For the past year, I've been forced to use my "secondary" (+ newer) workstation, as I woke up one morning to my former primary workstation (from 2018) having a severe bout of hardware-related instability.
Granted, this secondary workstation has actually been a vastly better machine in most regards, as it has actually been more of the setup that I'd originally been going for in 2018 (see details below). And also, this switch has helped postpone having to address a several "storage crunch" I'd been experiencing on the primary machine (i.e. I'd gotten to the point where my Photos/Media drive had actually *run out of space*, and I had to quickly offload as much "non-critical" stuff as I could to a mix of "other drives on the machine, for "more important" stuff, and only a secondary drive for all the "not very important" stuff)
Anyway, for "reasons", I've come to finally get around to trying to resurrect this machine now at last! (I'd actually mean to do it by this time last year... but then procrastination over finding "the right" replacement PSU and/or trying to fit that in against work/holiday schedules has meant that this has only happened now...)
This rambling post will be about the journey of trying to get this working again, and also the upcoming challenges of figuring out how this step now fits into my grand scheme for rejigging how I run all my computing setup (i.e. both the "Linux migration" project, but also the "two workstations / two locations, one big stash of data that needs backing up and syncing")
Secondary Workstation vs Old Workstation - Main Differences / Lessons Learned / etc.
So what differences were there between the two workstations? Among other improvements, I managed to get a bunch more of the things I'd actually originally been wanting when I got this older machine custom built:
* 1) It has a 1 TB NVME SSD for the primary OS drive, whereas I'd only gotten a 512 GB for the older machine. This in particular has been a decision that I've since come to regret actually, but at the time, I was like: "Eh... not sure I really need to spend a whole chunk more on that component".... Yeah, nah. I'm currently sitting at a small single-digit percentage of free disk space there (TBH, it has actually gone below ~600mb a few times!), with my other workstation constantly sitting just over that mark, despite keeping even less "project" stuff on there now!
* 2) It uses the case that I'd actually been eyeing up originally (i.e. one of the "beQuiet" ones) - Finally getting one of those, I can confirm they're much nicer in many regards (especially for opening it up for performing regular dust-control), except for 2 points:
i) It's really annoying how the large power button flashes brightly whenever the thing is sleeping
ii) The case is *MUCH* heavier. So heavy that I've learned I shouldn't use such heavy ones upstairs again...
* 3) Faster single-core CPU speed
* 4) It actually has built-in wifi, instead of being straddled with a USB-plug-in thingy that used to weirdly die after a few sleep-wake cycles, and need unplugging to fix.
* 5) Finally, the secondary "Projects" hard-drive is also a 4 TB like the "Photos/Media" disk
How it All Started: A "Sudden" Overnight Hardware Fault
So, how did this whole affair start in the first place?
Well, I woke up one morning last January to the sound of computer fans surging on and off. This isn't *entirely* unheard of, as there have been times when I've forgotten to fully put my machine to sleep before bed (or more likely, some pesky app blocked the sleep / shutdown, and then kept the machine running all night). So, I figured that it must've done what usually happens in these cases: 1) Several hours of idling had resulted in it starting to run some cleanup / maintenance tasks, which then coincided with, 2) A long delayed Windows Update forcing itself to install! Totally normal out-of-control Windows behaviour! 😅
When I got up to check what was going on, I realised that actually, the machine was actually turning itself on and off, *constantly*, many times barely managing to fully boot up (even after cutting power then turning the power back on at the wall without touching any buttons on the machine!). Oh, and there was a "sour" stench in the air, that I initially thought was a new leather chair "off-gassing"... (I later realised, it may have been more electrical)
It took several attempts, but I finally managed to get it to boot up for long enough, and stay alive, for me to grab some hard drives and make as much of a set of "emergency backups" as I could manage. I got most of the most critical stuff (or so I thought), before I came back from a quick food break to find that it had in fact rebooted again, and gone back to surging on and off!). Damn!
And that was that. I unplugged the machine, and left it off like that (assuming a major power-supply hardware failure / fault), where it would be best to not let it boot up again until the problem was solved.
(*) I may have also tried again the next day to check if the problem would go away by itself, but I think it had also started surging on and off ad-infinitum again then, so it would've been obvious that the safest course of action was to keep it powered off!
Interim Computing Solution - The Great Lifeboat-Switcheroo, and Decision Paralysis / Procrastination
IIRC, fortunately, I did still happen to have my laptop with me, so I limped through a few days using that. I can't remember if it was because I had it so I could sync things back across to this machine (following the 2-3 month long "exile" in my other place, while my usual den was getting renovated, and this machine was down for an extended period for the first time since I'd gotten it). In any case, I knew that relying on the laptop wasn't a good option, as performance-wise, it was kindof crippled by throttled performance (due to the HDMI output having to go via its anaemic "Intel Integrated Graphics", instead of via the discrete GPU).
Hence, very quickly I figured I had to move the newer workstation over, and move this old one over to the other place.
Resurrection Day
Just in time for Easter 2026, all the stars aligned for me to finally have an impetus for trying to get this machine running again this month. And so it was, that on the 20th of March, I called in some professional help to try to replace the PSU to try and have a chance of getting this working stably again.
