Your strange/bizarre computer repair stories

I just fixed an issue with a computer that I hadn't seen before and wouldn't have imagined. I figure plenty of you [gearheads] have had a similar moment. So, what was the problem/event, and how did you fix it/figure it out? Stories can be from any time, hardware and/or software related.

Here's mine: I diagnosed a stuck "Delete" key on an out-of-warranty Lenovo IdeaPad Flex 5-1470, where it acted as if the Delete key was continuously pressed (and thus would delete anything you clicked on, would delete any text in front of your cursor, etc). I thought it was a problem with the keyboard, but the issue persisted after disconnecting the laptop's keyboard, and was still there after plugging in an external keyboard. To rule out a software cause, I booted from a live CD and still had the problem. I also tried removing the main battery and the CMOS battery, and disconnecting the display/touchpad (also sources of input), but none of that isolated the problem. I ended up having to replace the motherboard (got one for $46 on eBay). It probably was a bad keyboard controller or some other gremlin. But never would I have imagined needing a motherboard replacement to fix that.

Charles

 

I was getting constant BSODs when trying to install Windows XP on my Clevo D901C. Tried everything - different GPU, different RAM, different CPU, different hard drive and nothing worked. Turned out that it was the battery which had gone bad and it installed fine after removing it. Wouldn't have thought that would throw up a BSOD in that way, especially since Vista installed without issue.

 

LOL, Vista out of ALL systems

funny timing of this topic, actually!

these past few days i've had weird freezes on my system and i couldnt figure out for the life of me what caused them. no bsod, no sound hangup/loops, just system freeze and that's it. would stay that way until hard reboot. fresh windows LTSC 2019 install, brand new SSD, brand new RAM, so what could be the issue?

i rechecked my cpu undervolt, to no avail (freeze at stock VCore). lowered my RAM OC back to SPD settings, still freezes. so then i opened up the machine and guess what i found!

apparently, the thermal pads used for my SSD heatsink were a tiny bit too long on one side (heatsink with backplate, thus two thermal pads used). that tiny bit actually got into the contacts of the M.2 port and likely disturbed the data transfer, thus causing the system freezes. so i thoroughly cleaned out the thermal pad residues from the contacts and voila - no more freezes!

so yeah, on a system where every single component is tweaked/overclocked/undervolted/non-stock its a frigging piece of THERMAL PAD that causes system freezes

 

A while ago friend at work had a dell that constantly froze or intermittently would not boot. brought it home and tore it down. All looked fine and worked on it several days. He finally remembered to tell me he replaced the old CD drive for DVD burner, an older IDE setup. Went back in and found he used the same 5v/12v power connector and almost all things were on the same leg from the cheap stock ATX supply as assembled. rewired the system splitting the load from the PSU on the two legs for the different devices and all became fine.

memories of the old days...……..

 

not that bizarre. upgrading m18x r1 from a 2630qm to a 2920xm. after replacing the cpu only 2 sticks of ram work and 1 of 2 GPUs aren't getting picked up in bios. swapped 2630qm back and everything is back to normal.. faulty cpu with faulty imc and half the link to GPU is hard to find, didnt think i'd see one. it was easy troubleshooting.

 

Before I begin: I have all the tools and I used most of them. I have a bench tester for PSUs, a fancy multimeter and everything was tested on a proper insulated surface while using anti-static gloves and tether.

I had a computer in which all of the parts tested fine, but it would not boot in its original case. Like anyone, I assumed it was either the PSU, the power switch or one of their wires shorting somewhere. I tested everything, removed every part, even booted another test system using that PSU. There was nothing wrong with the parts. So then of course I assumed something about the case was causing a wire to rub, static build up, whatever. So I went absolutely nuts checking every part of the case for stray voltage or heat when the power supply was engaged. Nothing. The switch in the PSU would engage, the rail would carry power, but nothing would happen. Tested the board, good. 100% nothing wrong with anything and all of it would boot outside that case with no issue.

Finally just grabbed an identical case off a refurb pile at my friend's warehouse (refurbisher, had hundreds of them). Put the parts in, booted fine. I was forced to finally, after thousands of repairs, accept that gremlins are a thing. In EVERY other case I have always been able to figure out the issue. This one was just cursed to not boot.

Also just a few days ago I bought a cheap busted Asus laptop off a dude on cL. I just picked it up because the shell was pristine and I have a line of the best main board for cheap (Asus X555L 2017). The guy even tried to lie about what components it had, because it was dead and he knew it would never boot. Said it was the i5/gt930m model when it was just the i3 one. Anywho, I decided to test the bad board for fun and holy god it lit up my test bed like a Christmas tree. I have never seen so many shorts in a laptop board. Turned out that some kind of conductive sand plus beer had been spilled on it and I guess the person must have shook it until it got under every single heat pad and adhesive, then tried to turn it on like 200x.

