Wire the FSB pin headers to SPDT switches mounted wherever you want. (i.e. a spare 3.5 or 5.25 bay face plate) I'm not really sure if you should or could switch them on the fly though -- they are probably make before break switches and it would likely lock-up the CPU at a minimum.

Yeah I wouldn't expect it to work on the fly, I'm okay with power off and restart. I think when I did it back in the 90s it just ignored it when I flipped it while on, and would only take effect after a restart. I'm sure it's motherboard specific behavior though.

So is there no way to split the turbo switch wires and plug them into multiple spots on the board? I'd rather have 1 switch control it all if possible. If it's not possible then I'll probably just abandon the idea.

Not with one SPST switch (1 in, 1 out). You have multiple FSB pin headers which would require 1 input (center pin) and 2 outputs (top & bottom pins) to switch between.

I was thinking of something like this, with a wire splitter on each line of the turbo switch, would this not work? Sorry for the horrendous drawing, but hopefully you get the idea:

You can't independently control two 3p jumpers with one switch like that. If you have a 3p turbo switch, you have one input and two outputs. (High, Common, Low) All you're doing with that switch is going between HC and CL.

Not sure how the settings are on your mobo, but on my Intel TX board, I've tested with 1 single rocker switch, while only changing the CPU speed, and keeping the FSB static. Which makes me able to switch between (only) 100mzh and 233mhz on the same CPU. Granted, a bit limited vs the speeds inbetween, but.... Depends on what the intended purpose would be of course... Using a different CPU, say a 200mhz, the range may differ (getting a lower minimum perhaps).

FYI, old style AT case Turbo switches tend to be DPDT switches, one to control the motherboard, and one to control the LED display. so you might be able to get away with it. though I doubt the cpu/motherboard/software will like it.

bestemor wrote on 2021-02-19, 22:38:

Not sure how the settings are on your mobo, but on my Intel TX board, I've tested with 1 single rocker switch, while only changing the CPU speed, and keeping the FSB static. Which makes me able to switch between (only) 100mzh and 233mhz on the same CPU. Granted, a bit limited vs the speeds inbetween, but.... Depends on what the intended purpose would be of course... Using a different CPU, say a 200mhz, the range may differ (getting a lower minimum perhaps).

I'm running a K6 3+, so I can change the multiplier with software already. I wanted the ability to change the FSB as well without opening the case and messing with jumpers.

I thought about this a good while ago as I *had* a build around an Asus P5A-B and a K6-3+ 450. I wanted to be able to change the FSB using the turbo but could not come up with a way to do it. In the end I picked up a Iwill XA100 Plus and this as FSB change in the BIOS.

The case which had the original P5A-B build in now houses a 233mmx system on a Gigabyte HX board which allows turbo to 50mhz by using the button. I'm going to order a 4 way DIP switch so I can change the FSB (60-66) and Multipliers and mount it somewhere on the case so I can change the Multi and FSB when needed without opening the case.

I've used an SPDT turbo switch exactly for this, although it's currently being used to toggle the multiplier. The principle is to toggle one of the jumpers between 1-2 and 2-3. Usually this has to be locked in via a reboot as it's programming a clock generator chip for FSB and setting the internal CPU multiplier and the jumpers are generally only probed at that time.

Yes that is exactly what I'm attempting, but I need to switch 2 jumpers instead of just one.

You say you need to switch two jumpers, yet settings for 66 and 100 differ only in one jumper setting. So I'm assuming you're running 112 MHz FSB? In that case, if you can't find a solution for two jumpers, you could still at least make it switch to 75 MHz FSB, since that only differs from 112 in one jumper setting.

weedeewee wrote on 2021-02-19, 23:44:

FYI, old style AT case Turbo switches tend to be DPDT switches, one to control the motherboard, and one to control the LED display. so you might be able to get away with it. though I doubt the cpu/motherboard/software will like it.

Just adding this image & hope the possible logic sinks in.

frudi wrote on 2021-02-26, 11:06:

You say you need to switch two jumpers, yet settings for 66 and 100 differ only in one jumper setting. So I'm assuming you're running 112 MHz FSB? In that case, if you can't find a solution for two jumpers, you could still at least make it switch to 75 MHz FSB, since that only differs from 112 in one jumper setting.

You are correct that only 1 jumper needs to be changed for the FSB to go from 66->100. The problem is, the computer will not post unless I adjust the sdram jumpers as well. The sdram jumpers must match the fsb speed for me to get it to post at either 66 or 100 (and they have different jumper settings for this).

From the "research" I've done (http://softlyspokenmagicspells.com/overclocking.html, check the bottom of the page) it appears like it's altering the PCI bus speed at either FSB/2 or FSB/3 with these additional settings. For whatever reason the computer will not post if I'm running 66 Mhz FSB with PCI bus = FSB/3, which would make the PCI bus run at 22 Mhz instead of 33. I guess the system doesn't like that. Maybe I just need to try some different RAM? The current RAM i'm using is PC-133 SDRAM (2 sticks x 256=512MB), which I figure would be able to handle any settings this board had.

Also, I can't seem to even get the board to run stable at 112Mhz with any possible settings. No idea why, maybe I need to disable L2 cache?

I figured I would need to replace the turbo button with a 3PDT switch to achieve this, and I really don't want to spend the time and money with a drop-in replacement, so I came up with this solution with a DPDT switch I already had.