Does this look like a 120mm fan port?



  • [img]http://i.imgur.com/c6Hzm.jpg[/img]



    I'm doing some upgrades very soon and I want to get a case fan for the back (because it's not a well ventilated case) and I'm hesitant to buy a 120mm because it looks like it wouldn't fit. Anyone here know if it would? It looks like there's JUST enough room but it'd be a squeeze.



    There's also 80mm mountings...



  • I'd go for a 92mm fan. I actually just bought a 120mm fan for a HTPC case, only to find it wouldn't fit grumbles. Turns out it was 92mm.



  • 120mm is the full width of the square body of the fan. Distance between the centres of two screw holes (along a 120mm side) is 105mm. Looks like what you have.



  • Remember, cutting off the grille results in better airflow and less noise due to reduced air turbulence. Even more so if you attach the fan with rubber screws.

     

    The more you know...



  • I see, thanks... it's not a great case, just a bottom of the line one. Looks to be ex-Dell judging by the blank silver circle on the front.



  • @DOA said:

    less noise due to reduced air turbulence.
     

    Air turbulence decreases noise. See owl wings. Some expensive fans have special notches to create controlled turbulence and reduce noise.

    But cutting off the grille probably isn't a bad move regardless. Free airflow.

     



  • Discourse touched me in a no-no place

     @nexekho said:

    I see, thanks... it's not a great case, just a bottom of the line one. Looks to be ex-Dell judging by the blank silver circle on the front.
    Negative. That's an HP/Compaq consumer line chassis.

    That grille is distinctive to them, as is the fiddly shit going on with the PCI slots and the usage of a standard pop-out ATX panel. Some Compaq machines in that class used circular logo inserts. It's also possible it's on the same metallic chassis, but with a different plastic fascia. All pretty standard issue. They're designed for either 80mm or 92mm fans. A 120mm retrofit is possible.

     

    Fun fact: CDs are exactly 120mm in diameter. You can use them to measure and mark grille cuttings for 120mm fans.



  • Thanks for that tidbit... I've never done any maintainence or "true" IT work, so I'm not massively experienced... just guesswork.

    The 120mm fan fit perfectly. Unfortunately, my RAM was DOA - Windows will boot but it's very unstable and when installing updates (gave me NO opportunity to say don't do it) it trashed itself by overwriting system files with corrupted copies.

    70000+ errors in Memtest86+. Had to wait until the next day to RMA because Ebuyer's return line opens at 9AM and closes at 6PM, making it difficult for anyone who has a job to return anything. Oh, and five step RMA (web, phone, email, post, receive) is TRWTF.



  • @nexekho said:

    70000+ errors in Memtest86+
    Please tell you wrote down "over 9000 errors in Memtest86+" as the reason for RMA



  • @nexekho said:

    Windows will boot but it's very unstable and when installing updates (gave me NO opportunity to say don't do it)

    Uh, wha? Why are you blaming Windows for any of this? Would it have made a difference, considering you have 70000+ memory errors, if Windows had let you skip the update? I have no clue why you mentioned this little tidbit... it's confusing me. Do you think if Windows skipped its update, your system would be stable somehow? (No wait, you already said Windows was unstable pre-update.) Also, why did you even bother installing Windows when Memtest failed? Did you think Microsoft has some technology to magically detect and fix bad RAM?

    I find it compelling the extent people will blame Windows for things that couldn't possibly be Windows' fault.



  • I don't blame Windows for it but you'd think it'd be a bit more careful, perhaps writing files out to a .tmp, checksumming, then swapping if correct.



    It was a previous install of Windows from before the update, I assumed the instability was some leftover driver messing everything up, it happens.



    Also, where did I say I got bad results from Memtest, THEN proceeded to run Windows anyway? Or that I did a fresh reinstall? Did you not think perhaps I ran Windows, noted how oddly it was behaving, suspected the RAM and then ran the test?



  • @nexekho said:

    I don't blame Windows for it but you'd think it'd be a bit more careful, perhaps writing files out to a .tmp, checksumming, then swapping if correct.

    Windows makes certain assumptions about its operating environment. For example, it assumes it FUCKING WORKS.

    @nexekho said:

    Also, where did I say I got bad results from Memtest, THEN proceeded to run Windows anyway? Or that I did a fresh reinstall? Did you not think perhaps I ran Windows, noted how oddly it was behaving, suspected the RAM and then ran the test?

    So you built a computer without running memtest?



  • I've never encountered bad RAM before, and never known anyone to have bad RAM.

    I realise there's only so much you can do with regards to checks but surely when updating the system files needed to boot the OS it'd read them back and make sure it hasn't bricked itself?



  • @nexekho said:

    I realise there's only so much you can do with regards to checks but surely when updating the system files needed to boot the OS it'd read them back and make sure it hasn't bricked itself?

