I might have accidentally said that the system was powered on when the RAM was changed during the live audio portion of the video - if I did, I meant to say that it was plugged in. The system was not powered on. My bad there!
At 21:05 you state: "the thing that most likely caused it to be broken by that procedure was the fact that Carrie left it plugged in while he did the ram swap" Plugged in, not powered on? Regardless, there are many power states to modern computers. I think we have a tendency to attribute causes that we are familiar with. And therefore I think you are searching for an electrical problem. Have you considered a firmware problem, i.e., that the BIOs is corrupted? As the RAM was removed while the computer was in an unconfirmed power state, and multiple OS levels store data in firmware that is implicated in POST, flashing the BIOs would be a good step to try! You can likely fix it with a USB Bios Recovery, even with Chinese Mini PCs, as they are based upon the Intel (now ASUS) NUC platforms. Enjoyable video. I liked seeing you trace the electronic paths.
@@DustinKleiboer That's a good point about the BIOS, and me focusing on an electrical issue. It hadn't crossed my mind whatsoever, but you're totally right about it possibly being the BIOS
you will have to track back to where that phase is getting an enable signal. (nightmare without schematics) a good channel to watch for that is northwest repair. GPU's use alot of phase enable sequencing. In the end i suppose the lesson is inplug and hold the power button in a moment, before changing components and any system. remember off IS NOT unpowered.
Beelink's problem is 99.9% a BIOS firmware failure. This type of Chinese semi-finished products has catastrophically problematic BIOS firmware. The HDMI output will not work until the OS boots and therefore no communication will be established between the GPU and the display. Try to copy, save and analyze the BIOS firmware, and it will be very good if you can find an “engineering bios”, try to find a similar Beelink and copy the BIOSa firmware and write a new, working firmware to your BIOS chip. Good luck!
Thank you for confirming my suspicions that arose as I listened to the diagnosis. I've become aware over time of how prevalent BIOS firmware failures are.
yeah a lot of new UEFI firmware especially on some laptops store settings in the same flash chip as the bios. so simpyl detaching battery won't reset bios defaults. change any setting wrong to make your system unbootable and you're fuxked, cause almost no laptops have a "cmos" (more like nvram??) reset button. had this happen on a laptop once, thankfully I had a full flash backup, flashing that also restored my working settings.
"The HDMI output will not work until the OS boots" - that's just false. HDMI is handled by the iGPU and it works on Beelinks exactly the same as it does on any laptop or desktop. One important thing is that if there is more than 1 HDMI out, only the primary one will be turned on initially. Figuring out which one is the primary is a fun exercise when nothing works.
@@GizmoTheGreen Had this issue with a craptastic Machinist x99 k9. Adjusted RAM timings, would no longer boot. BIOS battery removal didn't reset the bios, which is a known issue for many of the crap x99 boards. Solution was to buy a flash programmer and write the stock BIOS to the chip.
Wow I am so so impressed! Carey's channel administrator here - just had to say I so admire your skill, ability to detail all the steps and the fantastic work you did! Just so so good! Fantastic walkthrough and explanation!
A bit of information for you. The resistor that you replaced with that radial then bridged was probably .02ohms used as a current sense to check that the power button is pressed. Don't worry too much about the ripped pad happens to all of us. Good on you giving it a go. I wonder if the cap right next to that resistor went short and took the resistor out with it before it burned. That could potentially explain the odd solder and the burnt flux assuming someone else hasn't had a soldering iron on it. For the beelink you are probably right that the CPU is cooked although I could be wrong but typically voltage isn't present on the ram voltage rail with the system off. I have never worked on one of these mini-pcs but typically in laptops and desktops the ram voltage isn't a standby voltage meaning it should only be live when the PC enters a powered state.
That's interesting and makes a lot of sense with the resistor! Would you say that's something that should be replaced with a properly valued component or is it something that doesn't matter all that much? I tested the capacitor that was hanging out next to it and as far as I'm aware, it was fine so I think it might have just been the resistor that went. That or someone at the factory had a go at it, as Carey doesn't really do board level repair work. Good point on the RAM voltage as well, I'm pretty sure that the RAM voltage rail in this case was live only when the system was up, though I'll have to double check that when I pull it back out at some point.
@@CHWTTWith the state of that pad I probably wouldn't mess with it. If it is working then it should continue working. I suspect that the whole ram part might be a red herring. I think maybe if some damage did get done it was done on the board itself not from pulling the ram out. Diagnosis of dead CPU might be correct either way although I would probably look at that IC that drives the core voltages. Usually these aren't too expensive and if nothing else this would be a good chance to test out your hot air rework skills after you mess around a bit with other components.
Definitely going to look at some more things on the board in the future, somewhat apart from the RAM issue as it definitely could be a red herring as you said. I was considering going and replacing the power IC on this board, especially because I was also thinking it would be maybe a few dollars at most. However, I couldn't find that chip for sale anywhere! I may have to contact the manufacturer and see if I can get one from them, though I feel doubtful they'll entertain a random guy's request to buy a single chip.
I am quite convinced that the resistor was only a 0 Ohm resistor used as a jumper. The VGFX line is only active when the integrated graphics card is loaded. Otherwise, the entire CPU is powered by VCORE. It's going to be tense there. S No video output can be a bad ME region in BIOS. You need clean this part of bios but in modern pc is not that simple.Try measuring the resistance on pin 8 of the bios chip.
if the vendor supplies a full bios image. some new formats are upgrades only and need to be applied by their flasher onto the existing bios copy. also lots of new uefi bios has device unique keys that you need to dump, and thats if you know where to look and might require sketchy or hard to get softwares...
@@GizmoTheGreen where do you find hard to get softwares? forums?i am sure if i was inclined to know i would get them or locate them the same with all illicit things i suppose.
Try this: Connect a keyboard directly to the machine (no kvm switch). And when you turn it on try to press del, esc, f7, enter or other keys. The f7 is the one for entering bios in normal situation. In similar situation, After some trying, I noticed the keyboard started showing life and display also turned on Some of these mini pcs have bad bios and they does not turn display on boot screen even if they "show" a message and waiting for input to continue.
I have two mini-PCs and boy, there's a reason that they put a BIOS Reset button on them. With the unlocked do-whatever-you-want BIOS, you can do stuff that you shouldn't have access to, like enabling PEG graphics... (I don't think this applies to your case unless the BIOS really gets stuck without battery power)
@idan678 PCI Express Graphics (i.e., discrete / external GPU) I should have written "Set Graphics Priority to PEG" because it's not something you enable but a selection menu (also, exact wording change depending board manufacturer). It allows the user to select what graphic card should be used and have options like Auto, IDG (Internal Display Graphics), and PEG. This setting when set to PEG will ignore the internal graphics at boot.
@@TommyCrosbyYou can actually attach PCI Express graphics cards to these mini PCs through an adapter for the M.2 slot. I have such a setup on my desk at the moment. I wonder whether that setting in the BIOs would actually do anything if you had such a setup
If it were plugged in and the ram was changed without draining the power first, integrated graphics uses your system ram< 31:00 (shared) I can understand how that would effect the other.
If you have a thermal camera, check the board with the power on to see if anything is overheating. If you don't have a thermal camera you can put alcohol on the board and see if it evaporates quickly over a specific chip to see if there's a short. Northridge fix has videos on this method.
I was thinking about doing that isopropyl method (can't afford a thermal cam... wish they were cheaper!), but the resistances around the board and even on the dead rail all seemed like they were right, so I don't think there were any shorts in this case
Yeah, do NOT screw around with anything on these mini PCs with the power connected. I fried one of my two Intel NUCs by doing basically the same thing as the guy with the Beelink did, try to change memory with the power connected. Luckily the one I fried was a cheap one. Exact same symptoms as the beelink.
For debugging it's worth while investing in a post card, that way you'll know exactly where the machine fails to post. It's a PCIe card but you can use an m.2 to PCI adapter and plug it into the wifi slot
@@R4MP4G3RXD I haven't experienced it yet, I don't even own one of those disgnostic module.... i always try to connect that tiny speaker and it worked for me so far.
I bought one of these N100 minipcs for my parents and it came dead, very similar issue to the Beelink. I did not manage to fix it either: tried similar Voltage(and Amps - not that it matters) charger, probed some resistors and no issue there, tried HDMI to DVI and HDMI to VGA adapters, also removed the battery to reset BIOS, did packet capture with my firewall and it’s basically dead(no action, just NIC network discovery that happens anyway), tried with and without RAM and/or SSD. I was really hoping you are going to hit the nail on the head.
