UPDATE: Iv braved the jumpers and managed to get it to boot at 160mhz:

It won't make it into DOS however, simply freezing after "Starting MS-DOS...". Disabling L1 cache fixes this however?
It freezes regardless of memory timing speeds. It seems only L1 cache being enabled lets it boot fully.

So i guess this chip is a dud? It couldn't do 150mhz either mind you. Didn't even make it to POST.

Use DX2 66 3.3V Jumper setting for voltage, and frequency. Finally set MHz frequency to 40 MHz for the frequency generator jumpers. Hence 160Mhz.

Cache chips needs to be -15. Which is 15ns.

Cheers,

pentiumspeed wrote on 2020-05-24, 03:31:

Use DX2 66 3.3V Jumper setting for voltage, and frequency. Finally set MHz frequency to 40 MHz for the frequency generator jumpers. Hence 160Mhz.

Cache chips needs to be -15. Which is 15ns.

Cheers,

I'll have to give this a try later, what i ended up doing was jumping it up as a 586-133 and only changing the "CPU Clock Selector" jumper to 40mhz.

Im not sure if thats why it couldn't boot into DOS unless internal cache was disabled? It would POST and list properly as a 586-133 @ 160.
Cache chips are all -15. Its the internal cpu cache i need to disable inorder for DOS to fully boot.

There's a reason why it is a -133 and not a -160. If the L1 cache can't run at that speed, there's nothing you can do. Unlike with L2, there's no way to change the timings of the L1 cache. Your CPU does not work at 160 MHz. It's as simple as that. You can try a different one, they are quite easy to come by.

Not so fast... what voltage is the CPU running at? What other options could you select? By the sound of it, it's only marginally beyond this CPU to run at 160MHz at the voltage it's currently running on. 5x86 CPUs are specced for 3.45V +-0.15V. If you're running it at 3.3V it's in-spec, but at the very bottom of the range. Check for a 3.45V setting, if it's there those 0.15V might just make the difference to get it running at 160MHz.

If you're at 3.45V already, there's nothing to be gained there. If you have good cooling you could try 4V if you have that option, although it is risky.

Also, failing higher voltage for 160MHz, try setting the CPU to run at 3x50MHz. This would actually be faster than 4x40 in many things (RAM & L2 cache run 25% faster). A good plan B.

Can you post a photo of your motherboard jumper settings for 160mhz ?

Hello!
I'm also running two Am5x86-133s at 160. They were not stable with the 3.3 V setting on the board. I never actually tried disabling L1 though (which will make it slower than it is at 133).

What I did have to do is increase the voltage. Technically the Am5x86 datasheet I'm reading says 3.3V +/- 0.3V, so I'd be fine going up to 3.6.

Of course, to change the voltage you'll need to tweak the voltage regulator on your board. What I've done on both boards I've modified for this, is to find the datasheet for the voltage regulator and study your board and figure out what resistor values it is using for the voltage regulator. Then one of the resistors needs to have its resistance decreased by soldering another resistor in parallel with it (easiest to do on the underside of the board). The voltage regulator datasheet should give you an equation to determine what voltage it generates, and you know that lowering the denominator will increase the result. From there, if you add a resistor R' in parallel with existing resistor R the new resistance is 1/(1/R+1/R'). Of course, use a multimeter to check what voltage is actually being generated before putting a cpu back in. I have one board running at about 3.58V and another at about 3.8 V.

The more sophisticated way is to remove the correct resistor and replace it with a variable one (trimmer) so that you can set any voltage you want by turning a screw. But my soldering skills aren't up to that.

Make sure you have a good heatsink and fan, too.

Oh also, I'll add to double check your jumpers and make sure you didn't mess up anything with regard to the Am5x86-133 being a write-back SL-enhanced CPU.

jakethompson1 wrote on 2020-11-21, 04:03:

What I did have to do is increase the voltage. Technically the Am5x86 datasheet I'm reading says 3.3V +/- 0.3V, so I'd be fine going up to 3.6.

Of course, to change the voltage you'll need to tweak the voltage regulator on your board.

I was about to have a crack at this, but then, googling the part number on what I thought was the voltage regulator (the only large TO-220 package on the board), it turns out it is a transistor - it's a TIP31C. There is a fairly large electroyltic capacitor next to it as well. This is all just next to the 3.3/5v jumper.

Does that mean the board uses some type of switch mode power supply arrangement to derive the 3.3v (actually 3.4v according to my meter) or have I got the wrong end of the stick?