I've seen this problem before and it's usually caused by one of two things:
1) Bad OS installation - If XP hasn't been reinstalled in the past 6 years, there is a decent likelihood that something has become corrupted. Try to boot into safe mode by pressing F8 right after the BIOS screen and see if you can get in to the OS for more diagnostics. 2) Hardware problem - Two possibilities here. First reseat *every* connection in the system as small amounts of corrosion on the connections can affect things after long periods of time. If this doesn't fix the issue, it is possible that your RAM has degraded and is no longer stable. Try running with less memory and use different sticks on each attempt. If this solves the problem, you can just run with less memory or purchase replacement chips. Good luck.
I suspect that this is caused by a flaky video cable or connection, not anything to do with power. I would try opening the display and replugging each of its connectors. Doing the same for the logic board connection would also be a good idea.
While not a direct answer to your question, this Tom's Hardware article has a lot of pertinent information. Specifically: "If your SATA controller does not recognize PMs, it will run all attached hard drives as if they each were attached to a separate SATA port." "Port multipliers can be used to run up to 15 SATA drives via a single SATA connection. Hardly anyone would operate so many drives over a single connection, because the 300 MB/s link will become a bottleneck."
I would suspect that the RAID battery (similar to CMOS battery only usually larger) which allows the configuration to be stored has died. I would look for the battery (sometimes they are primary lithium coin cells and sometimes rechargeable NiCd/NiMH) and determine if it needs replaced. Remove the battery and check it with a voltmeter. Anything under 3V for lithium and under 1.15V for NiCd/NiMH (per cell) is probably dead.
I have seen similar problems on other laptops and multiple iPhones. Usually a sharp blast of air (from mouth or can) will cause the contacts to reseat properly. If not, I have also had success with somewhat violently inserting/removing a headphone plug repeatedly. If neither of those solutions work, opening the computer and fixing/replacing the jack manually is probably the only solution.
I'm not exactly sure what your problem is, but it is probably one of the following: CLOGGED PRINT HEAD
I usually remove the print head (I believe on your device this requires taking the printer apart and removing the carriage) and soak it in rubbing alcohol. I then use Qtips with rubbing alcohol to clean all of the remaining ink-paths that exist in the printer. It's messy but usually cleans out a clogged head. CARTRIDGE ISN'T BEING VENTED
First off, make sure that the little piece of plastic covering the vent hole is completely removed. On some models there are multiple holes being covered and it's easy to only pull the cover off of the first one or two. In an extreme case, I've used a small drill bit to create my own vent hole that I knew for certain was open. MANUFACTURER LOCKOUT
According to this blog, there is an ink-level sensor on your device that might be malfunctioning or otherwise be in need of bypassing. Something to check. DATA CABLE NOT WORKING
Each print head has a data cable going to it and s...
I'm not sure if "keep maximum power on to the hd" is the power mode it operates in or if incorporates other settings such as spin down time. Make sure the drive isn't just going to sleep by upping the sleep timer. This would cause both a click and a noticeable pause to your OS. David is correct though, first thing to do with any questionable drive is to pull your data from it.
I have seen many G5 iMacs fail due to both PSU issues as well as capacitor problems. There are two types of caps on the logic board and (like most caps) when they die, they bulge and leak electrolyte. Before checking the PSU, make sure that all of the capacitors look good. They /are/ replaceable if they're bad and you're handy with a soldering iron.