I'm sorry to hear that you received a dud PC. Did you try reaching out to support at all? I also really wish I had found a way to fix it, would have been useful info for someone like you, I'm sure! It may be possible that the CPU didn't properly solder to the board at the factory in your case, considering it was brand new and exhibiting these same issues I found and deemed to be CPU related.
@@CHWTT I bought this one from AliExpress and they did refund me. On a side note though, I gave it another try with the multimeter today(continuity tests) and found 2 capacitors and 1 resistor in front of them not closing the circuit, right on the other side of the RAM connections with the board. I think I found the problematic resistor but it's so tiny that I definitely can't repair that, can't even put it on a soldering plate because it's filled with components.
Glad to hear you at least got your money back, that would have been a real bummer if you hadn't. That's cool that you picked it up and did some more troubleshooting! I totally feel you on how those resistors are too tiny to repair. I guarantee I would not have been able to do it in my case without a microscope, which is extra cost that isn't very worthwhile if you don't use it often.
I think the Beelink issue could be as simple as a corrupted Bios, remember that the bios itself is responsible for the whole startup procedure and if this gets broken the motherboard gets bricked too. either a flashback function or a chip programmer to re-flash the bios again would probably fix this computer. since the bios is powered even with the system turned off it could cause issues if you yoink the ram while in hibernation or the like and the system expects a different chip to be in place it may corrupt the data in bios. make sure the bios is actually completely reset or restored before you quit this one
Makes sense to me, I honestly just never thought of the BIOS even though it makes sense now that people are pointing it out. Won't be giving up on it until I have an opportunity to try reflashing it!
Given the lack of video support, it may have been wiser to have first attempted a headless Linux boot from a USB stick with serial USB terminal before going so deep.
I've noticed some of the Beelink machines (and other machines from the no-name Chinese vendors) have a thing, where they don't display for like 2-3 minutes on the first boot after you swap out their memory. Maybe it's a memory training issue? On mine, it made me think the mini PC was dead until I walked away from it to make some coffee and I found it at the BIOS splash screen when I got back.
On servethehome forums it's pretty clear that the common oems of CWWK & BKHD have bioses that only configure the ram by doing training. And with big dimms that takes a while. Changing ram seems to often need a bios reset (what we used to call a "cmos clear") inorder to force the training to rerun. I'm not clear if they work that way to cope with cheap in-China ram, or if there's a big bucks licence needed to be allowed to read the embedded configuration from the dimm. People have also had issues where their specific device hated certain hdmi cables or monitors. And more oddly that could fix itself if you ever got it to boot once using say display port, or got an OS to boot. So there may be interactions in the BIOS/EUFI/OS pathway. Thinking about this kind of thing, we had an Asus Zenbook laptop die last year - all we did was open it up (far too hard a challenge), swap the sodimm from 16 to 32, and the laptop never booted again! So modern tech is getting super fragile again.
For the BeeLink - Recently replaced SODIMM RAM on a laptop (HP). Power is never really “disconnected” and I had no issues. I know there are “power off” options in some laptop BIOS to, as best as possible, remove power but didn’t see one. From other comments from people with boards from the same company, could it be a flaw that they treated power as a desktop instead of a laptop, so that, with plugged in power, that left power on that fried the channel? I definitely think you’re on the right track. Also seems others had issues replacing boards other than RAM, so could just be a manufacturer issue too. Thanks for the great troubleshooting process!
Very thorough explanation and very simple to understand. Really admirable skills you've got there. I tried something similar with an old Lenovo ThinkPad Motherboard but Alas! never got it working. Keep up the good work!
Putting in a POST card and seeing what is completing and where it stops in POST will be the best way to go on these. I have done a number of these this way.
I have chatreey T8 Pro and bumped into the same situation. Sometimes it boots from the first powering on, but never come back after few hundres of power cycles, works again, and never turned off from now. I didn't know what was the actual issue and now I can see many people in here pointing out the corrupted bios, I now found the manufactuter provides bios update and tried... my problem actually solved. Thanks for the video. I was sick of it.
A good tektronics oscilloscope with a good 1x probe for interpreting signal quality and signal over time would tell you a lot more about what is and isn’t happening. Keep challenging yourself. Looking at things you’re not familiar with is the best way to learn. Don’t be scared, it’s already broken. Well done dude.
I bought a Beelink Mini PC. All was well, until it refused to work with a brand new SSD mounted in a USB enclosure. At first I suspected the SSD, but my Beelink's USB ports are very flaky, and connected drives malfunction. Beelink were no help at all! Amazon gave me a full refund, and let me keep the Beelink. I then bought a Minisforum, which unlike the Beelink, works perfectly with all my USB drives! I like Mini PCs due to their tiny form factor, and their highly efficient energy usage, compared to my old home built desktop! Also, this allowed me to upgrade to Windows 11, love it or hate it, for very little money!
Something many troubleshooters seem to forget to try is boot with no memory. The system will of cause detect it and will beep (if there is a sounder) or flash a code on the system led's. If it doesn't do that its stuck before memory and that is CPU or before bios test starting AKA power problems or firmware.
It's the identical driver circuitry, but it's connected to a different rail output on the controller chip, so I think it would likely be a different voltage. 0.85V wouldn't be a bad place to start though, as it's almost certain to not fry something
@@CHWTT Question, I have a laptop that does not work and I want to rule out the power brick. I have no idea about electrical engineering. How did you probe that power cable's voltage, I know you used a multimeter but what settings?
For sure! I set my multimeter to DC Volts, and my multimeter is an auto ranging one. Since most laptops don't use a higher than 19V supply, if you don't have an auto ranging multimeter, I'd select the "20" option within the DC Volts setting. I then put the red (+) probe in the hole in the barrel jack and measured the negative pole by touching the black probe to the outer bit of metal. If your barrel jack is of the less common type that has a pin in the center of it, touch the red probe to that pin and the black probe should still go on the outer sleeve. Hope that helps!
Totally agree, Carey is one of the guys that really got me into PCs and building them custom. Watched a ton of his content, and was super thrilled when he reached out!
My Beelink just got back to life after was sitting disconnected about 2 months, wasn't posting, just turning on with no image like in the video so i packed it and forgot about it, now after it's been a while i was checking to see if there any new solution, i got you your video, i already tried most of the things you did but it didn't work so i figured i'll try once more to connect it - BOOM - Works! Go figure...
A lot of point of sales have gone to these compact units, but with PS/2 compatibility for legacy systems with sVGA, but often there's HDMLI ready to go. Our main issue is configuring it to use whatever the site's needs are and shut off the rest until the next service upgrade.
This video is 😂 funny, in a good way. When the resistor flew away and when he ripped the pads for the resistor. I hope you keep making videos and learning computer repairs. Also they sell very cheap memory tester to test open and shorts from memory to CPU
If you suspect that the iGPU are broken, maybe you can buy adapter for M.2 to Oculink adapter, and then connect external GPU using that oculing with external PSU.
POST order of checks is pretty different on all the hardware out there. I think your general thoughts are right about the order, but I've seen modern boards with different support for special hardware or just different OSes, that would change the POST order behaviour. Allways guessed there was a general order tree, but in special scenarios it would traverse other order than the actual tree order. And also: Very nice video!
Thanks for the compliment! And that's a good point about it possibly being a bit of a special board that doesn't go in a more common POST order. I wonder how one would go about figuring out the POST order a specific board uses
I love those Beelink PC's. I have four of them. They are all AMD CPU based. The oldest one will sometimes "forget" which SSD the OS is installed on and you have to boot to the BIOS and select the correct drive. Other than that one issue they work great.
I understand the reason for going to all that trouble is for the purpose of doing a video about how to repair one of these mini pc's, but you made the most important point at the end of the video: they are so cheap that it's worth chucking a dead one in the bin and buying a new one to replace it. I have had good luck with MinisForum, and I have a Xulu hopefully coming sometime this year from a Kickstarter project. Fingers crossed on that one. Beelink, though, seems like its reliability is not all that great. Great video, and you showed why these mini-PCs are a bit of crap shoot.
