Drugwash

It just occured to me: if the board only has one IDE port, it's possible that the BIOS assigns it as tertiary/quaternary/etc when SATA is in IDE mode, which may be the reason for the OS not to detect the devices (properly). How are the SATA/IDE ports displayed in BIOS when SATA are set to IDE as opposed to native (or whatever "normal" mode)? At POST time, can you actually see how the devices are detected and displayed on screen? Can you freeze the screen by quickly pressing the 'Pause' key on the keyboard and write down the IDE-related details or take a legible screenshot? Oh and just to eliminate one of the possible issues, please do set the devices manually as Master and Slave as there is a slight possibility that either the cable, devices or BIOS might not "like" Cable Select.

I just looked at the specs on the site. Ouch! video X3000? SATA II? Gigabit LAN? HD audio? I wonder what's actually usable in 9x on that board, except for half or quarter CPU… Also, Gigabyte are known for messing with the BIOS (maybe not as much as Soyo or Biostar but still) so it shouldn't wonder if the IDE issues stem from their "improvements". But I guess Mr. Loew might offer much better explanations and details.

Any BIOS update available for the board? Risky but may fix something. I'd say 95% you'll need the SATA patch but the rest 5% are reserved for BIOS quirks.

Yep, seems I overlooked the '2014' there.

Oh, you sure know how to get into trouble. If I were you I'd return it and ask for a refund. (not helpful, I know)

Go to Control Panel > Add/Remove Programs, there you'll see Windows XP Service Pack 4 to the bottom, if it's installed. There will be no mention of it in System Properties, you'll see the same as in the screenshot above. In the same Add/Remove dialog click on Add/Remove Windows Components. You'll see a list of components added by SP4, some will be unmarked (such as .NET 4.0). You may mark/unmark the components you want to install/uninstall respectively. I hope you did let it finish the installation. There is one process you need to watch (forgot its name but it's mentioned somewhere in the readme) and only when that process dissapears from the list of active processes in Task Manager the installation would have completed.

JFYI, it didn't work for me even when using No-IP services. There is/may be something in the ISP's infrastructure that disrupts communication. I can't access my HFS server from the inside using the external IP (82.137.x.x class) but I can access it through an intermediary IP (10.x.x.x class). However nobody from the outside can access it on either IPs. That's for a USB GPRS modem (ZTE model MF-190) but the same may be true for other (modem) connection types. Unless I'm really clueless and there's actually much more tweaking involved.

Unfortunately in situations like yours it's extremely hard to determine the faulty component(s) and maybe impossible to fix if any of the tiny SMDs (surface mounted devices) are damaged. The battery issue does strongly suggest something like this. A little elaboration: BIOS settings are kept in the volatile memory by means of stand-by 5V voltage from the PSU. The 3V battery is only meant as a backup during power shortage, unit displacement - anything that may interrupt the power for even the shortest amount of time. As such, there is a relatively simple circuitry that switches between stand-by voltage and battery voltage as needed, in order to keep the volatile memory of the BIOS powered at all times. Usually it's one or two transistors, a few resistors, maybe diodes and capacitors. Now, if that circuitry becomes defective for some reason, it can block the power from both the PSU and the battery, thus putting the BIOS chip in an emergency state. Sometimes it can deplete the battery in a very short time due to a shortcircuit. Such defect may only be repaired by qualified personnel, if there's anybody out there that still cares for fixing motherboards and other similar devices (such as mobile phones, for example). Unless you personally have the skills and the appropriate tools, of course. The fact that the board is flimsy as you say may have lead to PCB tracks breaking as I mentioned above. Those tracks are so thin that the slightest bend can break them and it's not only the visible ones on both sides but there are also a couple or more internal layers where tracks can break or blow due to high current (shortcircuit or heavy load). The northbridge fan may not be a problem, at least not in the early boot stage. That one can easily be cleaned and greased (usually), just be careful with the flexible safety (dunno its actual name) on the axle, not to break or lose it - it tends to spring away. It is kinda curious though to employ a fan but not a heatsink - it usually happens the other way around. I have a relatively new machine here donated by a friend (with a Gigabyte board, incidentally) which sports a quite large heatsink for its chipset but no fan; I've had intermittent crashes/reboots with it, inability to detect the IDE drive in BIOS and - coincidentally, maybe? - it all went back to normal after I attached a small fan to that heatsink which used to get really, really hot. Here it is: GA-M55S-S3 rev1.1. Anyway, since that board will be spare when the new one arrives, you may play with it as you like but still be gentle, try not to consider it dead until it actually dies. Good luck!

