Sounds like a FSB 0f 83 MHz - which is a common setting on Super Socket 7 boards.
3,5 x 66 = 233 MHz
3,5 x 83 = 290 MHz

https://theretroweb.com/motherboards/s/gigaby … -ga-5ax-rev-3.0

Please try to reseat JP7.

The manual https://theretroweb.com/motherboard/manual/ga … d1809228447.pdf has pictures on page 5/6 showing settings for 233 MMX. Looks pretty much as you have set them. But 290 MHz indicates it runs 83 FSB and 3.5x multiplier. Maybe JP7 doesn't connect right? I have seen jumpers like that with dodgy connection ... JP7 looks like it's the only difference between 83 and 66 MHz FSB.

No joy. I tried different jumpers to close JP7 2-3. They definitely seat well. No change. I also booted with the jumper removed. Still showed 290Mhz. Could that reinforce what's being suggested? Motherboard issue?

Since its working otherwise then pin 7 of U11 ICS9148 is soldered well, without it there would be no PCI clock for the Chipset. Whats left are those two tiny resistors between U11 and JP8 JP7. I suspect one of them links pin7 to JP7. With cracked resistor JP7 wont be able to ground pin7. You can just measure resistance between U11 pin 7 and middle pin of jp7, should be ~10Kohm.

Are you 100% you have the revision of the motherboard right?

Back then, there were a few boards that had the documentation for some of the jumpers backwards.

And yes, revision of the motherboard matters. They changed stupid stuff like this in different revisions.

ignore me

If you have an oscilloscope you can remove the cpu and test all the combinations to check your board's revision.

rasz_pl wrote on 2026-03-04, 10:20:

You can just measure resistance between U11 pin 7 and middle pin of jp7, should be ~10Kohm.

Are you able to see which resistor you want me to check in this image?

I can micro solder, multimeter, o-scope, etc.

Rev 3.0 is printed on the board. So I am assuming that is correct. I did notice different jumper settings in the Rev 5.2 manual. I am looking at what I'm pretty sure is a Rev 3.0 PDF.

I have another Pentium MMX 233 CPU. Is that worth swapping to see if anything changes?

Are you recommending to check for 10Kohm across these points marked in red?

I bet the fsb jumper instructions are wrong. No harm in trying the different combinations. Look up the datasheet for the ICS clockgen. There are usually multiple settings for each fsb selection, even if not all of them are wired up.

Most boards also do not document all the different settings available.

Some are different Fab settings and some are different dividers for PCI speed.

I'd echo just trying the different jumper combinations and seeing what FSB speeds are the result.

I've done this before with 486 boards and there are sometimes additional speed possibilities not covered in the manual.

I swapped to the other CPU without changing jumpers or dip switches. This one shows 233MHz at POST.

Ok confession. The image above of the bottom of the CPU, is the CPU that is in it now at 233MHz.

This image I am attaching in this reply, is the CPU that was in the system showing 290MHz. I lazily used the picture because... well it's the same model number. But, perhaps there is a difference afterall?

I did measure the resistance between U11's pin 7 and JP7. It's 10k. Without a jumper attached, it's JP7-3

Anyhow, I don't know if this means something is wrong with the first CPU. I sure would like to know if anyone has some theories or past experience.

Thank you

What is the weird coloration around the middle in the bottom of the picture? Almost looks like 2 pins are bridged?

It was dust/dirt. So I looked closer at this under the scope. Many pins are bent, but also bent at their tips. It was probably some connectivity issues.

bytesaber wrote on 2026-03-05, 02:57:

I did measure the resistance between U11's pin 7 and JP7. It's 10k.

good

bytesaber wrote on 2026-03-05, 02:57:

Without a jumper attached, it's JP7-3

wait what? what does JP7-3 mean ? 😀 pins 1 and 3 of jumper JP7 should be connected to ground and 3V

I have no idea how a CPU can influence clock produced by ICS clockgen. Failing global timer ticking slower could make bios display higher frequency.

rasz_pl wrote on 2026-03-05, 05:34:

wait what? what does JP7-3 mean ? 😀 pins 1 and 3 of jumper JP7 should be connected to ground and 3V

JP7-3 was my dumb way of saying, JP7 pin 3. I do see pin 1 and 3 are connected to ground. I was measuring 10k from U11 Pin 7 to Pin 3 at JP7 while the jumper was pulled off / all 3 pins open. So the same as if I measured to ground.

Does this clear up, and match what you are describing as well?

bytesaber wrote on 2026-03-05, 06:25:

JP7-3 was my dumb way of saying, JP7 pin 3.

that makes me question how that thing even runs in the first place?:)

bytesaber wrote on 2026-03-05, 06:25:

I do see pin 1 and 3 are connected to ground.

😮 , one of those pins MUST be connected to 3 or 5V, and for 83MHz FSB U11 Pin 7 needs to see pullup. Cant make a pullup with a ground on the other side of 10K resistor 😀

bytesaber wrote on 2026-03-05, 06:25:

I was measuring 10k from U11 Pin 7 to Pin 3 at JP7 while the jumper was pulled off / all 3 pins open. So the same as if I measured to ground.

thats wrong. look at GIGABYTE GA-5AA rev3.2.PDF diagram Re: 430VX or 580VPX schematics
your board has same clockgen circuit, just different part designations. Afaik your JP7 is JP1 on this diagram, and 10K R86 from the diagram might be one of those blue resistors between U11 and jumpers.
I have a suspicion there is a crack on the board and the process of changing CPUs flexed it a little temporarily fixing, or jp7 jumper sucks

rasz_pl wrote on 2026-03-05, 07:28:

I have a suspicion there is a crack on the board and the process of changing CPUs flexed it a little temporarily fixing, or jp7 jumper sucks

Much more likely than the cpu being able to change FSB, yes. Probably a cracked solder joint.