several power cicles to finish. Also, sometimes, if you have 3 or 4 modules, you have to install only 2 of them in dual channel, make memory training until you get to bios, and then install the other modules.
About the i7 pcb issue.
Skylake CPU had thinner PCB than Haswell and can get bent with the mounting presure of the cooler.
Direct contac heatpipes are fine when made correctly, the first Hyper 212 was pretty bad the EVO revision was the good one with the Heatpipes stacked together. Also, there are other issues with this type of coolers, like a not plain base, misaline pipes...
A well made copper base is better, more reliable and you can put much more heatpipes with a base.
You dont need that much thermal paste, its not about temps, its just a mess.
@@awalon2572 what was the final outcome? Ive had this power cycle since i removed my ram to upgrade.
I can't believe you brushed the CPU socket like that. Surprised you didn't bend any pins.
He can put them back.
Brush is soft and pins on a quality Mobo are not soft as a hair.
That poor computer seems to have had a very rough life and the "shake test" is not not an improvement.
Shake test or even hitting devices on possible weak points is part of standard QA testing for quality products. If it failed there then it's not going anywhere.
I've had same symptoms with the exact same EVO 212. Replaced the cooler and the PC worked perfectly... Never knew if it was a bad connection or too tight. :(
I had this same problem after I installed my hyper 212 evo. This video helped a lot thank you
In out shake it all about!
Satisfying sip of ☕ over the cpu.
"This works now.... 💦... Dang it!" 😅
Excellent point on screw drivers. Thank you.
interresting except some more research would have let you know on that board the 16x slot is actually easy to break off tons of people have done it and fixing it not a bad option but your quite lucky it wasnt the ram because another issue with the boot cycle is that the ram slots have failed on the board or broken solder joints this was not a good MSI product for its time MSI produced another board that was similar for the core 2 duo processors where the board would wipe hard drives power cycle and fail to post it was a faulty firmware update
The "correct" way to install different RAM modules is to put all the same spec ones in their own channels, that way the different models can be trained independently and run at different timings if necessary.
i"ve got a MSI z87-45 (same socket) with two faulty ram slots...broke the memory modules as well. The i7 4770k works fine on another nieuw mobo...
Thank you so much!... very similar behavior on my pc. Just waiting for the thermal gel to work on the cpu reseating. 🤞😁
thank you . has a working tech(not board level repair eyesight, hand shake) sometime I disagree with you
but I always learn something
Louis Rossmann - uses a high-powered microscope and heat gun to gently move a chip a hair's width to reseat in on the solder balls. AdamantIT - let’s shake this computer to see if we can giggle anything lose.
If you'd like a demonstration in the Dunning-Kruger effect, here it is. Not only did the owner build it, but they also built it wrong. Then, they fully broke it to where it needs to be sent off for repair. Gets it back & fucks it up again. Whenever a repair shop thinks about giving you a loyalty card, you maybe need to consider a refresher course of PC building.
My goodness, this boy is good! Love your vids. Keep up the great work 👍
"I will repeat it for the people in the back" lol
read this comment exactly the moment he was saying it xDD
Actually there were 2 issues: a bit of thermal paste on the cpu contacts and keep in mind that cooler master cpus have a really tight and agressive mounting system...it happened to me that mounting the cooler with uneven pressure would make my computer not run in dual channel, as the controller is in the cpu and somehow was not seating properly because of the cooler
16:23 The cpu had like you said, a few corners that look rounded. I think the cpu has experienced a few times Newton's law of gravitation.
Then you get slightly out of shape corners (not square, but swollen corners) , that's why the cpu didn't wiggle because the bottom wasn't flat pressed enough.
Also shaking a pc while it's on, will 0,0% help when there are bad Ram, Cpu or GPU connections, because when a bad contact gets back in contact, the fault is still there till the pc is power recycled and did a new initalization.
Initalisation of Cpu, Ram and GPU is not "on the fly", but only once at boot, those devices are not "hotplugged" like USB devices.
