You should make sure your power supply is up to the task. I need 400 watts to get mine going.

I am using an almost new 450W PSU. It works OK on the working Rev 1.1 Epox "P=" BIOS sticker board. (This board has 1 cap that looks in bad shape - the rest are OK however.)

Here's the seller's photo of one of the non working boards - see the replacement caps just east of the RAM slots.

I've just got this "non working" board to boot up Win98 OK - this is the board featured in the photo above. I did 2 things: I removed the Athlon XP 1700+ rated CPU, (and didn't try the T-Bird 1400 again), but instead tried a 3rd CPU - a "simple" Duron 600. The other thing I did was remove my anti-static bubblewrap base. That sounds a bit mad, but I did notice that on the underside of the board, there was "goo" on the solder points were the new caps have been fitted. Maybe this goo was causing some problems with the anti-static bubblewrap?

Also, I noticed on the BIOS POST screen that the BIOS is really "new". It's dated 04/09/2003. That's quite modern for this board. It's only an "E+ sticker" Revision 1.0 board. (With that dated BIOS, I guess it should be able to handle the XP 1700+ without any problems.)

I tried 5 CPUS. The BIOS settings were set to "failsafe" mode. The FSB was set to 100, not 133. The RAM used was 133, and the BIOS POST screen identified it as 133 speed.

Duron 600 = OK
Athlon T-Bird 1200 (200 FSB model) = OK
Athlon T-Bird 1400 (266 FSB model, running at 200) = the PC froze after about 90 seconds. I tried several times. Same result.
XP 1700+ (running at 200 FSB) = PC froze after about 15 seconds.
XP 2000+ (running at 200 FSB) = PC froze after about 5 seconds.

All 5 CPUs were identified correctly on the BIOS POST screen. It seems that the faster CPUs caused the PC to freeze in the shortest time. There is no fan on the chipset heatsink. I wonder if the chipset chip is damaged? I have a good heatsink on the CPU, and a decent fan. I have also temporarily placed a fan over the heatsink chip.

Who knows if they put in the right capacitors...
Matching voltage and capacity is not enough in such high-current cirquits.

I have now got the second "non working" board to successfully boot up windows 98. Just like the first "non working" board, I had to abandon using a fast CPU such as XP 1700+, and instead rely on using a slow Duron 600 CPU.

Curiously, if I "upgrade" the Duron CPU to something a bit faster (but not too fast), such as a T-Bird 1200 (200 FSB model), the onboard BIOS POST error LED unit says "FF" immediately when the power is switched on. I get no BIOS POST message on the screen - then I switch off the power.

This second board also has an "E+" sticker on its BIOS, but the BIOS POST screen reports this BIOS to be older than the first board. The screen says: 02/05/2002. (BTW, both of these boards are Rev 1.0.)

> I see that some of the caps are domed.

Why do you keep hitting a dead horse?

These Epox boards are notorious for having "bad caps". I've got 2 of these Epox boards, both with at least 1 "bad looking cap", but they both seem to work OK.

Edit: The board that seems to accept "slow" CPUs, such as 200 FSB CPUs, but will not accept fast CPUs such as XPs, has no domed caps.

As mentioned previously, I'm testing a third Epox EP-8KTA3+PRO. Unlike the other 2 Rev 1.0 boards, this one is Rev 1.1. Its BIOS is dated 02/05/2002, and has a "P=" sticker on the mobo's BIOS chip. Two of the large caps are a bit domed, but the rest look OK. This third mobo seems to behave itself the best. I tried an XP 1700+ CPU in it, and set the BIOS settings to 133 CPU clock, and the RAM to 133 speed. The BIOS POST screen correctly reports that its an XP 1700+ CPU. I get to the windows 98 desktop OK.

But if I replace the XP 1700+ CPU with a 2000+ CPU, I don't get to the windows 98 desktop. I either get a protection error or a fatal error just before the desktop is about to appear, and I have to switch off the power.

Edit: I removed the nVidia 4 series MX 440 AGP card, and replaced it with a Voodoo3 2000 PCI card. I was then able to get to the desktop for about 8 seconds, before the screen went black.

What's going on? Is the board unable to cope with the XP 2000+ CPU because of failing components, or do I need to increase the voltage settings inside the BIOS setup area? It's a pity that I've got no other PSU to hand. (However, the 450W I have should be OK.)

Don't different cpu's have different power requirements (which might explain why higher speed cpu's fail)? Might be a symptom of failing caps.

retro games 100 wrote:

These Epox boards are notorious for having "bad caps".