 

A week ago, I wanted to clean my external keyboard.
I violently shook the keyboard, over the sink, while turning it upside down.
Lots of shaking, beating, smacking. It lasted for minutes.
I found some "things" trapped under the keys. Removed them and more excessively violent shaking, beating, smacking.
After that, I plug it back into my laptop.
A few keys seemed to be stuck or not working, but after a few hours, the keyboard "settled down' and all keys work fine now.
It's good to clean out keyboards sometimes, but I was lucky there was no permanent damage.

 

i've got a few problems here maybe someone here can help me find a solution or whats wrong with it.

XPS M2010:

1. SATA I so 1.5Gbps ~ 150MB/s at best but when I overclock T7600G and then do benchmark Crystal disk mark or HD Tune, it'll bench at higher speed say 175-200MB/s depending on the overclocked frequency but actual speed of course capped by SATA I. Restart computer with the higher frequency (saved by TS profile) and then down clock, reverse will happen I get lower MB/s reading from benchmarks.

2. its got a unique ODD bay you can look it up on google. swap out the ODD (IDE connection) for a SSD and caps out at 80MB/s read and 60MB/s write. Write no issue, read has a ton of problem. HD Tune test run at 80MB/s for like 2 sec and it'll drop down to 2-3MB/s and eventually lock up. Put a game, play from disk & it'll freeze up and forced to restart PC. install windows from original ODD bay has no issue, no lock up and definitely faster than 2-3MB/s read. lastly, if I use a software called FastCopy and transfer file read/write from/to disk, no issue at all, at full 80MB/s speed.

3. T7600G is like T7600 at base clock 2.33ghz but G means unlocked so I can overclock with throttlestop. for those who uses throttlestop knows the clock modulation is always at 100% and if that is lowered it'll lower frequency and performance etc. I can overclock cpu with 1 more multiplier and have no issue ~2.5ghz but anymore than that I start running into issue where clock modulation reading on TS is at 25% and computer is extremely sluggish, this can't be changed/fixed until a hard restart. other weird thing is that it passes TS stress test, but browsing/gaming will cause this down clock very quickly.

if you know a solution/ workaround to these problems especially 2/3 i'd really appreciate it.

 

1. is not a problem, its just a fact CPU and storage performance are indeed linked together and looks like u have a cpu bottleneck at stock clocks.

2. Just wild guess this might have something to do with power delivery. writes usually use up more power than reads, so could be that the ODD connector gets to its limits when u do write intensive operations.

3. have u checked TS throttle reasons? maybe ure running into a power limit of some kind.

Sent from my Xiaomi Mi Max 2 (Oxygen) using Tapatalk

 

for 2. its giving issue when its read not write lol which is the mess up part. but fastcopy reading files from the ODD bay at max speed gives no problem at all.

for 3. i'd have to check with latest version of TS but i seriously doubt it will give me the info i need, cause this is old ass cpu. as to how old.. T7600G, x7800, x9000, x9100, QX9300, 940xm (first gen i7).

 

@ole!!! did you try different ODD caddy? I read that some crappy IDE to SATA ODD caddies are in fact IDE to USB to SATA, which may introduce all kinds of problems.

Regarding CPU performance - what OS are you running, what settings do you use? Your T7600G should be about on par with quad-core modern Celerons e.g. Apollo/Gemini Lake, bottom line - sluggish but usable. GPU performance of your XPS, on the other hand, is probably outright pathetic and way below modern standards.

As for software - consider using CPUGenie, worked like a charm for me back in C2D days.

 

ODD thing that im not sure but interesting to see if that could be the reason. caddies I got were from ebay, what should i be looking for when looking at the PCB? USB controller?

also for CPU Genie, is that an overclocking software?

 

I'm talking about the thing that allows you to use SSD in ODD bay. You should be looking at all controllers on its PCB and googling their details. If your friends have similar caddy but with different hardware, try it in your machine, maybe the one you got is buggy.

Also, look into ODD power management in BIOS if you didn't yet, it may well be the culprit of your problems. You can post photos of BIOS menus here if you're not sure, I'll look into it.

CPU control, not just overclocking. Similar to Throttlestop and XTU. It's shareware, but you can always pull a Bartholomew Roberts on it, if you catch my drift.