Compare it with the old PS2 keyboard and mouse back in the old days, when they where not connected at bootup, they didn't work if you plugged them in while the pc was on,
but only when you rebooted the computer they got working.
you should read about the correct way (and amount) to place thermal paste to avoid air pockets and overflowing. the thinner the layer the better
There IS no CORRECT way. There are only different ways.
@@chrismoule7242 did you read the part about air pockets and overflowing?
The tightening actually depends on the mounting system I have a Scythe Ashura on the same chip and it has no springloaded screws or anything. First screwed it down to a point where I gots cared and stopped. CPU failed the stresstests. Lived a while with it then took all my heart and tighted the screws to a point I thought I am breaking the board, i.e. fully down so metal touched metal. The power to achieve this was immense and it felt horribly wrong. yet in the stresstest after I ran quite a couple degrees cooler and passed the test without thermal throttling.
I had a similar boot cycle problem a few years ago with my MSI X58 and Xeon setup. Initially suspected power supply after pulling all cards out and moving RAM around. Bought new power supply and still had the same problem. In the end, I pulled the CPU and noticed slight brown corrosion on some of the pads. Cleaning the CPU pads with isopropyl alcohol fixed it.
Hello, kinda need help rn nd potentially having similar issue like u did. If youre still active, I want to ask. Do you mean the cpu pad as in the gold pin? I notice one of the many gold pins under my cpu it brown
I would not be able to sleep at night with the Cooler Master logo facing away from you lol id have to turn that bastard around
Adamant IT... I've had this problem before with one of my MSI Boards. My older MSI P67A-GD65 1155 i7 2600k. (Zalman 9900 max copper air CPU cooler has never been removed or touched since new) Earlier this year I was installing a USB hub into a spare 2.5 front facing bay in tower. I had to remove the GPU, 4 sticks of ram, and moved a mechanical hard disk to install front header. Long story short... the system boot looped every single time unless only one stick of ram was used closest to CPU. I did get it working with all sticks by using electronics cleaner on the ram stick"s edge contacts. My system doesn't like being messed with or this boot loop could happen again. I'm thinking since it's 9+ years old and chip creep and age has made a weak connection? Ram on these MSI Boards seems to be pesky problem as well. I also have a "Semi working" Z77A-GD45 with only on stick of ram since it boot loops or doesn't posts with any more installed. Might give the ole' college try and clean that board to post with more ram. Adamant IT thanks for the informative video.
Power cycling is an annoying problem to diagnose. I had a used board that power cycled, and the problem was that the CPU cooler mounting screws stripped the motherboard PCB coating. That specific board had traces right next to the cooler mounting holes, so the underside of the cooler's mounting screws ended up shorting a couple of those traces. I covered those traces with kapton tape and it never power cycled again.
pretty sure the problem was thermal paste on the cpu lol
19:00 I wonder if direct touch heat pipes are all that great. Most of the more expensive coolers have solid cold plates.
It depends - cold plates are better if the soldering between the cold plate and heatpipes is really good. But on lower-end coolers that are much cheaper, direct-contact is better than badly soldered heatpipes.
Next time, you have to shake it with an NH-D15 mounted instead. :D
Great channel btw, i can´t get in my head why you not are bigger here since you are very pedagogic!
/L
Very well done, to the point, clear
I just noticed that at 6:20 the cpu cooler back plate is on upside down. This would cause more pressure on the cpu. the three bolts thats hold the lga bracket down should line up with those 3 holes
Another good job. Thanks again for sharing Adam
Really enjoying this channel. Better than Louis Rossman cos it's just about fixing stuff and less about the right to repair stuff. The right to repair is a real thing though and Louis does great work there I just don't wanna hear it when I'm trying to learn something.
mute your discord mate.
