Nice,
I had built a celeron 333 computer back in 1998
I think I had it over-clocked to 400mhz. Maybe even 500mhz.
I was using a Stock Slot-1 CPU cooler
I think I was using an Abit BH6 motherboard.

Wow the Abit BH6 it's a very good Motherboard.
I have heard it can manage better vcore increase...i tried to find one online, but it costs so much!
From a "real test" today, with a friend, the Celeron at 525 Mhz beat a Pentium II 450 in most cases. I suppose that in 1998, this CPU with this clock, was a very good solutions.
I am so young, i I didn't experience that era consciously (being from 1994), but i can now understand why those CPU are so famous among overclockers.

Retroinside wrote on 2024-10-14, 10:15:

Hi guys! […]
Show full quote

Hi guys!

I was working in this days on a Slot 1 Platform (Iwill VD133Gold board), tryng to perform a decent overclock for a Celeron Mendocino 333 MHz.
Had a 90 radiator and an universal chipset water block aside. I easily recovered a universal pump from China and created a small, but efficient liquid cooling system for my Celeron Slot 1.
I'll be honest, I was hoping to get more. Unfortunately, however, the motherboard does not allow me to increase the CPU voltage as I would like. With some Jumpers , it allows me to increase the vcore by 10% or 5%. Same thing for the VIO too.
However I was hoping to be able to push it to something beyond 2.3 (I know that up to 2.5V is manageable, if properly cooled).

I attempted pin mod of pins a119 and b119. Without any success alas. The motherboard just won't POST with this change. I also tried covering the pins individually but again, no success.

I therefore ask for advice from any of you, if you have ever tried this practice, and can provide me another solutions.

At the moment I am stable with a poor vcore of 2.2 V, at 525 MHz (105 MHz FSB X5.0) but I can pass POST even at 550 (110 MHz X5.0) but it won't load OS.

I am therefore sure that already with a decent voltage I would see at least 550 MHz stable.

At the moment I can even get close to the performance of a 500 MHz K6-2+

here a video of my tests of this weekend

https://youtu.be/ozpZ1SLEroc

Have you tried it at all at 550 with the +5/10% jimper setting? It may be enough? Alternatively one may be able to modify some resistors before those jumpers to alter voltage.

Also be sure your ram is up to the task, don't have a stick of PC66 in there, preferably some 133 rated stuff. Also try and lower all memory settings to slowest to begin with.

Yes of course all tests was made with +10% VCOREsettet by jumper in the mobo. Also the VIO are working currently at +10%. RAM i am using is A PC133 module, still good at 166 MHz (tried many time for other oc's projects) and is still stable with CL 2 and timing 2-2-2-6.

Celerons are ~30W, maaaybe up to 50W with OC to 550MHz. Watercooling is an unnecessary burden/risk.

I find more easy to build this watercooler in a few hours with some parts scattered around, that find a solution with somelse different air thing to adapt.Slot 1 have the problem of the RAM in front of the CPU, and the little space area around. Unnecessary? Maybe. Useful for sure and cool also. But at this point, the issue is not how to cool down the CPU. If could find a way to put 2,3 V at least or more (as I had originally planned) a good cooler is for sure necessary. And don't forget this hw is more than 20 years old! Traces inside could be just worn from previous experiments by some other tester. A pretty low temps just help to avoid risks of breaking traces.

Your Iwill board is a Via Apollo Pro 133A wich puts it firmly in Pentium Coppermine territory.
The VRM for that generation is designed along the 8.4 specification which covers the voltage range of 1.30v - 2.05v.
According to the Intel specification sheet the voltage range 1.8v-3.5v that supports the earlier generation of cpu's is just 'optional' -see page 9 of attached sheet.
If that voltage range is disabled on the voltage chip on your board then you may be out of luck.

Retroinside wrote on 2024-10-14, 10:15:

Hi guys! […]
Show full quote

Hi guys!

I was working in this days on a Slot 1 Platform (Iwill VD133Gold board), tryng to perform a decent overclock for a Celeron Mendocino 333 MHz.
Had a 90 radiator and an universal chipset water block aside. I easily recovered a universal pump from China and created a small, but efficient liquid cooling system for my Celeron Slot 1.
I'll be honest, I was hoping to get more. Unfortunately, however, the motherboard does not allow me to increase the CPU voltage as I would like. With some Jumpers , it allows me to increase the vcore by 10% or 5%. Same thing for the VIO too.
However I was hoping to be able to push it to something beyond 2.3 (I know that up to 2.5V is manageable, if properly cooled).

I attempted pin mod of pins a119 and b119. Without any success alas. The motherboard just won't POST with this change. I also tried covering the pins individually but again, no success.

I therefore ask for advice from any of you, if you have ever tried this practice, and can provide me another solutions.

At the moment I am stable with a poor vcore of 2.2 V, at 525 MHz (105 MHz FSB X5.0) but I can pass POST even at 550 (110 MHz X5.0) but it won't load OS.

I am therefore sure that already with a decent voltage I would see at least 550 MHz stable.

At the moment I can even get close to the performance of a 500 MHz K6-2+

here a video of my tests of this weekend

https://youtu.be/ozpZ1SLEroc

Slot 1 watercooling, i approve!
Tho that Celly might need a coat not to catch a cold =)

I managed to push a S370 300A (also Mendocino) to 600MHz using a normal cheapo S370 HSF.
Can't remember what MB I used but probably a BH6 or a BX6-2.

bofh.fromhell this is awesome!!! Thanks for sharing this !

You know, these days I was just thinking about the PGA370 variant, which thanks to a slotket adapter would allow me to set the voltages directly from the PCB and I could literally make the little Mendocino fly. The fact that I have to reinvent a way to fix the water block to the socket is braking me a bit, as it is much more complicated than the good Slot 1. If I can do something decent, I'll use my dear MSI 6905 card and see what I get from it.