Just for testing: download and run rejetto's HFS (HTTP File Server) from here, switch to Expert Mode and poke around Menu > IP address (use Find external address), see what IPs it finds. Then try to reach HFS using each of those IPs (if more than one). That may offer a hint on what type of network you're connected to.

Then there may be a problem with the DSL modem and/or the router not forwarding certain HTTP(S) calls. There are also certain services (such as Web server HTTP and Secure web server HTTPS) that may have to be enabled/tweaked. Not sure how to access them in your OS version - in XP they're under Windows Firewall > Advanced > (select your WAN connection) > Settings > Services. Also do take into account that the ISP may be bluntly lying about that, but keep it as the last possible explanation. Replace (temporarily) the DSL modem and router, if possible. Eliminate the router (temporarily) from the configuration, see if it works with a direct connection to the modem. Unfortunately this is not my expertise so do keep asking around and testing different configurations. Good luck!

You may have to get in touch with your ISP and ask them firmly if they somehow block this type of setup on purpose or there is some technical reason for it. I myself have a GPRS connection only and having the same problem not being able to reach a locally hosted server from the outside. It's possbile that any connection going through any type of modem (as opposed to direct cable connection) cannot support this setup. But better check with your ISP.

Correction: want to. They don't actually need to survive. At all. They're done for.

Careful with the CPU socket, the contacts inside are very fragile and the cover can easily break as I said above. Since the CPU itself has been moved around, do examine its pins too with a magnifying glass, one or more may be bent. Those are fragile as well. What I forgot to mention: - if the BIOS chip is in a socket, you may try to pull it out (using a couple of needles inserted in opposite corners) and insert it back a couple times (check the position of the cut corner) - make sure you insert the battery the right way; I have a board with a vertically mounted battery socket and it's easy to insert it the wrong way. Measure the voltage on the battery (or replace it - it could be shortcircuited) - worst possibility is for the board to be damaged mechanically, which may happen when mounted too tight in a flaky case which is then twisted whilst handling it or when inserting RAM/ATA or floppy cables/power + LED + speaker wires; that could explain the faint/missing sounds too

Never met that beast and hopefully never will because I may have to pay the owner after smashing the thing into pieces. Oh and remember AMI for regular desktop machines (but not only) - guess they come second, very close.

Yeah, the right tool may be the key; what I used to have could only flip the removable bit off but not back to on. There are many tools that can prepare a bootable pendrive: Unetbootin, YUMI, Easy2Boot are the ones I have at hand in downloads. There are still problems in certain BIOSes though. I have a couple relatively new machines here: one would boot with a certain Linux USB stick, the other wouldn't - regardless of which option I set in BIOS. Laptops may be even more selective. More USB preparing tools have to be tested before finding the one that produces a usable drive for the particular machine being worked on.

Because power comes hand in hand with paranoia, that 's why. Regular, normal people mind their own business, they don't stick their nose in someone else's lives because they don't have the knowledge, the technology and the spare time for that. And they usually have no reason either.

There is a trick to make a removable flash drive appear as a fixed drive but that is irreversible AFAIK. Your BIOS probably doesn't accept booting from removable drives. What exactly are the boot options in regard to selectable devices (USB Floppy/USB HDD/USB CD-ROM/USB ZIP/etc)? There is a DOS driver that can detect USB storage devices, you may have to load it in autoexec/config.

Not turn off, but set explicit compatibility. I had an issue with some application where it wouldn't work with compatibility disabled and required strict 98SE compatibility. I just checked, apparently sys files can use KernelEx compatibility mode. Do try setting it to 98SE. Such kind of issues may be the reason I stopped using NTFS driver long long time ago - can't even remember if it was Sysinternals' or the other one (forgot its name).

Probably the driver and the utility check for the presence of certain NT-only components in the system and they modify their behavior on-the-fly accordingly. Which means you can't have both the wireless and the NTFS driver at the same time. Do you happen to have KernelEx installed on that machine? If so then try to set explicit 98SE compatibility to the utility and any related driver library that allows it. But that's a long shot.