That is why I prefer AMD sockets and processors - CPUs may be a little more fragile but sockets don't have these weird connection issues and It's not easy to damage them. I don't get how people pull out AMD CPUs with coolers, that never happened to me because I care about my computers :)
It's not the problem of the socket. It's rather the fact that the customer screwed the cooler too tight and the hyper212 is a pain in the ass to install.
Im guilty of over tightening. Spent 23+ yrs as a mechanic. Military nickname was to tight/snap off. After 2nd deployment I lost my finesse lol
I have a pc with the same issue, been using it for 4 years. If I transport it it will power cycle and sometimes post but it will give me the symptoms of bad ram if i manage to boot into windows and play games. All i need to do it power off, reseat my ram. Maybe a few times, and it's good to go.
@@awalon2572 yeah
i saw this problem two times in my life as pc technician and this symptoms was always a power button that was closing the circuit whenever it likes, just my two cents as a fellow technician
But he had the mobo out and was shorting pins to start it and it was still intermittent
That's why you do a bench test with everything else out of the equation.
3 sticks of ram and broken pci-e slot. I would have removed the white module and installed the blue ones in dual channel. #fail
When the cpu cooler is too tight, it won't post.
The dual channel for that one was 2 and 4. 1 and 3 Which he did, and he can't ditch the white ram because it's not his computer, he just has to make it go.
checking the memory should be step 1...
Yeah memory is fukedup...
The tighter the better with water blocks,now because those 212 HSF use thin sloopy mounting hardware I always modify them for more pressure and use the correct tools.
I only chipped a tiny piece on a P-3 chip from over tightening and chip still worked.
In my experience power cycling is either 1. Power Supply 2. RAM reseat 3. Case ground.
If it's not those motherboard or one/two components are damaged.
Got a simulair problem also with a z97 gaming 3 board, also with the pci-e ripped out, and the same cpu cooler. Before watching this vid I thought it might have been cpu seating issue or thmermal paste or smth on the bottom of the cpu not making a proper connection. After some troubleshooting it seems 2 of the ram sockets seem to cause the power cycling. The other 2 work ... but then the ram is in single channel, not quite optimal.
very good job m8
Power supply not delivering the correct voltages.
Have to take out my cpu to see if that is a problem on mine also.It has never been over 60 degree C temp.But corrosion can happen.Specially when it often has run at a temp that is close to room temp.
over tightened HSF causing the PCB to bend, which probably caused a short or bad connection somewhere
thanks @adamant IT i fixed a pc from this video
LGA don't like that heavily screwed cooler that why is bootloops
Can I ask your advice. I have an AM3+ motherboard which works absolutely fine...EXCEPT! When I shut the system down the cpu fan keeps running....and simply will not fully turn off !! Any suggestions ? Thanks for the very informative videos by the way . Top job.
I hate when custom built system come in that I did not build. Never know what the frack was done to it during the assembly process.
great video as always
Wow. I've only had a few out of thousands of systems where the cpu was the issue. Interesting this one.
So, it's safe to just drop the cooler onto the CPU without thermal paste, while you are doing these kinds of tests, without risking frying the CPU? I've got a Gigabyte Z87X-UD3H board has been working fine since 2014, but now it'S power cycling like yours... I tried a new power supply, same result... Yesterday I drove 1 hour to buy a used LGA1150 motherboard from some guy, drove back and reinstalled the CPU with cooler in that new motherboard, and.... I still get power cycling... it's driving me insane :( I'm glad yours seemed to have a problem with the CPU seating, I might just try again and remove the CPU, use a small brush like you did, etc... thanks for making that video though :)
quite a few likely culprits here, but if it was cpu mounting issue wouldn't it be likely the cause wasn't shaking but that you were running it laid flat and the customer took it home and stood it up
I recently had a power cycling problem - from post code 00. Research said that it's mosly a motherboard problem so I RMA'd the board and got a replacement. But then the next board failed too with a 98 post code (GPU). I took the PC to a shop that had compatible CPUs who confirmed that both the motherboard and CPU were faulty. I've now got a 3rd new motherboard and am waiting for a replacement CPU to arrive from AMD after RMA/testing.