    Windows makes certain assumptions about its operating environment. For example, it assumes it FUCKING WORKS.



  • @nexekho said:

    I've never encountered bad RAM before, and never known anyone to have bad RAM.
    One of the problems I've seen over the years is RAM chips that aren't actually "bad" but they can't run at the speed advertised (would cause a BSOD). Going into the BIOS and manually setting a lower speed, rather than auto-detecting, would make everything work OK.


  • Discourse touched me in a no-no place

    @nexekho said:

    I've never encountered bad RAM before, and never known anyone to have bad RAM
    It's the single most common problem I've encountered. It's endemic to certain make/models of PC at certain times - Dell GX270's had a very bad run for about a year.A few HP/Compaq business lines as well.

     

    Bad memory [i]controllers[/i] are also common. I have no fucking idea how. This particular problem has gone away since memory controllers went CPU-side - so it probably had something to do with lousy northbridge cooling.



  • @Weng said:

    Bad memory controllers are also common. I have no fucking idea how. This particular problem has gone away since memory controllers went CPU-side - so it probably had something to do with lousy northbridge cooling.

    I've had bad disk controllers more than once.

    The real tragedy there is how Windows handles (what it perceives as) bad disks: it keeps a counter of bad (but non-fatal) reads/writes and every 5 it lowers your DMA level. If you're already at DMA level 1, it goes to PIO mode which, as you know, is the slowest thing EVER. If it encounters bad reads/writes at PIO level, it doesn't bother telling you, instead it just does nothing but tally them.

    The reason this is such an annoying behavior is:
    1) When your disk slides down a level at a time, you don't notice anything wrong because it's so gradual-- until it flips from DMA to PIO, where your performance turns to shit
    2) Windows logs these events in the System Log, but doesn't have like a notification bubble to notify you when your disk goes down a level. At the least, it should notify you before it goes to PIO mode
    3) If the read/write errors stop, the error counter never resets, and your DMA levels never go back up. Since the DMA/PIO counter is associated with the disk controller and not the actual disk, even replacing the HD will not reset the counter
    4) There's actually no good way to tell if the problem is due to the disk controller or the disk itself, short of replacing the disk, then deleting the disk controller from Device Manager and letting Windows reinstall it, then checking the DMA level each week or so.

    (Note: this gripe is from Windows XP. I haven't yet experienced a bad disk controller on Vista or Windows 7.)



  • @blakeyrat said:

    If you're already at DMA level 1, it goes to PIO mode which, as you know, is the slowest thing EVER.
     

    Had that once. Kept wondering why opening notepad stuttered audio playback.

    I "fixed" it by uninstalling the DMA channel and rebooting. The computer was in use for a few months after that; then bricked. Thank god for backups.



  • As largely an outsider to all this, is this an explaination as to why some machines just run dead slow after a while? IO errors causing a drop in chosen bandwidth? With no warning or automatic raising? That's... silly. Just real silly.

    In other news, EBuyer told me it's gonna be 3-5 days before they even LOOK at the RAM, then a further 3-5 to return it, so I told them to forget it and just refund. Bought some other DDR3 1333 from another store. Some people actually rely on computers for work, ya know? I still have this laptop but I'd like to be able to use my desktop thank you.



  • @nexekho said:

    As largely an outsider to all this, is this an explaination as to why some machines just run dead slow after a while? IO errors causing a drop in chosen bandwidth?
     

    That... may actually be something.

    I always chalk up poor performance to poor maintenance, i.e. the user's fault, but I have only data points from myself (always a fast system even after 5 years) and a few non-power users (always crap out and always the complaints about how slow it is.)

    But perhaps there's some merit to checking out the DMA status and drive health.



  • @dhromed said:

    @nexekho said:

    As largely an outsider to all this, is this an explaination as to why some machines just run dead slow after a while? IO errors causing a drop in chosen bandwidth?
     

    That... may actually be something.

    I always chalk up poor performance to poor maintenance, i.e. the user's fault, but I have only data points from myself (always a fast system even after 5 years) and a few non-power users (always crap out and always the complaints about how slow it is.)

    But perhaps there's some merit to checking out the DMA status and drive health.

    Believe me, the PIO slowdown is an order of magnitude slower than the "crampacked with malware" slowdown. If you've worked with computers, it should be obvious when it happens...

    The real danger is the more subtle downgrading of DMA levels, which you really wouldn't notice unless you frequently copy large files, and keep track of the speed you're getting. DMA-1 is significantly slower than ... whatever drives ship with now. (DMA-5? DMA-6? Whatever.) But it's not "your computer is crippled" slower.