I was writing something as you were explaining your theory. By the end, i deleted it. The board is cooked. You could try an external GPU unit to see if you get a bios but thats not a certainty
Tip to try, does the light or fan reacts different when the RAM isnt seated? Like longer or shorter maybe.. You didnt own DDR5 but you also never checked if that could be it, take out the ram, see if the PC reacts different, theres your possible answer, another part to this answer, order the cheapest compatible ram stick for this device (check online if there's a CVL for this board) and write that $30 as repaircosts if it works or doesnt
the serial port, jtag, spi bus, and post cards are your friend. try putting the beelink in a reflow oven, could be stuck transistors or a dry solder joint.
You're almost their , Check the IC that controls the HDMI if power is their then yes you have a dead super IO and will need to be replaced using hot air rework, but if your not confident in micro soldering I would not attempt. But your close ,Good job troubleshooting
If the 4th phase was not controlled or broken, the fans would not spin. The 4th phase can also be the turbo booster, which is normally turned off. The 1.8V phase is probably the PCH, not the IGPU, although the GPU is supposed to be in the PCH of the N100. Test all your phases in ohms mode. If they are not close to 0 ohms, say 20 ohms, they are fine. If they are lower, you can confirm that the CPU is dead by applying voltage to the phase (0.5 V, 500 mA) with some alcohol on the DIE. If the whole area gets hot, the CPU is fine. If one corner lights up, the DIE is broken. My guess is that the N100 is fine. Instead, I suggest you focus more on the 3.3V PCH rail, as this is likely to be the problem. Test pin 8 of the bios SPI IC, which should bring 3.3 V. In most cases in these small PCs, the ram SPD SPI IC is in parallel to the bios SPI IC, both powered by the 3.3 V rail of the PCH. My guess is that the ram swap triggered voltager on the SPI lines of the bios IC (through the rams SPD / SPI). Try dumping the bios and open it in the UEFI tool and look for any errors. Also reflash the bios with a clean (ME) dump. Since the pc turns on shorly and the fans spin, my guess is that it 's somehow bios / SPI related.
I would never pay $100-$200 to repair a $200-$300 (not incl RAM, storage) PC! If I can not get it working, it's a PCB issue and not worth paying someone to fix an old turd.
Too bad they don't have adequate protections for the CPU - thanks for the attempt! I've seen multiple of these Beelink N100's fail - so will be avoiding these.
You are probably right about the BeeLink. If you wanted to test further, you could see if you could get an adapter to hook up a PCIe video card to the M.2 slot. I doubt it would yield any results other than more confirmation.
Fascinating! I am most certainly nowhere near as smart as you, but I had a thought about the remaining dead one! I have a similar issue with a NUC of mine, essentially dead beyond the power light and fan. but I noticed that trying to boot it with the RAM removed, the power buttons light had changed, AND with speakers plugged in to its headphone jack I DID get a beep! so there is some kind of a status check still operating! I'm planning on getting a thermal camera so I can see if anything is abnormally hot when attempting to boot, RAM installed or not. Could be helpful! Great video pal, keep up the good work! 😁🧡
Thanks for the kind words man! That's really interesting that you got beeps out of the headphone jack, I'd never thought of that myself! I'll definitely be giving that a try when I revisit it, thanks!
Since I am mainly a software guru, I definitely can't help with the hardware part of things. I will give some insight on the software side of things though. Mini PC's are annoying to fix because the integrated bios chips are poorly designed and a hassle to diagnose. Although, I have gotten myself many times into a situation where digital video output, like HDMI or DP wouldn't work until the OS boot phase, after the communication protocol between the GPU and display has been set. If I were to restart the computer, the display would work just fine, outputting the post message and everything, until a complete power down happened. Then, I would have to repeat the blind process of selecting the boot drive from the boot menu. I believe this could be the problem in your case, so try plugging in a keyboard first. It is possible that an "input device missing", or "system configuration change" (since the RAM was changed) prompt could stop the boot process and therefore not completing the display output handshake. Also, have a bootable drive connected to the computer and see if there are any signs of life. These solutions helped me multiple times, especially on crummy intel atom devices.
6:13 It's actually rather simple. Everything which isn't carrying data, is carrying current, therefore anything attached to a current rail is there to manipulate it. Since in electronics there are only a few principles of altering the current and voltage albeit in a billion combinations (like series or parallel arrangements, RC LC RLC configs, etc), you can safely deduce that whatever is on a power rail, is there to either drop voltage, increase it, drop current, increase it, or lastly, to filter it in some manner. Any congregation of elements are there to do most of those things. What we are interested in, is whether the voltages meet the power requirements of the components they're plugged into, so we read the spreadsheet for the component and go from there. You look at the voltage drops, but it's worth looking into the IC power inputs as well, then diagnose backwards.
The NIC did light up when a cable was connected, though I feel like I've heard somewhere that this can just be the NIC doing its own thing rather than a major sign of what's going on with the rest of the system. Still something to look into, though!
@@CHWTT next thing to try then is hook it up to a PC running Wireshark and see if it's trying to connect to a DHCP server. If so, you definitely have a running OS.
You need to invest in a hot air rework station for dealing with surface mount devices. Also look for tin whiskers when looking at the board when looking at it through the microscope.
Is there oxidization or metal eaten away on the RAM connector pins? RAM unplugging and replugging while the system is on could make some pretty little sparks. Deoxit and if that doesn't work, soldering in a new RAM slot might be an interesting thing to try. Also, something to consider, if there were pretty sparks created, then maybe it could be physical slot damage due to human jumping.
"Finding schematics for __THIS__ computer was never going to happen" - i like the approach :), sad but __TRUE__ ;) Thanks for pointing that "CMOS" reference - cmos technology was a great break in the 70s but nobody now remembers how great it is, i dont even remember when have it started.
Very interesting ! I'm having the same issue with a Fitlet2 Intel mini pc board. After changing something in the bios the screen is dark. I think a bios problem. So keep trying to fix it !
14:04 btw you do not need to worry. you can just leave these dies "empty" when testing. just make sure to put a finger on it occasionally to feel if it's overheating or not. these modern cpu's are so safe you do not need a heatsink most the time lol. theres a video of someone pointed a thermal imager onto a ryzen CPU's bare die with some program stressing each core.
Ahh okay, good to know it's less of a massive deal than I imagined. I was aware that the CPUs are pretty good at protecting themselves from overheating by shutting down/throttling, I just wasn't sure whether that would happen so fast with a bare die that it could interfere with POST
The Beelink one can be the same problem from Xulu Mini PC, do a checkup arround the Power Circuit and this same resistor from the Power Circuit is the frist thing to do on this board. Edit: I noticed two corroded SMD Capacitors behind the memory slots, this may be the problem source.
If the Blink system supports a direct flash BIOS image it might be worth while to see if maybe it got corrupted and just needs to be flashed with a working image. It's logical that something like that would happen if the BIOS were active because their normally connected to the memory in that the BIOS image get's written to memory when the system boots. It does this because the RAM is generally faster than the BIOS chip. At least that was the case for older systems. Point of the matter is this. Attempting to flash a good image to it might be worth a shot. Like you said, "it's already broken."
While putting the RAM in while the power was on it is possible a circuit got more power it than was designed for and the chip holding the CMOS or BIOS got zeroed out which means the computer has forgotten how to turn on (POST). Install a new CMOS chip or use the box for parts.
your theory makes sense about the blink pc (one you troubleshooted first ddr5 ram that had
7 หลายเดือนก่อน
if the coil is completely shorted to ground, forget about it, but since it was operated on while the pc was powered, there is the easiest thing to do with high probability of fix, reflash the bios
Gonna try the BIOS at some point for sure, and the dead phase's coil had about 32 ohms on it so I don't think it was shorted in this case, but definitely something I checked for that reason!
See if there are any TVS diodes near the RAM connector that look compromised. It's possible that hot-plugging the RAM with the PC on was enough to blow one of them short. Conveniently enough most integrated graphics solutions use the system RAM as their VRAM. It's possible that a short (or partial short since it's not drawing an insane amount of current) may be causing a fault in this power rail shared between the iGPU and system RAM which now won't turn on as a protection mechanism. You can probably check this pretty easily first by measuring the resistance of this power rail to ground. Anything less than a few kOhm would be pretty suspicious. A bit of an out there theory but something worth checking if all else fails.
I saw someone do video fixing a graphics card and they just took off or connected across one of those square things and he said the cpu or something figures it out.
On the Zulu, that resistor is likely acting as a fuse. Most times they are very small values. Less than 1 ohm, or maybe even zero. Shorting with solder is fine for test, but putting in a 0 - 1 ohm link is preferred.