 

i'll try the software for sure, as for ODD bay, i'll take it apart and look it up, you would think ODD bay with moving parts use more power than a power saving SATA SSD that runs way below its spec so even if its USB i seriously doubt it'll be power related issue, at least not the board/pcb.

as for the ODD power management that one is a goner, this is laptop we talking about here and there isnt any unlock bios thing so outta luck.

but yep definitely trying that CPU software.

oh btw, its been about 2-3 yrs i havent turn on the computer and they have SATA SSDs in them, i'll be one of the first to test "SSD loses data if not powered on after 1yr"

 

Because ODD may draw a lot of power, BIOS may have aggressive power saving features in place. Control over this feature is typically available in standard BIOS. But then again, maybe crappy ODD caddy is the culprit, e.g. the circuitry gets hot or overwhelmed, and speeds drop dramatically.

 

A couple of oldies...this one's not really interesting on its merits, and it's nothing that we'd recognize as a computer today, but it happened before a lot of you were born. My father was a chemical engineer Way Back When, and his office had a bunch of HP calculators around. This was back when HP built real stuff, like oscilloscopes, logic analyzers, and things that were in general built to last (when William Hewlett and David Packard were around and Carly Fiorina and Mark Hurd...weren't). They built a lot of nice calculators, but they did have one little vulnerability if they dropped, particularly if they fell on their side. Whenever that happened they'd throw the offending bit into a drawer somewhere and forget about it, despite the fact that those were very expensive back then (the HP-35 cost $400 in 1970, which would be about $2800 today). He brought one home on day for me to play with, sure enough, it either didn't turn on or it just flashed a weird display but was useless. I opened it up (a bunch of Phillips screws in a solid plastic case), and lo and behold, discovered that it had a large board, with a smaller mezzanine board connected by means of a row of split gold-plated connectors soldered to the mainboard with matching plated through holes in the daughter card. The split connector was supposed to act as a spring to maintain a solid connection, but if it fell on its side, the pins could be squeezed together. I noticed that it didn't feel like a solid fit, so I took my penknife and spread them apart. Reattached the daughtercard, and it turned on just fine! He gradually brought home a couple more HP-35's, an HP-21, and the real gem, an HP-45, which was my calculator throughout high school and college. Back in high school, there was only one other student who ever figured out how to use it, so I rarely had to lend my calculator out (everyone else had cheap TI junk).

(The HP-45 was an interesting beast. It was a hopped-up 35, but it also had a hidden stopwatch that I somehow found out about. It wasn't very accurate, running about 11% fast, but if you knew about that, you could use it, along with its 10 splits. In the later HP-55, the stopwatch was real. But I never could score one of those, nor the real top of the line, the HP-65, which actually offered external storage. No matter any more, if and when I need this, I have an HP-48 emulator on my phone. A couple of them actually, I think.)

The other story's a bit more interesting. I was working for a supercomputer vendor, first job out of college, I had been there about 5 years. One of our customers (sorry, I never use customer names, even from decades ago) was having a lot of trouble with errors from their machine and escalated matters higher and higher because our customer support never could figure out what was going wrong. I had a pretty good rep as a debugger, and they wanted someone from engineering out there, so I got sent out. My manager told me he was hoping it would prove a software error.

So I fly out there, meet up with our support people, and drive out to the customer. I wanted to take the measure of the errors, so I wrote a quick and dirty little Fortran program that would create and operate on an array of data. Fortran's a good language for that purpose because the data layout is very well defined. Anyway. So I run it, and sure enough, it generates data corruption at a particular array element. Mapping things back, I blurted out "board X is bad" (whatever index number of the board). Saying it out loud like that was putting my neck a bit on the line; it committed us to swapping out an expensive board and would look bad if I were wrong...but we go around to the back of the machine to access it, and it has a red fault light glowing! The customer support people were half in awe and half...not too happy. In retrospect I should have quietly called one of our support people over and told him what board it was and let him figure out how to say it, but I was young and a bit tired at the end of the day. So we replaced that board, then some others, then we needed someone else to fly out with more because there weren't enough spares on site. But by the end of the week, the machine was in a lot better shape and the customer a lot happier.

 

well it definitely isnt the heat though, cause like i mentioned. if i copy files in window OS using explorer when reading files from ODD bay it will freeze/lock or slow speed. if i play games from the drive it'll do the same (read files from it).

however if i use FastCopy software to copy instead of explorer then it is fine, copy for hours at max 80MB/s read no problem at all, hence the strange part also why im looking for help lol

 

I have this clock on my desk at work. We were running low on room in the power strip, so I connected it to my desktop via USB.

All was fine until a few days later, I noticed one of the programs on my computer would freeze on startup (a tool pathing program for a 5-axis water jet - not something you can find a lot of support for on forums, etc.).

I eventually put two and two together (what changed recently??), unplugged the clock, and lo and behold, the app started right up. So I took an x-acto blade and damaged the data pins on the USB cable so it only draws power now. Problem solved!