I also played at home - trying different ram sticks in different slots, removing the mobo to check for errant screws, trying a different GPU, disconnecting all drives, CMOS CLR, checking CPU pins, change cpu cooler. Hopefully it'll all be OK when I get the replacement CPU.
So.... was is your CPU?
@@davidbergewaytogo both the motherboard and CPU were faulty - R9 3900X, Gigabyte X570 Aorus Master.
@@i7baby814 damn... talk about unlucky... power surge maybe?
@@davidbergewaytogo It wasn't a power surge. The CPU and the mobo are protected. I suspect it was a faulty CPU that messed with the mobo - I had to replace a mobo before this one (on warranty).
I find I get this when my graphics card is not at the proper angle, so i have to sort of move it around a bit till its finally happy, its also left slightly loose as if i try to screw it down to the rear mounts on the case it forces it to almost come out of its slot, I also had that same power supply loved it until i accidentally shorted it by blindly putting the display port cable into the HDMI slot did not think much about it until it kept power cycling, after lots of testing i realised I had damaged the power supply and the graphics card HDMI slot, i remover all the pins from the slot (as to stop the shorting) and it powered up but the power supply cut out under load got another one works perfect!
Mine is having the same problem but I did everything you did in this video and checked out my cables everything is good but it still power cycling? Can my ram not be conoatable with my mother board?
I got that Gold Rated CS750M ex demo for £40... built by Seasonic or Great wall i think...its brand new with a little dent on it..amazing discounts on out the box stuff...
Only X58 should run triple ram..thats dual channel with 3 dimm kits in...
It could be his goddamn sheep sweater that is providing a lot of static electricity and that is why the motherboard is fuck up.
You should use some thermal pads as your temporary thermal solution you can reuse them which is nice
Doesn't the battery have to be out for a longer amount of time? Like 10-15 minutes or so?
I've never put it to the test, but 3-5mins should be enough. I've seen 3-5mins work enough times that when it doesn't work, I can assume that a BIOS reset wasn't the answer.
Where is the first PCIe slot? This PC got a brutal gaming time for sure :\
0:54 he explaint it
Mine X97MX gaming 5 pc also have started doing this.I have to remove a ram stick and it start normal and restart as normal.But if i turn it off and let it be off 30ish min it start to power cycle again.And i just have to turn power of press power button to empty whats left of power.Put a ram stick more in and it start and restart normal again.I have to change the amount off memory to make it start,in other words
Is there a way to create some sort of test device for each component to test it right away outside of the whole system? So you don't have to A/B test it like this?
Theoretically yes, but you'd need a huge amount of test gear on hand. Diagnostics with whatever's in the box alone can be more time consuming, but are also more feasible when you don't have shelves of test components.
@@Adamant_IT gotta work towards it then. maybe contact the manufacturers and hope they have some old stuff they don't use they can donate.
usually Power-cycling/bootloop/not power etc the first thing check is ram not GPU (provided all psu cable connected correctly)
if ram still no luck then reset the cpu will do the job most of the time
Don't suppose you have any tips for this board cycling through some debug codes and not posting? I tried a usb stick with the latest bios file on it but the board just cycles on forever. Bought 2nd hand off ebay but it was listed as working. Some people are just so dodgy
I have similar issue and I installed everything one by one to check out which part is causing that issue.
And in the end I managed to figured out that when I plug in the 8 pin power connector to the motherboard my computer starts power cycling. And when I plug it out n then start my computer it starts without power cycling but without power connector plugged in I can't run my computer as it just starts without power cycling but won't show anything on screen as the power connector is not plugged in.
Now what could be the reason behind it? Is it my bad power supply as when ever I plug the 8 pin power connector to motherboard it starts the issue or is could be my motherboard 8 pin socket which is faulty?
this was awesome. thank you for showing how you diagnosed. my only question for my problem, my computer tries to turn on right from the switch of the power supply. i dont even tough the on/off switch. it just begins the power cycle from that point. would i still go through a similar breakdown of checking each components?