I guess the only thing I could think off that can be done for the n100 is to power on the PC without ram installed and check for heat/power draw differences and hope it's a case of multiple ddr5 modules being bad/damaged by the powered scenario.
Getting all the voltages on the various rails is a good way to diagnose some possible problems. Don't forget to also find the on board fuses and 0 ohm resistors between the voltage rails and the actual power using chips. You have found a few, but you did not explicitly say in the video you checked all. You may have done so off camera?
Well, yes, a CMOS battery that's dead will often make the BIOS reset to defaults, you're right there. However, I've had several systems such as OptiPlex 7010s and 7020s which entirely refuse to POST and spit out a motherboard error if the CMOS battery is dead, so it most certainly can cause POST issues.
for the pc that has the graphics problem, couldn't you use a external graphics card setup, you can get external graphics card dock, maybe that would fix that problem.
I am quite convinced that the resistor was only a 0 Ohm resistor used as a jumper. The VGFX line is only active when the integrated graphics card is loaded. Otherwise, the entire CPU is powered by VCORE. It's going to be tense there. S No video output can be a bad ME region in BIOS. You need clean this part of bios but in modern pc is not that simple.Try measuring the resistance on pin 8 of the bios chip. Sorry for my english.
i somehow broke the bios on an intel nuc, that was set up with ubuntu linux. i had the task of adding a sata ssd and setting it up for dualboot. but then i had to reboot a hundred times through tangled threads and corporate crimes because windows by default holds the pc hostage until configured properly. at some point i force rebooted it while it was on it's way to boot into windows again. that's when it broke because after that, the system wouldn't post anymore. no display, no beep, nothing. i found that there is a way to update the bios automatically with an FAT formated USB stick, containing nothing but the bios file and i think i had to long press the powerbutton or something to put it in update mode. because there was nothing on the display, i did not know what it was doing but otherwise the system acted like it should according to the online guide provided by intel. the guide stated that this process could take like half an hour, but i let it sit for more than 4 hours. there was no indication that it was done, but after waiting that long it's either done or dead. i rebooted and the monitor came on again and it booted into ubuntu like normal. that was not the first thing i have tried. that pc was like almost a month at my house for an upgrade that was supposed to take a few hours tops, fml.
Thanks for sharing that experience, I find it quite interesting! I'll be giving a BIOS flash a shot on the Beelink at some point, I'll have to see if it has an automatic BIOS reflash feature or if I'll have to get an external programmer and flash the chip itself with another PC. Should make an interesting video!
@@CHWTT it works with modern mobo, but sometimes that process fails... if you repeat it a couple of times, it should work... and yes, it can take a lot of time (up to an hour)
Please send the broken machine on to another youtuber! Th3c0der (or however it is spelled) in the UK would be able to replace the CPU. Not that there is commercial value to doing that, but having him swap it out in a video will be good content, collaboration and it might just prove your diagnosis correct.
I have a question that is a bit off topic. The mat that you are working on, the gray mat with a grid pattern and the different sized circles on it. Do you have a link for it? I'd like to know more about it. I'm just starting to delve into this "kind of stuff" myself, and I am still getting my work area set up. That mat looks like it would be a nice area to do my work on. Thank you and greetings from Alaska.
Bad RAM might stop you from getting a post screen, but zero RAM shouldn't, should have tried booting with no RAM. And with it plugged in & plugged into the network, check it you saw it on your router. To me only 3 minutes in I'd guess the i/o was blown, meaning half the m/b was dead. But seeing if the Ethernet port was blinking would tell me it most likely is still good.
I've never had a system POST without any RAM, and I don't think it would have any chance of working in this case, because of the fact that the iGPU in this system requires system RAM to work
They must have a way of testing those boards during manufacture. To that end I saw several test pads ready to take spring probes in a jig. What if you try any of those pads with a scope. Is there any lilfe in them?
I'm not so sure that changing the memory had anything to do with it. I bought 2 Beelinks, an EQ12 for a router and a SER5 Max for mostly emulation gaming (I'm old) and some older titles. The Ser5 Max died after 3 days in the exact same way that your friend's EQ12 did. I read somewhere that it is a BIOS issue (can't find it now) . I'm convinced that it's software and/or BIOS. Neither of these computers would update past 21H2 out of the box. It is blocked by Windows Update. Beelink acknowledges it on their forum but apparently isn't doing anything about it. There is some reason for that. The Ser5 died after I attempted to force it to update. I don't know if that's the cause, but it was late, I forced the update with a USB stick and an in-place upgrade. I played around with it a bit, I shut it down, went to bed, and it never booted again. Same issue as you're seeing. I pulled it apart, put the SSD in an external case, wiped it, packed it up and sent it back to Amazon with a note. I wasnt leaving my personal info on it. The EQ12 won't be running Windows but I wanted to make sure that it was OK. When it wouldn't update, I found the drivers, dug the key out of the Registry and attempted to do a clean install. It wouldn't let me. I couldn't get past the MS account screen. There isn't anything that you can just click and bypass anymore but you can hit F10 and type a command. That didn't work. F10 did nothing. I gave up and attempted to sign in to a secondary MS account. MS blocked it. Beelink has done something to their Windows images that Microsoft really doesn't like. I don't know if that's causing a hardware issue but it's still another thing wrong with these Beelinks. I replaced the SER5 Max with a GMKTEC M7. Much better. All the drivers are available. I did a clean install that went smoothly. The key activated OK. I definitely recommend the GMKTECs. I'll get another if the second Beelink dies.
I might have accidentally said that the system was powered on when the RAM was changed during the live audio portion of the video - if I did, I meant to say that it was plugged in. The system was not powered on. My bad there!
At 21:05 you state: "the thing that most likely caused it to be broken by that procedure was the fact that Carrie left it plugged in while he did the ram swap"
Plugged in, not powered on? Regardless, there are many power states to modern computers. I think we have a tendency to attribute causes that we are familiar with. And therefore I think you are searching for an electrical problem. Have you considered a firmware problem, i.e., that the BIOs is corrupted? As the RAM was removed while the computer was in an unconfirmed power state, and multiple OS levels store data in firmware that is implicated in POST, flashing the BIOs would be a good step to try! You can likely fix it with a USB Bios Recovery, even with Chinese Mini PCs, as they are based upon the Intel (now ASUS) NUC platforms.
Enjoyable video. I liked seeing you trace the electronic paths.
@@DustinKleiboer That's a good point about the BIOS, and me focusing on an electrical issue. It hadn't crossed my mind whatsoever, but you're totally right about it possibly being the BIOS
@@DustinKleiboer was an ISO approved grounding strap-on worn?
@@DustinKleiboer True. Shutting down a windows system does not power it down, unless you held the shift key while clicking "shutdown"
you will have to track back to where that phase is getting an enable signal. (nightmare without schematics) a good channel to watch for that is northwest repair. GPU's use alot of phase enable sequencing. In the end i suppose the lesson is inplug and hold the power button in a moment, before changing components and any system. remember off IS NOT unpowered.
The old saying: It's never the CPU except when it is.
Yes
the CPUs are the toughest electronic component on any board.
@@jimw7916 Agree hence my comment.
@@colinreece3452 agreed.... Hence MY comment.
@@jimw7916🤣
Beelink's problem is 99.9% a BIOS firmware failure. This type of Chinese semi-finished products has catastrophically problematic BIOS firmware. The HDMI output will not work until the OS boots and therefore no communication will be established between the GPU and the display. Try to copy, save and analyze the BIOS firmware, and it will be very good if you can find an “engineering bios”, try to find a similar Beelink and copy the BIOSa firmware and write a new, working firmware to your BIOS chip. Good luck!
Thank you for confirming my suspicions that arose as I listened to the diagnosis. I've become aware over time of how prevalent BIOS firmware failures are.
yeah a lot of new UEFI firmware especially on some laptops store settings in the same flash chip as the bios. so simpyl detaching battery won't reset bios defaults. change any setting wrong to make your system unbootable and you're fuxked, cause almost no laptops have a "cmos" (more like nvram??) reset button.
had this happen on a laptop once, thankfully I had a full flash backup, flashing that also restored my working settings.
"The HDMI output will not work until the OS boots" - that's just false. HDMI is handled by the iGPU and it works on Beelinks exactly the same as it does on any laptop or desktop.
One important thing is that if there is more than 1 HDMI out, only the primary one will be turned on initially. Figuring out which one is the primary is a fun exercise when nothing works.
