VOGONS

Socket 7 Tillamook 300 (4.5x)

Topic actions

First post, by Sphere478

Posted on 
User metadata
Rank l33t++
Rank
l33t++

At long last,

It looks like I have finally found a (working/usable) tillamook 4.5x for my socket 7 collection. And it comes in a package I had previously never heard of. (see also Pentium 1 366mhz The Infamous Golden Tiger/Warrior)

The following is a information and image dump of what I know so far. Info/pics Provided by Batyra.

And this is what ChatGPT found after deep search on archival japan forums. So the same as I did.

HP5-MX300-L – Multiplier Switch (SW1) and FSB Settings
CPU Multiplier Configuration via SW1 (DIP switch)
The Buffalo HP5-MX300-L is a CPU upgrade module featuring a mobile Intel Pentium MMX 300 MHz processor, designed for Socket 5/7 motherboards. On the PCB, there's a set of four DIP switches (labeled SW1) used to configure how the CPU runs .

The first three switches (SW1-1, SW1-2, SW1-3) control the CPU multiplier, adjusting how the CPU clock relates to the base FSB frequency . The factory-default configuration is OFF, ON, OFF, OFF (in order) —this sets the module to the highest supported multiplier.

By changing SW1-1 to SW1-3, you can select lower multipliers (useful for alternate CPUs or lower FSB boards). The available multipliers range from around 1.5× up to a maximum of 4.5× .

The fourth switch (SW1-4) does not affect the multiplier—it likely has a different function (not clearly documented in sources) and is usually left in its default position .

Multiplier Limits and FSB Dependency
SW1 settings allow for a maximum multiplier of 4.5×, which results in ~300 MHz when used with a standard 66 MHz FSB . Higher multipliers (e.g., 5×) are not supported—Japanese sources confirmed no DIP combination results in more than 4.5× .

This limitation affects users installing the module on boards with lower FSB. For example, in a PC-98 system with a 60 MHz bus, the max achievable clock would be 270 MHz (60 × 4.5) .

The multiplier is set entirely by hardware (via the SW1 DIP switch) and doesn't rely on BIOS support. This enables the module to function even in systems that don't officially support MMX CPUs.

NOTE: Some systems may require additional software or settings to support the MMX write-back cache mode, but that is unrelated to the multiplier itself.
Last edited by Sphere478 on 2025-07-22, 03:44. Edited 2 times in total.

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 2 of 6, by RayeR

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

And what's the best O'C result you can get with it? Does it better that slower-rated models with 4x multi?

Gigabyte GA-P67-DS3-B3, Core i7-2600K @4,5GHz, 8GB DDR3, 128GB SSD, GTX970(GF7900GT), SB Audigy + YMF724F + DreamBlaster combo + LPC2ISA

Reply 3 of 6, by RetroPCCupboard

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

Interesting package format. So, in theory, it could do up to 450Mhz on a super socket 7 motherboard. Thats assuming the chip can take that speed. I have a Pentium MMX 233Mhz that runs stable at 3 x 100Mhz at 2.9v. So, this CPU, which I think is a smaller fabrication process, should be able to exceed 300Mhz easily. Especially if you increase the voltage.

Reply 4 of 6, by RetroPCCupboard

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

BTW, I think I have read somewhere that its not unusual for Tillamook CPUs labelled as 300Mhz to actually be overclocked 266Mhz models.

Reply 5 of 6, by old school gamer man

Posted on 
User metadata
Rank Member
Rank
Member
RetroPCCupboard wrote on 2025-07-22, 17:26:

Interesting package format. So, in theory, it could do up to 450Mhz on a super socket 7 motherboard. e.

That would be one fast mmx

Reply 6 of 6, by Sphere478

Posted on 
User metadata
Rank l33t++
Rank
l33t++
RayeR wrote on 2025-07-22, 12:57:

And what's the best O'C result you can get with it? Does it better that slower-rated models with 4x multi?

Not sure yet. I’m going to find out though 😀

It does kind of seem that the plastic pin grid array chips always seem to have an advantage though doesn’t it.

Interestingly 🤣 I don’t think I’ve ever been able to max out any of my PPGA chips. The board I always seem to be using is a 66, 100, 105 or 112 mhz and at 4x the chip just runs 🤣. Doesn’t seem to matter if it’s a 166, 233, or 266 ppga tillamook.

I do happen to have recently set up a mvp4 test bed though…. Heheheh

RetroPCCupboard wrote on 2025-07-22, 17:26:

Interesting package format. So, in theory, it could do up to 450Mhz on a super socket 7 motherboard. Thats assuming the chip can take that speed. I have a Pentium MMX 233Mhz that runs stable at 3 x 100Mhz at 2.9v. So, this CPU, which I think is a smaller fabrication process, should be able to exceed 300Mhz easily. Especially if you increase the voltage.

Yeah, I had never heard of them or seen them before I found this one. (Don’t have it yet, it’s on its way)

The ppga ones are just rock solid at 400mhz every time I try one. But maybe I have good luck with them.

I have had really bad luck with the film chip ones though (what was that package called again?)

9times out of 10 by the time I get them someone has already damaged the film legs and they are so tiny, I can’t fix them. :,(

I have a whole colledtion of golden warriors, i was testing them when I discovered this problem they were made poorly in a way that the heat spreaders fall off and the previous owners ended up touching the gold wires. Once I found that out I haulted all further testing for fear of damaging a board or CPU until I find a way to verify and repair the gold wires. Best to preserve them, rather than risk them, or other equipment. But unfortunately that means they aren’t much fun.

Most of the time I run the tillamooks at 2.0 or 2.1v because that’s the lowest the board can do. (BTW)

Bonus tidbit
In intel documents there is mention of a 60 mhz 5x model. But to my knowledge, none have ever been documented to exist or still exist. And also, no one that I know of has yet been able to find a sample that allowed 5x to be set when they fiddled with jumpers.

Maybe this one will have it?

It would be nice to attract a tillamook engineer to one of these threads and get some inside deets.

BTW, RayeR, I am very interested in getting a CPU from you still if you ever decide you are willing to make me one. 😀 I even have one on the MFG holder I can send you. (266 film chip)

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Go to top of page Go to top of page
Back to General Old Hardware
Close window
Close window