Same prob mate. Have you managed to fix it?
Just fixed mine replaced the cmos battery.
Can you look at my PC? Same build just about and same issue... Spent $160 already at a local shop for them to tell me the wiring of my house is bad...
What do you think it is if mine will boot, run for a minute with the CPU fan at full speed then shuts off? It doesn’t supply any power to the USBs or Display but will run for a bit longer than a normal power cycle. Then reboot
I have cooler master t400i , kind of same as this one in this video. But mine doesnt have a flat surface , in fact there is a aluminium between 4 heat pipes which are not flat all together. I replaced ryzen 5 3600 stock cooler with this which was working terrible. And had no problem with this one before buying new case and also hot weather I started to see high temps. And I was wondering if I need to flatten the contact surface of my cpu cooler. I can tell that there is no good contact between heat pipes and cpu surface.
why wont you replace pcie slot just get a dead board
Never thought of that tbh! The GPU from this build won't mind an 8x slot, but that's definitely something to try in the future - although broken PCI-E slots are pretty rare I think.
Sorry to criticise your customer, but how in the hell do you snap off a PCI slot accidentally? Honestly, I thought they started making metal slots just to justify price creep.
Also: 18:32 "mirror finishes" are usually nickel-plating, which is not only more chemically inert than copper, but when bonded, the heat transfer coefficient of that bond will far exceed that of the interface between the cpu die and the heat spreader, or of the heat spreader and cold plate.
It doesn't really sacrifice any performance to nickel plate a copper cold plate, but it does allow you to use gallium-based thermal interface materials.
the old msi boards the plastic becomes very brittle overtime .. had it happen to me on the same board . release the clamp and the damn pcie plastic came away with the gpu lol
I had a power cycling fault when inserting new ram.i had to leave one of the old ram modules inserted to stop the power cycling.
the ram was in the wrong configuration when you first started and when it was finnished the ram was in different slots
Thanks for the video, it took me 8 hours to fix my brother's PC and I could not figure it out until I took out the ram then re-seated the cooler and didn't screw it down as much, although it still needed a new PSU I'm glad I don't have to buy a new mobo because I though that was dead.
Oh, dear, I've got an I7 in that exact case. I hope they didn't both come from the same company.
The customer built that computer
I just got rid of my GB H61 MOBO. POS it was , bought an unlocked used CPU and it had every BSOD possible upon trying to get into Windows. Was also showing many RAM errors on Memtest 86. Well guess what ? The RAM was fine , so was the CPU.
I got a USED MSI H81 and everything works now. Never Gigabyte again. Had corrupt BIOS several times , and eventually would not use the backup BIOS to recover and would not boot. Best day in a long time is when I tossed it into trash.
Possible failure on power supply rail to RAM. Most likely a capacitor losing it's capacitance which can cause unstable power to RAM modules. It will be one of the bigger capacitors like tantalum or electrolytic.
"car journey"?! That was more like a dirt road trip... still, desperate times require desperate guys looking for an answer. Also, it made me doubt but ¿Isn't pulling the sata connectors just stop data transfer? If you want to prevent the drives from powering on you have to disconnect the power connectors ¿Right?
He's out in the countryside so not great big roads, and I guess you can pull the sata cables out of the motherboard from the front if you don't know if you're taking it out.
Have tried everything with memory but it is something with ram.Have the same cpu also
Where can get schematic brd files for open board view app?
I have the same motherboard with the i5 4690k and it was doing the same thing I did a lot test like you never find out I reas some where could be bios problems .. it happens to me that I install the fan connector to the wrong pin in the motherboard .. I restart the computer and it start doing that.. I bough another mother board and the cpu still working I have to try one day to see if is working .. thank you for the video
You didn't wait for the thing to do the thing. Rookie error there.
The common issue of CPU heatsink is too tight... That causes a lot of strange issues.