@@GizmoTheGreen Had this issue with a craptastic Machinist x99 k9. Adjusted RAM timings, would no longer boot. BIOS battery removal didn't reset the bios, which is a known issue for many of the crap x99 boards. Solution was to buy a flash programmer and write the stock BIOS to the chip.
That is true , most of the problem of beelink from bios🤦
Wow I am so so impressed! Carey's channel administrator here - just had to say I so admire your skill, ability to detail all the steps and the fantastic work you did! Just so so good! Fantastic walkthrough and explanation!
Thank you Maralina! I really appreciate the kind words!
But who is Carey? - google user id hash ?
@@TymexComputing not sure if you're kidding - see video notes for Carey's channel
@@MaralinaDG Oh sorry - i havent registered that these machines were sent from Carey - now i know thanks :)
@TymexComputing gotcha much appreciated and no worries!
A bit of information for you. The resistor that you replaced with that radial then bridged was probably .02ohms used as a current sense to check that the power button is pressed. Don't worry too much about the ripped pad happens to all of us. Good on you giving it a go. I wonder if the cap right next to that resistor went short and took the resistor out with it before it burned. That could potentially explain the odd solder and the burnt flux assuming someone else hasn't had a soldering iron on it. For the beelink you are probably right that the CPU is cooked although I could be wrong but typically voltage isn't present on the ram voltage rail with the system off. I have never worked on one of these mini-pcs but typically in laptops and desktops the ram voltage isn't a standby voltage meaning it should only be live when the PC enters a powered state.
That's interesting and makes a lot of sense with the resistor! Would you say that's something that should be replaced with a properly valued component or is it something that doesn't matter all that much?
I tested the capacitor that was hanging out next to it and as far as I'm aware, it was fine so I think it might have just been the resistor that went. That or someone at the factory had a go at it, as Carey doesn't really do board level repair work. Good point on the RAM voltage as well, I'm pretty sure that the RAM voltage rail in this case was live only when the system was up, though I'll have to double check that when I pull it back out at some point.
@@CHWTTWith the state of that pad I probably wouldn't mess with it. If it is working then it should continue working.
I suspect that the whole ram part might be a red herring. I think maybe if some damage did get done it was done on the board itself not from pulling the ram out. Diagnosis of dead CPU might be correct either way although I would probably look at that IC that drives the core voltages. Usually these aren't too expensive and if nothing else this would be a good chance to test out your hot air rework skills after you mess around a bit with other components.
Definitely going to look at some more things on the board in the future, somewhat apart from the RAM issue as it definitely could be a red herring as you said. I was considering going and replacing the power IC on this board, especially because I was also thinking it would be maybe a few dollars at most. However, I couldn't find that chip for sale anywhere! I may have to contact the manufacturer and see if I can get one from them, though I feel doubtful they'll entertain a random guy's request to buy a single chip.
It's almost certainly a 0ohm link.
I am quite convinced that the resistor was only a 0 Ohm resistor used as a jumper. The VGFX line is only active when the integrated graphics card is loaded. Otherwise, the entire CPU is powered by VCORE. It's going to be tense there. S No video output can be a bad ME region in BIOS. You need clean this part of bios but in modern pc is not that simple.Try measuring the resistance on pin 8 of the bios chip.
When I was fixing circuits we would use through hole resistors to replace SMD as a POC all the time. Great job.
That's cool to hear, thanks!
try desoldering the bios , flash it outside of system .
if the vendor supplies a full bios image. some new formats are upgrades only and need to be applied by their flasher onto the existing bios copy. also lots of new uefi bios has device unique keys that you need to dump, and thats if you know where to look and might require sketchy or hard to get softwares...
@@GizmoTheGreen where do you find hard to get softwares? forums?i am sure if i was inclined to know i would get them or locate them the same with all illicit things i suppose.
Try this: Connect a keyboard directly to the machine (no kvm switch). And when you turn it on try to press del, esc, f7, enter or other keys. The f7 is the one for entering bios in normal situation.
In similar situation, After some trying, I noticed the keyboard started showing life and display also turned on
Some of these mini pcs have bad bios and they does not turn display on boot screen even if they "show" a message and waiting for input to continue.
I'll definitely give that another shot, it didn't make the cut but I did give that a go (though I didn't try the F7 key)
I have two mini-PCs and boy, there's a reason that they put a BIOS Reset button on them. With the unlocked do-whatever-you-want BIOS, you can do stuff that you shouldn't have access to, like enabling PEG graphics...
(I don't think this applies to your case unless the BIOS really gets stuck without battery power)
what does peg graphics means and what can it cause?
@idan678 PCI Express Graphics (i.e., discrete / external GPU)
I should have written "Set Graphics Priority to PEG" because it's not something you enable but a selection menu (also, exact wording change depending board manufacturer).
It allows the user to select what graphic card should be used and have options like Auto, IDG (Internal Display Graphics), and PEG.
This setting when set to PEG will ignore the internal graphics at boot.
@@TommyCrosbyYou can actually attach PCI Express graphics cards to these mini PCs through an adapter for the M.2 slot. I have such a setup on my desk at the moment. I wonder whether that setting in the BIOs would actually do anything if you had such a setup
If it were plugged in and the ram was changed without draining the power first, integrated graphics uses your system ram< 31:00 (shared) I can understand how that would effect the other.
If you have a thermal camera, check the board with the power on to see if anything is overheating. If you don't have a thermal camera you can put alcohol on the board and see if it evaporates quickly over a specific chip to see if there's a short. Northridge fix has videos on this method.
I was thinking about doing that isopropyl method (can't afford a thermal cam... wish they were cheaper!), but the resistances around the board and even on the dead rail all seemed like they were right, so I don't think there were any shorts in this case
Yeah, do NOT screw around with anything on these mini PCs with the power connected. I fried one of my two Intel NUCs by doing basically the same thing as the guy with the Beelink did, try to change memory with the power connected. Luckily the one I fried was a cheap one. Exact same symptoms as the beelink.
For debugging it's worth while investing in a post card, that way you'll know exactly where the machine fails to post.
It's a PCIe card but you can use an m.2 to PCI adapter and plug it into the wifi slot
I usually add a tiny pc speaker to hear any POST codes bips... it helps diagnose at what stage it fails.
@@armirol not all motherboards have a speaker pin unfortunately, this is where a post card comes in handy
@@R4MP4G3RXD I haven't experienced it yet, I don't even own one of those disgnostic module.... i always try to connect that tiny speaker and it worked for me so far.
I bought one of these N100 minipcs for my parents and it came dead, very similar issue to the Beelink. I did not manage to fix it either: tried similar Voltage(and Amps - not that it matters) charger, probed some resistors and no issue there, tried HDMI to DVI and HDMI to VGA adapters, also removed the battery to reset BIOS, did packet capture with my firewall and it’s basically dead(no action, just NIC network discovery that happens anyway), tried with and without RAM and/or SSD.
I was really hoping you are going to hit the nail on the head.
I'm sorry to hear that you received a dud PC. Did you try reaching out to support at all?
I also really wish I had found a way to fix it, would have been useful info for someone like you, I'm sure! It may be possible that the CPU didn't properly solder to the board at the factory in your case, considering it was brand new and exhibiting these same issues I found and deemed to be CPU related.
@@CHWTT I bought this one from AliExpress and they did refund me.
On a side note though, I gave it another try with the multimeter today(continuity tests) and found 2 capacitors and 1 resistor in front of them not closing the circuit, right on the other side of the RAM connections with the board. I think I found the problematic resistor but it's so tiny that I definitely can't repair that, can't even put it on a soldering plate because it's filled with components.
Glad to hear you at least got your money back, that would have been a real bummer if you hadn't.
That's cool that you picked it up and did some more troubleshooting! I totally feel you on how those resistors are too tiny to repair. I guarantee I would not have been able to do it in my case without a microscope, which is extra cost that isn't very worthwhile if you don't use it often.
I think the Beelink issue could be as simple as a corrupted Bios, remember that the bios itself is responsible for the whole startup procedure and if this gets broken the motherboard gets bricked too. either a flashback function or a chip programmer to re-flash the bios again would probably fix this computer. since the bios is powered even with the system turned off it could cause issues if you yoink the ram while in hibernation or the like and the system expects a different chip to be in place it may corrupt the data in bios. make sure the bios is actually completely reset or restored before you quit this one
Makes sense to me, I honestly just never thought of the BIOS even though it makes sense now that people are pointing it out. Won't be giving up on it until I have an opportunity to try reflashing it!