    If you want a preview of what we're talking about, you can manually disable DMA. I have no idea why that option exist, or what possible reason anybody could have for using it, but it's there.



  • @blakeyrat said:

    Believe me, the PIO slowdown is an order of magnitude slower than the "crampacked with malware" slowdown. If you've worked with computers, it should be obvious when it happens...
     

    It also happens on computers without malwarespywarevirusetcetera.

    But then again, fucked software like F-secure and the plethora of HP and Canon drivers and software that the common user obliviously installs might conceivable be categorized as malware. That shit really piles on your system.

     


  • Discourse touched me in a no-no place

    @blakeyrat said:

    (Note: this gripe is from Windows XP. I haven't yet experienced a bad disk controller on Vista or Windows 7.)
    I'm told you can actually turn DMA back on now - and that it'll automagically flip it back on for a given channel when all the drives attached to that channel have different identifiers. Functionally, this means that a SATA channel will reset its counter when a drive of any other model is plugged in - PATA you have to change both drives.

     

    In the interests of being informational, here's a screenshot of my computer, showing what appears to be bitrot in progress:
    [img]http://pic.phyrefile.com/e/ey/eyemwing/2011/03/04/dma2.png[/img]

    The drive in the lower-left should properly be operating at UDMA6 or 7. Since the drive will theoretically do throughputs around 90MB/sec, it's operating in a degraded mode at anything below UDMA5 (up to 100mb/sec). Multiword mode 2 is a maximum of 16.7MB/sec - which, incidentally, is the same as the fastest PIO mode. Frankly, this sounds absolutely incorrect - that's got my boot and primary storage partitions on it - I'd most certainly have noticed the whole damned thing sucking that hard.

    So in the interests of science, here's me showing you that you can't always trust everything you read in Device Manager:
    [img]http://pic.phyrefile.com/e/ey/eyemwing/2011/03/04/devicemanagerfail.png[/img]

    Yep. Throughputs of around 90MB/sec.

    Multiword 2 also supports a minimum response time of 120ms. Benchmarks reveal an average random response time of about 13.

     

     

    ... So how does one report a bug in Windows, anyway?



  • @Weng said:

    ... So how does one report a bug in Windows, anyway?

    I'd bet dollars to donuts that the bug's in the ATA driver or firmware and not Windows.

    But the first step is to narrow it down and figure out exactly what's causing it-- that's a toughie. Unless you work for Microsoft and have a vested interest in it, I'd suggest leaving well-enough alone.


  • Discourse touched me in a no-no place

    @blakeyrat said:

    I'd bet dollars to donuts that the bug's in the ATA driver or firmware and not Windows.
    These drivers are the reference AHCI shipped-with-Windows ones, so it's still their problem. Of course, this might just be a string bug if the drivers predate Mode 7's availability and it rolled around to the first string in the array or whatever. (NT4 is exactly the right age for its 'minimum' to be Multiword mode 2)



  • @Weng said:

    @blakeyrat said:

    I'd bet dollars to donuts that the bug's in the ATA driver or firmware and not Windows.
    These drivers are the reference AHCI shipped-with-Windows ones, so it's still their problem. Of course, this might just be a string bug if the drivers predate Mode 7's availability and it rolled around to the first string in the array or whatever. (NT4 is exactly the right age for its 'minimum' to be Multiword mode 2)

    Interesting! Part of me wants to check my own computers for the same bug. A much more demanding part of me wants to play Far Cry 2, which I recently got cheap off Steam.



  • Just wanted to point out that the OP is not dumb to ask this seemingly simple question. I'm in a sprawling apartment complex in the USA that's only a few years old, and wanted to get a fitted shower curtain rod. Knowing that the American standard for bathtubs is 60", I decided to measure just to be sure it wasn't a shorter tub, and came out to just over 59". I made the mistake of thinking that's probably good enough, bought the 60" rod, and found it was indeed an inch too long and I should have not given the benefit of the doubt. It turns out the developer bought a tub that uses European standards, which is 150cm, coming out to be 59.05" or so.

    Live and learn.



  • Thanks for that...



  • @nexekho said:

    Thanks for that...

     

    Yeah, RHuckster condescending post totally changed my opinion about you, now in my eyes you are not asking a dumb question at all, in fact you are almost smart.

    So thanks RHuckster for showing me the light



  • @serguey123 said:

    Yeah, RHuckster condescending post totally changed my opinion about you, now in my eyes you are not asking a dumb question at all, in fact you are almost smart.

    So thanks RHuckster for showing me the light

     

    Sorry, I really didn't mean to be condescending. This just happened to me yesterday and my mistake just reminded me of this thread, and decided to post about it out of boredom and partially to vent about my apartment.


Log in to reply
 

Looks like your connection to What the Daily WTF? was lost, please wait while we try to reconnect.