I can't speak for Your exact motherboard, but I found the post I made here years ago regards performing the 5th bit enable mod on my KT7-RAID KT133 (non A) motherboard.
I've pasted the important part.
I soldered a switch to enable and disable this mod, otherwise You lose the original multiplier settings.

BIOS MULTI= CPU MULTI
5= Does not boot
5.5= Does not boot
6= Does not boot
6.5= Does not boot
7= 15 for 1500mhz (100fsb)
7.5= 22 for 2200mhz
8= 16 for 1600mhz
8.5= 16.5 for 1650mhz
9= 17 for 1700mhz
9.5= 18 for 1800mhz
10= 23 for 2300mhz
10.5= 24 for 2400mhz (was running 3dmark99 until it crashed)
11= Does not boot
11.5= 19 for 1900mhz
12= Does not boot
12 above (12.5)= 20 for 2000mhz

"Does not boot" can be resolved by removing the AC cord for a few seconds, then holding down "delete" during boot, sets
it to a multi of 20 for 2000mhz

Thanks. Yeah i was planning on installing a switch, I don’t currently have the correct patch wire in order to do this. I could see the wires I was planning to use had many near-invisible strands, and was worried about short circuits.

I’ve found a few lists for the new multipliers with the mod enabled but thanks for sharing it. For my board, all of the available multiplier combinations are accessible for me, between 13-18x, I haven’t tried above this as I want to keep 133+ FSB. I didn’t find a multiplier combination that didn’t require me to pull the power and disable/enable the dip switches though.

I will try pressing del during boot to see if it does anything!

My other thought is, now I’ve settled on a multiplier for this CPU, could bridge the other pins on the bottom of the board that I need to, in order to lock the multiplier at what I want. It’s not ideal and shouldn’t be necessary because these dip switches on the board should be doing the same job? I did try testing the dip switch behavior via the pins on the socket but the results were inconclusive, so perhaps the are not directly connected to the socket and are instead doing something via the chipset? I think I’ll try this anyway and report back. I can always reverse it if it doesn’t work

Ok I have made some progress here. I went back to 100MHz FSB and discovered that the computer now POSTs normally up to 22x multi

It will not post without the trick described in the first post, at any setting above 120MHz FSB. 112 and 115MHz post but are unstable. 120 tries to post but I get a blank screen. This is more than I get at 124, 133 and up. The monitor just goes to sleep.

120 is the highest you can go with PCI = FSB/3 everything above that is PCI=FSB/4

So I am wondering what could cause this behaviour. It can be made to post at 133, 140 and 150 FSB with the trick mentioned in my first post and 133 and 140 are both stable and I can go to about 145MHz before I start to get any stability issues. I am wondering if there is an issue somewhere on the board where the PCI divider isn’t being set correctly. A bad solder joint perhaps?Does anyone have any ideas? I’ll have a poke around to see if I can find anything in the meantime

Could be clockgen. Some were better than others.
Look for a flat ~28-48 pin chip with a bunch of tiny components and 1 silver metal thing.
Take a photo of it that is readable.

This might take a few attempts. Many old markings are faint

Assuming trw is correct, it should be this.

https://theretroweb.com/chip/documentation/pl … 03147121309.pdf

Thanks for the replies. Yes I’ve seen this chip. It’s next to the first DRAM socket and is a Phaselink chip. I will get a photo later on and upload

Well, assuming that is it, look at page 3.

Your jumpers /softbios are seeing those speeds.
If it has jumpers, dont be surprised to see additional settings that aren't documented.
If it's soft,since it overlaips, they may have chosen settings that you wouldn't choose.

Yes I’ve just been looking at it and comparing with the motherboard manual. As you mentioned, several options are undocumented in the motherboard manual.

It looks like there is another way of setting 133MHz. I2C byte 0 bit 2 is enabled in the manual for 133MHz but it can also be disabled. It’s disabled on all of the frequencies that currently work without issue

I’m trying to understand what this bit is doing. It’s influencing the clock output when in some combinations but not in others.

I assumed this bit was the FSB:PCI divider

I shall try setting this to disabled, see what happens

Eyeball math
+5% on some frequencies?

The chip is a Phaselink PLL205-13. I forgot to take a pic (I will sort later) but I found another datasheet.

Disabling the last bit didn’t work. I checked a few of the other clock speed options and they don’t line up. I need to make a spreadsheet to compare properly

Remember, someone manually typed the speeds you see in the bios.

I’ve made my spreadsheet and spotted a few misalignments. The switches are in a different order in the manual vs in the datasheet. Once I figured out which line was which I could see that all of the frequencies up to 120MHz line up, one of the 133MHz options line up, but most of the rest do not seem to match

I am wondering if this chip has been replaced and the original was slightly different. I guess there are multiple versions of this chip with different options and perhaps behaviours?

Let's see your table

But also, try an undocumented one and compare a simple benchmark like superpi

Apologies about it being a photo of a screen. Took this earlier so I could refer to it on my phone later

The ones that don’t match seem to match another slightly higher frequency

I’ve been checking with CPU-Z and Everest what speeds I’ve been getting together with benchmarks in 3dmark99. This CPU tops out around 2200MHz before I start having stability issues on stock voltage. The chipset seems ok up to about 145Mhz, though I’ve not found the absolute limit yet.

Forcing to boot at 133+ frequencies is yielding better results. My best 3dmark 99 scores are just over 15,000, with best CPU scores about 35000. This was at 144 x 15

The best scores I’ve had without a boot issue (by using the lower FSB options) is around 500 points below the above GPU and 1000 below the above CPU scores. This was at 110 x 19 with the memory running async at 143MHz

I tried a few undocumented options yesterday but nothing worked. This was before i figured out the jumpers are not in the same order in the manual as they are in the datasheet, so I’ll be trying more later! If I can get just one of the higher ones to work (where the PCI divider is 4 instead of 3) I can do the rest of the o/c in the bios

There's no particular reason to trust the usually hand-typed lookup table values in the bios.
Datasheets are pretty reliable.
Clock crystals can go out of spec or be different to specified by the clockchip maker.
Jumpers are wired in whatever order is easiest to route for the boardmaker.

The gold standard here will be a pure cpu benchmark like superpi to see the actual changes.
They should line up with the datasheet even if the bios says otherwise.

I tried the other settings from the datasheet but none of them work for some reason, with or without the reset trick. I only tried the ones in my table as most of the other additional settings aren’t particualrly useful

I should be getting an oscilloscope for christmas(!) so I can check actual PCI clock frequencies (and infer the FSB from that) and I’ve also ordered a post code analyser