+1 on this. It's easy to reflash the bios and simple to rule this out.
Put a terminal based linux on an ssd and try to access the system remotely to check if the igpu took down the whole cpu or just the igpu part.
Now THAT is a great idea! Imaging a drive with some headless Linux install and trying to SSH into it or something is actually genius, thanks for this!
@@CHWTT Is the drive activity LED turning on at any point?
@@CHWTT did it work?
Given the lack of video support, it may have been wiser to have first attempted a headless Linux boot from a USB stick with serial USB terminal before going so deep.
I've noticed some of the Beelink machines (and other machines from the no-name Chinese vendors) have a thing, where they don't display for like 2-3 minutes on the first boot after you swap out their memory. Maybe it's a memory training issue? On mine, it made me think the mini PC was dead until I walked away from it to make some coffee and I found it at the BIOS splash screen when I got back.
On servethehome forums it's pretty clear that the common oems of CWWK & BKHD have bioses that only configure the ram by doing training. And with big dimms that takes a while.
Changing ram seems to often need a bios reset (what we used to call a "cmos clear") inorder to force the training to rerun. I'm not clear if they work that way to cope with cheap in-China ram, or if there's a big bucks licence needed to be allowed to read the embedded configuration from the dimm.
People have also had issues where their specific device hated certain hdmi cables or monitors. And more oddly that could fix itself if you ever got it to boot once using say display port, or got an OS to boot. So there may be interactions in the BIOS/EUFI/OS pathway.
Thinking about this kind of thing, we had an Asus Zenbook laptop die last year - all we did was open it up (far too hard a challenge), swap the sodimm from 16 to 32, and the laptop never booted again! So modern tech is getting super fragile again.
For the BeeLink - Recently replaced SODIMM RAM on a laptop (HP). Power is never really “disconnected” and I had no issues. I know there are “power off” options in some laptop BIOS to, as best as possible, remove power but didn’t see one.
From other comments from people with boards from the same company, could it be a flaw that they treated power as a desktop instead of a laptop, so that, with plugged in power, that left power on that fried the channel? I definitely think you’re on the right track. Also seems others had issues replacing boards other than RAM, so could just be a manufacturer issue too.
Thanks for the great troubleshooting process!
Very thorough explanation and very simple to understand. Really admirable skills you've got there. I tried something similar with an old Lenovo ThinkPad Motherboard but Alas! never got it working. Keep up the good work!
Thank you for the kind words!
Putting in a POST card and seeing what is completing and where it stops in POST will be the best way to go on these. I have done a number of these this way.
Is there a M.2 POST card? Or how are you plugging it into a Mini PC?
@@xumatyt m.2 is pcie, can always use an adapter
I have chatreey T8 Pro and bumped into the same situation. Sometimes it boots from the first powering on, but never come back after few hundres of power cycles, works again, and never turned off from now.
I didn't know what was the actual issue and now I can see many people in here pointing out the corrupted bios, I now found the manufactuter provides bios update and tried... my problem actually solved.
Thanks for the video. I was sick of it.
You can sort of test the mem CTRL in cpu, put + probe on ground and with negative probe go over ram datalines
But its like 98% that cpu is done
I might have to give that a shot, even if it's very likely it's the CPU
A good tektronics oscilloscope with a good 1x probe for interpreting signal quality and signal over time would tell you a lot more about what is and isn’t happening. Keep challenging yourself. Looking at things you’re not familiar with is the best way to learn. Don’t be scared, it’s already broken. Well done dude.
I bought a Beelink Mini PC. All was well, until it refused to work with a brand new SSD mounted in a USB enclosure. At first I suspected the SSD, but my Beelink's USB ports are very flaky, and connected drives malfunction. Beelink were no help at all!
Amazon gave me a full refund, and let me keep the Beelink. I then bought a Minisforum, which unlike the Beelink, works perfectly with all my USB drives!
I like Mini PCs due to their tiny form factor, and their highly efficient energy usage, compared to my old home built desktop! Also, this allowed me to upgrade to Windows 11, love it or hate it, for very little money!
beelink are turning out to be unreliable in the long term.
i've got many tiny Beelink PCs and never experienced bad one... those are pretty good. I guess it's bad luck
Something many troubleshooters seem to forget to try is boot with no memory. The system will of cause detect it and will beep (if there is a sounder) or flash a code on the system led's. If it doesn't do that its stuck before memory and that is CPU or before bios test starting AKA power problems or firmware.
Introduce a signal to the igpu power driver to jump start it.
If it starts there’s something in the power circuit
If it doesn’t the igpu is dead
🤷🏻♀️
I could definitely give that a shot! I just don't know what voltage would be good to inject there, any ideas?
@@CHWTTis the driver the same on the vcore drivers? I’d mimic that input if they’re the same.
It's the identical driver circuitry, but it's connected to a different rail output on the controller chip, so I think it would likely be a different voltage. 0.85V wouldn't be a bad place to start though, as it's almost certain to not fry something
Wow! Superb job! I did not understand much of what you said towards the end of the video, but it was fun to watch either way!
Glad you enjoyed it!
@@CHWTT Question, I have a laptop that does not work and I want to rule out the power brick. I have no idea about electrical engineering. How did you probe that power cable's voltage, I know you used a multimeter but what settings?
For sure! I set my multimeter to DC Volts, and my multimeter is an auto ranging one. Since most laptops don't use a higher than 19V supply, if you don't have an auto ranging multimeter, I'd select the "20" option within the DC Volts setting. I then put the red (+) probe in the hole in the barrel jack and measured the negative pole by touching the black probe to the outer bit of metal. If your barrel jack is of the less common type that has a pin in the center of it, touch the red probe to that pin and the black probe should still go on the outer sleeve. Hope that helps!
@@CHWTT It does! Thank you!
did you try unplugging it and plugging it back in? ( sorry had to lol ) great video.
Carey is a legend, one of the OG guys I learned how to build a PC from way over a decade ago back when TH-cam was a very different platform.
Totally agree, Carey is one of the guys that really got me into PCs and building them custom. Watched a ton of his content, and was super thrilled when he reached out!
My Beelink just got back to life after was sitting disconnected about 2 months, wasn't posting, just turning on with no image like in the video so i packed it and forgot about it, now after it's been a while i was checking to see if there any new solution, i got you your video, i already tried most of the things you did but it didn't work so i figured i'll try once more to connect it - BOOM - Works!
Go figure...
A lot of point of sales have gone to these compact units, but with PS/2 compatibility for legacy systems with sVGA, but often there's HDMLI ready to go. Our main issue is configuring it to use whatever the site's needs are and shut off the rest until the next service upgrade.
This video is 😂 funny, in a good way. When the resistor flew away and when he ripped the pads for the resistor. I hope you keep making videos and learning computer repairs. Also they sell very cheap memory tester to test open and shorts from memory to CPU
Im amazed to see u already have sponsors at that small of a channel size!
If you suspect that the iGPU are broken, maybe you can buy adapter for M.2 to Oculink adapter, and then connect external GPU using that oculing with external PSU.
POST order of checks is pretty different on all the hardware out there. I think your general thoughts are right about the order, but I've seen modern boards with different support for special hardware or just different OSes, that would change the POST order behaviour. Allways guessed there was a general order tree, but in special scenarios it would traverse other order than the actual tree order. And also: Very nice video!
Thanks for the compliment! And that's a good point about it possibly being a bit of a special board that doesn't go in a more common POST order. I wonder how one would go about figuring out the POST order a specific board uses
I love those Beelink PC's. I have four of them. They are all AMD CPU based. The oldest one will sometimes "forget" which SSD the OS is installed on and you have to boot to the BIOS and select the correct drive. Other than that one issue they work great.
You did your best brother and accepted defeat gracefully. Well done.
I understand the reason for going to all that trouble is for the purpose of doing a video about how to repair one of these mini pc's, but you made the most important point at the end of the video: they are so cheap that it's worth chucking a dead one in the bin and buying a new one to replace it. I have had good luck with MinisForum, and I have a Xulu hopefully coming sometime this year from a Kickstarter project. Fingers crossed on that one. Beelink, though, seems like its reliability is not all that great. Great video, and you showed why these mini-PCs are a bit of crap shoot.
I was writing something as you were explaining your theory. By the end, i deleted it. The board is cooked. You could try an external GPU unit to see if you get a bios but thats not a certainty
Use an m.2 (x4) to pci-e adapter and an external video card.
