A more detailed view of my processor

If this page hasn't displayed properly (the pictures should all fit together to make a big block) then please read the notes below the diagram.

The silvery clip around the heatsink and the fan is made from a blanking plate that was bent to shape by banging it flat with a knife and then bending it using a drawer as a makeshift vice.

The heatsink is truly vast.  It extends a good inch above the actual processor and never seems to get even slightly warm.

The fan was ripped out of an old PSU.  It's powered from the PSU, not from the motherboard, as it might be too powerful for the motherboard (I'm not sure what power the motherboard can supply without breaking). l I happened to have an old cable that was perfect for connecting the thing.  It's a short length of four pin drive cable, with a male connector at one end and a female at the other end.  In one end there are two wires spliced into the 12 V and a ground wire.  It's to these wires that I've connected the fan.  It is incredibly quiet, and blows rather well.

Both fans blow air from right to left, as does the PSU fan.  The PSU is just a few inches above the picture.

The wire sticking into the heatsink is the thermistor probe that came with the motherboard.  It's the only way that I can have any indication of the processor temperature at the moment.  The Abit BX6-2 motherboard has wiring on it to enable it to read the thermal diode within the processor cartridge itself.  Unfortunately, later revisions of the board, mine included, are missing a capacitor and a resistor from the motherboard, as Abit claimed that the thermal diode gave inaccurate readings.  This is contrary to many people's beliefs, and it would seem that the diode gives the most accurate reading of the core temperature that you can get.  It is possible to solder on the missing components; this would enable the Winbond temperature monitoring chip to read the thermal diode.  However, I've not got the skill or the courage to solder something onto my motherboard.  I might ask someone better at soldering if they can help.

You'll notice that the location of the power connector on the motherboard is particularly unsatisfactory.  It means that the power lead interrupts the airflow over the RAM, and is stretched quite tightly over the processor.  I wish the effort had been made to locate the connector along the side of the motherboard nearest the PSU so that it wouldn't get in the way.

Browser difficulties with this page

I don't think that this page works at all in Netscape.  The HTML is all correct, and the page works in Opera and Internet Explorer.  It's just one of those things, I guess.  Netscape 4.51 doesn't seem to be at all capable of displaying the page; it doesn't even work when I use Netscape Composer to create the page.

However, Internet Explorer might show the page buggily; the version provided with Windows NT 5 beta 2 certainly has slight problems.  In that browser, and perhaps other versions too, to make the page display properly, you must do the following.

Go to View...Internet Options.  Go to the Advanced tab.  Enable and then disable any option, then click Apply.  The page should be redrawn, and it gets redrawn properly.

On Internet Explorer 5 for Windows 98, this method doesn't work.  However, what does work is pressing refresh a few times.  Eventually it gets it right.  I really don't know what's wrong with it.  If I build the table in FrontPage Express or Netscape Composer it still goes wrong.  I draw the table, and it's fine.  Then I start adding a few pictures.  Still OK.  Then I add one more, and bang!, it breaks.  Take the picture away and it's fine again.  There's nothing wrong with the pictures or anything, nor the HTML.

Some builds of Internet Explorer work fine, others work after a fashion.  I've not tried it with IE 4, as I don't have a copy to hand, unfortunately.

This makes me rather cross, as I've recoded the page several times this afternoon to get it to work, not suspecting that it was a bug at work; I assumed that something was wrong with the code but then I noticed that Opera worked properly, so I suspected foul play from my browser, which appears to be the case.

Back to Top