Yeah, I have a pretty old MSI board - something with 845P Neo - that used to boot only once in a few dozen attempts, then more and more rarely. At some point I tried a BIOS update which I can't recall if it went through or not, at which point the board died completely, the diagnostic LEDs on the bracket stay constantly lit green. BIOS chip is soldered on the board so no way to take it off and do a hot flash. While it did work, the external SATA controller chip used to get quite hot which I don't know if it was normal or not. Disabling it in the BIOS appeared to improve board's stability but only temporarily. Capacitors were all looking good, no bumps or leaks, but no way to know if any of them were dry unless I had unsoldered them, which I didn't. Other than that there's a weird chip on the board called "MSI core cell" with a small cooler on top, which I have no idea what it does. First of all you should measure all the voltages from the PSU and make sure they're within the acceptable margins. Then examine carefully all the (large) capacitors on the board, see if there's any gonflated and/or leaky; those would have to be replaced before damaging anything else. The BIOS chip may be failing, but you can't know that for sure unless you replace it with same or directly compatible model flashed with the correct BIOS version. Other possibilities may be bad contact in RAM or CPU sockets. Sometimes RAM pins may get burned due to such bad contacts so pull them out and examine the pins at both RAM sticks and board memory sockets. Clean RAM stick contacts with a pencil rubber. Loose contacts in CPU socket could also lead to weird behavior - take the CPU out, find yourself a good magnifying glass, gently and carefully remove the CPU socket cover and examine the contacts, see if there's any loose ones and if found use a thin needle to bend very gently the loose pins back to their normal shape. Please be extremely careful when you remove the socket cover, it can easily break. Same with the contacts. So don't try it if you're not sure you can handle the operation and the possible loss of the board.

I just watched again last night the 1999 movie "Pirates of Silicon Valley". Reminded me how Billy got the richest "kid" (at the time). Thing is, Microsoft are greedy - they don't want only the revenue from the OS, they want it all! Their apps, their office suite, their browser, their everything should be used exclusively and all other software - usually much better/bug-free/feature-rich/etc - should just die by simply shoving it off the users' systems every few months. Isn't that how cancer cells work, by eating the good ones? There's a reason why alternative software exists: because all people are different, think differently, act differently, have different needs and different preferences. When an operating system is so anal that it doesn't even allow a large diversity of themes unless it's being hacked - as was the case starting with XP - then that should already raise a red flag as large as Billy's house. Reverting all user settings, file associations and uninstalling or blocking legitimate applications should raise a red flag as large as the continent itself. But is God still watching - assuming He does exist…?

OK, now approximately how large would the Windows 95 OSR 2.5 installation kit would become if all of its icons would be replaced with scalable 256x256px images (assuming it could decode PNG)? No, better let's take Windows 98SE which is already larger. Would it get to the current size of Win10 which is about 6GB if I'm not mistaken? Of course, you'd say there are lots of new drivers inside and all… but what else is it there that is actually useful? Actually, what else should it be in an operating system, other than rudimentary means to access local or remote resources? Because the OS is not obligated in any way to provide a full set of professional tools. It should only offer the ability for the user to navigate their internal/external fixed/optical/flash/etc drives as well as the local network and the Internet in order for them to retrieve the desired software and install it, at which point that software would take over the file types designated for it. That would be a normally set up system. The user should have unlimited options to install/uninstall any applications as they see fit, without the operating system denying it or reverting the changes through a forced update, as Windows 10 does. When an operating system acts as if itself should be the only software running on a give computer then something is very wrong with whoever designed it. It's a cancer!

I started with a Z80 computer and stuck my nose in a lot of its games. Those were complex enough already and they also fit in 48kB of RAM (as the rest of 16kB were reserved for BASIC and the video memory). I always considered machine code as the best language because there is no room for error there, everything must be perfect and done manually instruction by instruction. Nowadays, with all different instruction sets and even hardware bugs in CPUs, everything relies on compilers which is the first possible source of bugs apart from human error in the code itself. If it weren't for the money people would create out of sheer passion and creativity, they'd compete for the best, smallest, feature-richest and so on while at the same time helping each-other. Everything would be better, close to perfection. Or at least let me dream it would.

I remember the times when programmers actually could code. Anybody recalls kkrieger, the 96kB 3D first-person shooter game? Bigger is not better - bigger is lamer!

Programming is not a walk in the park. I'm an amateur programmer (at best) and even simple scripts I build for my main 98SE system sometimes require major tweaks in order to be able to run (correctly) under XP, 7 or later. A project as complex as a modern browser would require much more for an (almost) perfect behavior under different OS versions, especially considering security. Of course, some of us don't care much if at all about that, we don't do online banking/shopping/etc., we don't keep critical/private data on our (browsing) computers but many people do and they need to be safe. For now there is the Atom/XP build of Pale Moon and that should be enough for us. Alternatives are provided on the very page linked above. Anyone up to maintaining a full-XP compatible fork of main-line Pale Moon in a long term…?