Tip to try, does the light or fan reacts different when the RAM isnt seated? Like longer or shorter maybe.. You didnt own DDR5 but you also never checked if that could be it, take out the ram, see if the PC reacts different, theres your possible answer, another part to this answer, order the cheapest compatible ram stick for this device (check online if there's a CVL for this board) and write that $30 as repaircosts if it works or doesnt
He definitely did try booting with and without the SODIMM stick installed.
the serial port, jtag, spi bus, and post cards are your friend. try putting the beelink in a reflow oven, could be stuck transistors or a dry solder joint.
Its never ever the CPU. Until it is. Side note, I love those little Beelink PCs. I have two! They work great for home office type environments.
You're almost their , Check the IC that controls the HDMI if power is their then yes you have a dead super IO and will need to be replaced using hot air rework, but if your not confident in micro soldering I would not attempt. But your close ,Good job troubleshooting
If the 4th phase was not controlled or broken, the fans would not spin. The 4th phase can also be the turbo booster, which is normally turned off. The 1.8V phase is probably the PCH, not the IGPU, although the GPU is supposed to be in the PCH of the N100. Test all your phases in ohms mode. If they are not close to 0 ohms, say 20 ohms, they are fine. If they are lower, you can confirm that the CPU is dead by applying voltage to the phase (0.5 V, 500 mA) with some alcohol on the DIE. If the whole area gets hot, the CPU is fine. If one corner lights up, the DIE is broken. My guess is that the N100 is fine. Instead, I suggest you focus more on the 3.3V PCH rail, as this is likely to be the problem. Test pin 8 of the bios SPI IC, which should bring 3.3 V. In most cases in these small PCs, the ram SPD SPI IC is in parallel to the bios SPI IC, both powered by the 3.3 V rail of the PCH. My guess is that the ram swap triggered voltager on the SPI lines of the bios IC (through the rams SPD / SPI). Try dumping the bios and open it in the UEFI tool and look for any errors. Also reflash the bios with a clean (ME) dump. Since the pc turns on shorly and the fans spin, my guess is that it 's somehow bios / SPI related.
I would never pay $100-$200 to repair a $200-$300 (not incl RAM, storage) PC!
If I can not get it working, it's a PCB issue and not worth paying someone to fix an old turd.
That's reasonable, figuring out where to draw the line between worth fixing and not worth fixing is something that's always gotta be done.
Too bad they don't have adequate protections for the CPU - thanks for the attempt! I've seen multiple of these Beelink N100's fail - so will be avoiding these.
You are probably right about the BeeLink. If you wanted to test further, you could see if you could get an adapter to hook up a PCIe video card to the M.2 slot. I doubt it would yield any results other than more confirmation.
Fascinating! I am most certainly nowhere near as smart as you, but I had a thought about the remaining dead one! I have a similar issue with a NUC of mine, essentially dead beyond the power light and fan. but I noticed that trying to boot it with the RAM removed, the power buttons light had changed, AND with speakers plugged in to its headphone jack I DID get a beep! so there is some kind of a status check still operating! I'm planning on getting a thermal camera so I can see if anything is abnormally hot when attempting to boot, RAM installed or not. Could be helpful!
Great video pal, keep up the good work! 😁🧡
Thanks for the kind words man! That's really interesting that you got beeps out of the headphone jack, I'd never thought of that myself! I'll definitely be giving that a try when I revisit it, thanks!
My similar Beelink refused starting due to the glue under BIOS battery became conductive and shorted IO signals. After cleaning it works fine again.
Any JTAG in the motherboard you could use to see the boot process? Thanks for the video!
I'm not sure, I'll have to check at some point!
Since I am mainly a software guru, I definitely can't help with the hardware part of things. I will give some insight on the software side of things though. Mini PC's are annoying to fix because the integrated bios chips are poorly designed and a hassle to diagnose. Although, I have gotten myself many times into a situation where digital video output, like HDMI or DP wouldn't work until the OS boot phase, after the communication protocol between the GPU and display has been set. If I were to restart the computer, the display would work just fine, outputting the post message and everything, until a complete power down happened. Then, I would have to repeat the blind process of selecting the boot drive from the boot menu. I believe this could be the problem in your case, so try plugging in a keyboard first. It is possible that an "input device missing", or "system configuration change" (since the RAM was changed) prompt could stop the boot process and therefore not completing the display output handshake. Also, have a bootable drive connected to the computer and see if there are any signs of life. These solutions helped me multiple times, especially on crummy intel atom devices.
Thanks for that insight, I'll likely give a few of those suggestions a go when I revisit this PC!
6:13 It's actually rather simple. Everything which isn't carrying data, is carrying current, therefore anything attached to a current rail is there to manipulate it. Since in electronics there are only a few principles of altering the current and voltage albeit in a billion combinations (like series or parallel arrangements, RC LC RLC configs, etc), you can safely deduce that whatever is on a power rail, is there to either drop voltage, increase it, drop current, increase it, or lastly, to filter it in some manner. Any congregation of elements are there to do most of those things. What we are interested in, is whether the voltages meet the power requirements of the components they're plugged into, so we read the spreadsheet for the component and go from there. You look at the voltage drops, but it's worth looking into the IC power inputs as well, then diagnose backwards.
You should plug in a network cable and see if it lights up. It might be booting with no video.
The NIC did light up when a cable was connected, though I feel like I've heard somewhere that this can just be the NIC doing its own thing rather than a major sign of what's going on with the rest of the system. Still something to look into, though!
@@CHWTT next thing to try then is hook it up to a PC running Wireshark and see if it's trying to connect to a DHCP server. If so, you definitely have a running OS.
Some integrated graphics use ram as of vram so thats probably how it broke
You need to invest in a hot air rework station for dealing with surface mount devices. Also look for tin whiskers when looking at the board when looking at it through the microscope.
Great video. And I agree with your diagnosis on the no post n100
Is there oxidization or metal eaten away on the RAM connector pins? RAM unplugging and replugging while the system is on could make some pretty little sparks.
Deoxit and if that doesn't work, soldering in a new RAM slot might be an interesting thing to try. Also, something to consider, if there were pretty sparks created, then maybe it could be physical slot damage due to human jumping.
Good thoughts, thanks for sharing them!
sometimes, the gravity.test and shotgun tribulation makes your day a little bit better.
your skills are neat, and somehow these two Mini-PCs would be great as a Server for Media, Files, etc.
"Finding schematics for __THIS__ computer was never going to happen" - i like the approach :), sad but __TRUE__ ;)
Thanks for pointing that "CMOS" reference - cmos technology was a great break in the 70s but nobody now remembers how great it is, i dont even remember when have it started.
What you could try is an external gpu by using a m.2 to pcie adapter to see if it does work and if the igpu is the problem.
I think you can extract video signal from that wifi slot with external gpu wifi to pcie adapter
Very interesting ! I'm having the same issue with a Fitlet2 Intel mini pc board. After changing something in the bios the screen is dark. I think a bios problem. So keep trying to fix it !
That could be a boost phase, ie, off until it's needed.
Since most voltages are ok, I think I would go for bios.
A couple people have pointed out the BIOS, definitely going to give that a shot in a future vid.
14:04 btw you do not need to worry. you can just leave these dies "empty" when testing.
just make sure to put a finger on it occasionally to feel if it's overheating or not.
these modern cpu's are so safe you do not need a heatsink most the time lol.
theres a video of someone pointed a thermal imager onto a ryzen CPU's bare die with some program stressing each core.
Ahh okay, good to know it's less of a massive deal than I imagined. I was aware that the CPUs are pretty good at protecting themselves from overheating by shutting down/throttling, I just wasn't sure whether that would happen so fast with a bare die that it could interfere with POST
The Beelink one can be the same problem from Xulu Mini PC, do a checkup arround the Power Circuit and this same resistor from the Power Circuit is the frist thing to do on this board.
Edit: I noticed two corroded SMD Capacitors behind the memory slots, this may be the problem source.
If the Blink system supports a direct flash BIOS image it might be worth while to see if maybe it got corrupted and just needs to be flashed with a working image. It's logical that something like that would happen if the BIOS were active because their normally connected to the memory in that the BIOS image get's written to memory when the system boots. It does this because the RAM is generally faster than the BIOS chip. At least that was the case for older systems. Point of the matter is this. Attempting to flash a good image to it might be worth a shot. Like you said, "it's already broken."