Thanks for the heads-up, I was thinking to buy one but maybe I shan't bother.

I ran pcpbench on the one "half working" mobo - the one which will not accept a XP 2000+ CPU, but will accept an XP 1700+. I was pleased with the pcpbench mode 107 (1280x1024x8 LFB) score of 66.5 fps. That was just using a "crappy" Radeon 7500 AGP card. Quake 2 @ 1024x768 timedemo gives 157.8 fps, which I thought was good.

I just DL'd the free "3DMark 99 Max" benchmarking utility from 3dfxzone.it, and ran it on this "dodgy" board I'm testing. As mentioned above, it's just got a "crappy" Radeon 7500 in it. Using the ATI control panel, I left all settings on their installed default values, apart from ensuring that the v-sync value was set to Off. The 3DMark 99 Max settings were also left alone on their installed default settings, and it ran the benchmark test at 800x600x16. I get these scores -

8031 3DMarks
24565 CPU 3DMarks

Do you need to change some jumpers or DIP switches on the boards depending on what type of CPU you're using?

I don't think so. When the XP 2000+ CPU is installed, the BIOS POST screen displays this CPU information correctly. (But the board freezes about a minute later.) I checked the manual for any interesting jumper options, and I see -

CPU Vcore Adjustable (0 = default, then +.1, +.2, etc)

I also see -

STR Function (I don't know what function this jumper is for)

I'm not sure if an XP 2000+ CPU requires the Vcore jumper to be set to something higher than it's default setting of 0v. There's also a similar voltage setting option inside the BIOS setup area, that probably does the same thing as this physical jumper on the mobo.

Apparently STR is "Suspend to RAM".

retro games 100 wrote:

I'm not sure if an XP 2000+ CPU requires the Vcore jumper to be set to something higher than it's default setting of 0v. There's also a similar voltage setting option inside the BIOS setup area, that probably does the same thing as this physical jumper on the mobo.

I has been some time since I used a KTA3+ Pro (mine is dead, rotten caps of course 🙄), but I think the onboard jumper can be used to add 0.1V or 0.2V to the voltage selected in the BIOS.

Which cores are your Athlon XP CPUs, Palomino or Thoroughbred?
Also, did you look at the hardware monitor in the BIOS? Is the core voltage and temperature OK?

Reading this make me feel pretty lucky mine works as well as it does. I'm happy to see that you were using a good power supply, I had some really strange results with an underpowered one when I started out. These motherboards have that nifty 80 port feature, why don't you cross reference the manual to figure out where in the POST sequence it's failing. Might have some good insight there.

The XP 1700+ CPU (which works) is a Palomino model 6 CPU -

http://www.cpu-world.com/CPUs/K7/AMD-Athlon%2 … X1700DMT3C.html

The BIOS setup "PC health status" section says -

CPU temp 35C
System temp 22C
Vcore 1.82 V
AGP Vcore 1.57 -> changes to 1.58 V, then back to 1.57 V again. Etc.
3.3V 3.50V
5V 4.83V
12V 12.60V

The BIOS setup "Frequency/voltage control" section says -

CPU ratio (Default setting) = x 11
Vcore voltage (Default) = 1.750
Vagp voltage (Default) = 1.5
Vio voltage (Default) = 3.4

The XP 2000+ CPU (which freezes the mobo at the windows 98 desktop) is also a Palomino model 6 CPU -

http://www.cpu-world.com/CPUs/K7/AMD-Athlon%2 … X2000DMT3C.html

The BIOS setup "PC health status" section says -

CPU temp 41C (and rising)
System temp 22C
Vcore 1.81 V
AGP Vcore 1.58 V
3.3V 3.50 V
5V 4.78 V
12V 12.66 V

The BIOS setup "Frequency/voltage control" section says -

CPU ratio (Default setting) = x 13
Vcore voltage (Default) = 1.750
Vagp voltage (Default) = 1.5
Vio voltage (Default) = 3.4

Update: with this XP 2000+ CPU installed, I got to the windows 98 desktop OK for the first time. I ran SiSoftware benchmark, and it froze. Reboot time...

As for the BIOS POST LED "debug panel", as far as the board is concerned, it thinks everything is OK, because I just see "FF" - thunderbirds are GO! - let's see some action! I think my next best move is to try an XP 1900+, "one notch" down from an XP 2000+. I'll report back when I've got one...