As I should have probably anticipated, when we went to power it up (for the first time in over a year now by this point), it proceeded to largely work! Well... there was a "corrupted files" warning, which is to be expected from the system having shut down abruptly many times in quick succession back then like that, PLUS you've also got an SSD that has been sitting idle without power for the past year
But yes... surprisingly, everything "just worked" again (and TBH, if I'd dared to plug it in again to check, a few months after that fateful day / week, I may have ended up just taking a punt on it too!)
But since I was still spooked by the whole "it's likely something physical popped in the power supply, which has exceeded its expected warranty period (i.e. 5 years)", PLUS I had also sourced a replacement PSU with someone around who could install it, I figured it was still better to just get that component replaced for peace of mind.
And so, with that, the PSU has now been replaced out for a new model + different supplier (i.e. 1000W vs 850W now). I must also say, that it's been quite instructive seeing this done in real life in front of me. As someone who has relatively little exposure to physical hardware stuff, this was very useful to just get a good grasp of "how it's actually done" vs "how folks say it is done, making it sound very complicated"
Return of the Random Black Screen Flicker
Initially, all seemed well enough... That is until I started getting frequent random bouts of black / blank screens - the type that used to only happen when trying to play a Netflix video in Firefox, and Windows Updates were available and overdue! Except now, there weren't any updates to apply, but the screen kept blanking out for 1 second pauses at a time. Oh, and then some of these started to have a bit of flickering in there too!
Oof... this was getting way worse than before, and I ended up shutting down the system to try to stabilise it before anything got any worse!
As usual, the first step was to Google to see if there were any know issues + easy fixes.
* 1) A search for Nvidia and black screen flickering issues pointed to a potential interaction with the Color Management and the "Auto" Setting that was incorrect flipping between "Desktop App" and "Full Screen Video" cases - Unfortunately, this didn't seem to have any real impact.
* 2) The next most frequent recommendation was the usual "reinstall drivers" thing - Err... wasn't keen, but if it came down to it, I may have to!
* 3) Then there were the mentions about RAM going bad, and/or maybe some of the connectors fouling up - Hmm... RAM going bad is something the guy who replaced the PSU suggested as having been the original cause of all the problems, so I added to my list: "Run some memory tests" to rule out any issues there
* 4) And on that topic of bad / worn connectors, one suggestion I saw was to try using a different PCI slot, in case it was the connectors there going bad! (hmm... again, might be forced to try that if nothing else works)
A Dust Buildup / Blockage Being the Real Culprit?
At this point, when I mentioned these problems on my socials, someone chimed in to say that it sounded a lot like a dust / cleaning-needed problem! Now that's a salient hint that I can do something about, as there's still quite a lot of obvious cleaning to be done (e.g. particularly with the CPU cooler, but also quite a few of the fans!)
From my experience with the other newer workstation, I'd come to understand for example that the filter in front of the front fans is particularly prone to building up a dust layer. But thus far, I couldn't find an easy way to access that panel / filter (unlike on the BeQuiet case, where it's a simple matter of bonking off the front panel (by nudging upwards from the bottom)) - and I still haven't!
That lead me to looking up all the documentation + promo videos for this thing. Interesting things I learned:
* 1) There was also a top filter, that was apparently "easily removable". But how do I access it?!
Answer: Turns out, there's a slight ridge about an inch from the top at the back of the case. Pushing upwards on that with a bit of force leads to the whole top-plate/shell coming off, revealing the top filter! As expected, that screen (held in place with a magnetic frame) did have quite some dust buildup - i.e. a clearly visible "white" layer through all the fan holes!) Removing that and trying to wipe it clear quickly revealed a whole bunch of dark lumpy clumps of dust that fell out! (Good thing I had a respirator on for this!)
* 2) From the various pictures, it looks like the front panel can be opened like a hinge! But how/where he release mechanism lives is again a mystery! So far I haven't found anything! (But if/when I eventually do, it will most likely be great to find it, as it would give easy access to cleaning the filter-screen that I *know* is clogged with dust right now
Failing finding that flap again, next-up, I finally worked up the courage to try clearing some of the obvious matted dust inside the CPU cooler (i.e. on the grill, just behind the fan) using a small brush. Yeesh... just touching that ended up bringing the whole mat out in clumps pretty easily! (That said, there's still a lot of the dust stuck on the fan blades themselves, along with who knows how much inside all that grid fin structure... a good blow out using compressed air will be needed at some point, but I don't have any right now, and also, that would require moving the machine up and down the stairs again, which is NOT easy for me at least)
Next steps:
* Get some compressed air, and try blowing out a few areas sometime in the next week or two
* Keep looking for a way to open that flap
* Look into why some of the fans don't seem to be producing much in the way of airflow (i.e. esp that rear fan, which should IMO be giving a lot more airflow)
* Try to figure out again if anything can be done about the CPU cooler's fan noise (i.e. a constant somewhat irritating growl), which may/may not be related to all this
No comments:
Post a Comment