Well one thing that is allso helpful is to setup a postcode monitor. There you get at least an indication where the cpu is hanging.
While putting the RAM in while the power was on it is possible a circuit got more power it than was designed for and the chip holding the CMOS or BIOS got zeroed out which means the computer has forgotten how to turn on (POST). Install a new CMOS chip or use the box for parts.
Nvram reset keyboard combos, those minimalistic bioses can easily get stuck.
the beelink ones are dropping like flies apparently
You could even say that they're dropping... like bees!
Sorry, I had to, haha!
your theory makes sense about the blink pc (one you troubleshooted first ddr5 ram that had
if the coil is completely shorted to ground, forget about it, but since it was operated on while the pc was powered, there is the easiest thing to do with high probability of fix, reflash the bios
Gonna try the BIOS at some point for sure, and the dead phase's coil had about 32 ohms on it so I don't think it was shorted in this case, but definitely something I checked for that reason!
See if there are any TVS diodes near the RAM connector that look compromised. It's possible that hot-plugging the RAM with the PC on was enough to blow one of them short. Conveniently enough most integrated graphics solutions use the system RAM as their VRAM. It's possible that a short (or partial short since it's not drawing an insane amount of current) may be causing a fault in this power rail shared between the iGPU and system RAM which now won't turn on as a protection mechanism. You can probably check this pretty easily first by measuring the resistance of this power rail to ground. Anything less than a few kOhm would be pretty suspicious. A bit of an out there theory but something worth checking if all else fails.
I saw someone do video fixing a graphics card and they just took off or connected across one of those square things and he said the cpu or something figures it out.
On the Zulu, that resistor is likely acting as a fuse. Most times they are very small values. Less than 1 ohm, or maybe even zero. Shorting with solder is fine for test, but putting in a 0 - 1 ohm link is preferred.
I might have to see if I can get a replacement resistor, just so that the repair is a little more clean/professional.
I guess the only thing I could think off that can be done for the n100 is to power on the PC without ram installed and check for heat/power draw differences and hope it's a case of multiple ddr5 modules being bad/damaged by the powered scenario.
Most modern motherboards have a lpc header which can be used with appropriate post debug card to see what is really going on
Getting all the voltages on the various rails is a good way to diagnose some possible problems. Don't forget to also find the on board fuses and 0 ohm resistors between the voltage rails and the actual power using chips. You have found a few, but you did not explicitly say in the video you checked all. You may have done so off camera?
Nice video. Hope there's a follow up soon.
Whats going on next to the ram slot on the side with the audio jack. Looks like two caps have corrosion/burn/blow out there.
Electrostatic grounding wrist straps were a thing back in the day for a reason.
Dislike:
A dead CMOS battery would NOT "create POST issues"!
It would simply go back to DEFAULTS!!!
Well, yes, a CMOS battery that's dead will often make the BIOS reset to defaults, you're right there.
However, I've had several systems such as OptiPlex 7010s and 7020s which entirely refuse to POST and spit out a motherboard error if the CMOS battery is dead, so it most certainly can cause POST issues.
That was insane in the membrane 😮
for the pc that has the graphics problem, couldn't you use a external graphics card setup, you can get external graphics card dock, maybe that would fix that problem.
I am quite convinced that the resistor was only a 0 Ohm resistor used as a jumper. The VGFX line is only active when the integrated graphics card is loaded. Otherwise, the entire CPU is powered by VCORE. It's going to be tense there. S No video output can be a bad ME region in BIOS. You need clean this part of bios but in modern pc is not that simple.Try measuring the resistance on pin 8 of the bios chip. Sorry for my english.
Wonderful video.
Try the beelink without ram and see if you get an error beep or something.
Maybe there are bad cpu pads 😮 but cant know till you know great insight on broken things to bad many things can break them and not enough safety 😅
i somehow broke the bios on an intel nuc, that was set up with ubuntu linux. i had the task of adding a sata ssd and setting it up for dualboot. but then i had to reboot a hundred times through tangled threads and corporate crimes because windows by default holds the pc hostage until configured properly. at some point i force rebooted it while it was on it's way to boot into windows again. that's when it broke because after that, the system wouldn't post anymore. no display, no beep, nothing.
i found that there is a way to update the bios automatically with an FAT formated USB stick, containing nothing but the bios file and i think i had to long press the powerbutton or something to put it in update mode. because there was nothing on the display, i did not know what it was doing but otherwise the system acted like it should according to the online guide provided by intel.
the guide stated that this process could take like half an hour, but i let it sit for more than 4 hours. there was no indication that it was done, but after waiting that long it's either done or dead. i rebooted and the monitor came on again and it booted into ubuntu like normal.
that was not the first thing i have tried. that pc was like almost a month at my house for an upgrade that was supposed to take a few hours tops, fml.
Thanks for sharing that experience, I find it quite interesting! I'll be giving a BIOS flash a shot on the Beelink at some point, I'll have to see if it has an automatic BIOS reflash feature or if I'll have to get an external programmer and flash the chip itself with another PC. Should make an interesting video!
@@CHWTT it works with modern mobo, but sometimes that process fails... if you repeat it a couple of times, it should work... and yes, it can take a lot of time (up to an hour)
Please send the broken machine on to another youtuber! Th3c0der (or however it is spelled) in the UK would be able to replace the CPU. Not that there is commercial value to doing that, but having him swap it out in a video will be good content, collaboration and it might just prove your diagnosis correct.
I have a question that is a bit off topic. The mat that you are working on, the gray mat with a grid pattern and the different sized circles on it. Do you have a link for it? I'd like to know more about it. I'm just starting to delve into this "kind of stuff" myself, and I am still getting my work area set up. That mat looks like it would be a nice area to do my work on.
Thank you and greetings from Alaska.
Absolutely! Here's the Amazon link to the mat that I'm using: www.amazon.com/dp/B0BJKSX6J2/ref=twister_B09MMXCJNQ
@@CHWTT . . Thank you!
Bad RAM might stop you from getting a post screen, but zero RAM shouldn't, should have tried booting with no RAM. And with it plugged in & plugged into the network, check it you saw it on your router. To me only 3 minutes in I'd guess the i/o was blown, meaning half the m/b was dead. But seeing if the Ethernet port was blinking would tell me it most likely is still good.
I've never had a system POST without any RAM, and I don't think it would have any chance of working in this case, because of the fact that the iGPU in this system requires system RAM to work
Now some overclocking to go the extra cooling would be nice
I would try putting in a usb to VGA cable and see if I get a display with that
They must have a way of testing those boards during manufacture. To that end I saw several test pads ready to take spring probes in a jig. What if you try any of those pads with a scope. Is there any lilfe in them?
I'm not so sure that changing the memory had anything to do with it. I bought 2 Beelinks, an EQ12 for a router and a SER5 Max for mostly emulation gaming (I'm old) and some older titles.
The Ser5 Max died after 3 days in the exact same way that your friend's EQ12 did. I read somewhere that it is a BIOS issue (can't find it now) . I'm convinced that it's software and/or BIOS.
Neither of these computers would update past 21H2 out of the box. It is blocked by Windows Update. Beelink acknowledges it on their forum but apparently isn't doing anything about it. There is some reason for that.
The Ser5 died after I attempted to force it to update. I don't know if that's the cause, but it was late, I forced the update with a USB stick and an in-place upgrade. I played around with it a bit, I shut it down, went to bed, and it never booted again. Same issue as you're seeing. I pulled it apart, put the SSD in an external case, wiped it, packed it up and sent it back to Amazon with a note. I wasnt leaving my personal info on it.
The EQ12 won't be running Windows but I wanted to make sure that it was OK. When it wouldn't update, I found the drivers, dug the key out of the Registry and attempted to do a clean install. It wouldn't let me. I couldn't get past the MS account screen. There isn't anything that you can just click and bypass anymore but you can hit F10 and type a command. That didn't work. F10 did nothing. I gave up and attempted to sign in to a secondary MS account. MS blocked it.
Beelink has done something to their Windows images that Microsoft really doesn't like. I don't know if that's causing a hardware issue but it's still another thing wrong with these Beelinks.
I replaced the SER5 Max with a GMKTEC M7. Much better. All the drivers are available. I did a clean install that went smoothly. The key activated OK. I definitely recommend the GMKTECs. I'll get another if the second Beelink dies.
I would love if you could explore the Core Ultra and Arc GPUs more from a media server perspective.