Possible stupid question related to G4 Dual 867

[nanopico]

So  I have this Dual 867 which when working is awesome.
But some times it doesn't.  As in I will shut it down and the next day I boot it up and the fan's just turn on then they turn on really fast and the power button will no longer turn it off if you hold it.
So I have figured out that when this happens I pull the power cord to turn it off (yeah bad idea number one), then i plug it back in, then (here comes bad idea number 2) I remove the video card and then plug it back in.  At this point plugging it back in causes it to power on and actually boot correctly.

Any one had a similar situation with some of the MDD?

[mrhappy]

Try running the diagnostics here on the site. Maybe you have some funky ram or something else that might be identified.

http://macos9lives.com/smforum/index.php?board=33.0

[MacOS Plus]

  Yikes!  Are you saying you're pulling and re-seating the video card with standby power present?  Even if you pull/push perfectly straight on an AGP card there's a good chance of shorting the pins because of the close-pitch staggered pattern.  I managed to cause quite a spark even with a PCI card once.

  The kind of behavior you're experiencing should be related to one hardware device/component that is confusing the system.  While it may be a sign that something is failing, often it's a mild compatibility issue with RAM or an expansion card, or even just something slightly mis-seated.  If I cause a hardware-related freeze and then reset the computer from the front reset button, it almost always does the same thing you're seeing, and then I have to use the internal reset to shut it off first before powering up again.  Because of all the crazy experiments I do with hardware I've seen this happen a lot with MDDs.

  I'd check pin oxidation and seating of the video card first, replace it with another one second, then look at RAM and any PCI cards third.  If you get as far as checking the RAM, install only one stick at a time in the first slot and see how the machine responds.  The RAM timing is determined by the first stick in the group, which can result in strange issues with unmatched ones being pushed past their reliable timing limits (no matter what their own actual speed chips say about them).  I had one machine working for months, when suddenly one day it insisted one of the sticks was incompatible.

[nanopico]

  Yikes!  Are you saying you're pulling and re-seating the video card with standby power present?  Even if you pull/push perfectly straight on an AGP card there's a good chance of shorting the pins because of the close-pitch staggered pattern.  I managed to cause quite a spark even with a PCI card once.

  The kind of behavior you're experiencing should be related to one hardware device/component that is confusing the system.  While it may be a sign that something is failing, often it's a mild compatibility issue with RAM or an expansion card, or even just something slightly mis-seated.  If I cause a hardware-related freeze and then reset the computer from the front reset button, it almost always does the same thing you're seeing, and then I have to use the internal reset to shut it off first before powering up again.  Because of all the crazy experiments I do with hardware I've seen this happen a lot with MDDs.

  I'd check pin oxidation and seating of the video card first, replace it with another one second, then look at RAM and any PCI cards third.  If you get as far as checking the RAM, install only one stick at a time in the first slot and see how the machine responds.  The RAM timing is determined by the first stick in the group, which can result in strange issues with unmatched ones being pushed past their reliable timing limits (no matter what their own actual speed chips say about them).  I had one machine working for months, when suddenly one day it insisted one of the sticks was incompatible.

No not with stand by power. Just fully powered down, but still plugged in, but like I said my fix is a really dumb idea.
It's not the first time though I've had "brilliant" ideas for fixing stuff.   Primary reason I'm not a hardware guy and more a software guy.

When it first started happening I bought all new RAM and tried it, with no luck in stopping it. So I'm fairly confident the actual ram is not bad, but the sockets could def be oxidized.  I'll clean all connections.

Try running the diagnostics here on the site. Maybe you have some funky ram or something else that might be identified.

http://macos9lives.com/smforum/index.php?board=33.0

I haven't run hardware diagnostics on it lately, but probably a good idea to try again.

[MacOS Plus]

  There's potentially still some live power to the slots in that state.  Harder to do major damage though.  I'm more inclined to suspect the video card as the culprit than the RAM.  Power supplies on these models were extremely prone to failure, but since you can boot it properly that's not likely your problem.  (Two out of three MDDs that came to me second-hand had dead power supplies.)

[nanopico]

  There's potentially still some live power to the slots in that state.  Harder to do major damage though.  I'm more inclined to suspect the video card as the culprit than the RAM.  Power supplies on these models were extremely prone to failure, but since you can boot it properly that's not likely your problem.  (Two out of three MDDs that came to me second-hand had dead power supplies.)

I'll mess around with it more over the weekend. I Just found it funny the steps I take to make it work again.