shamino wrote on 2023-07-27, 18:39:

In the picture it looked like there are 5 jumpers. If there are only 4, then it means the high order bit is soldered, so then the minimum will be 2.0V. But if all 5 jumpers are present then it can go below 2.0V. On my board (different brand/model), early revisions only had 4 jumpers but later revisions added the 5th.

Yeah, on second look, the motherboard manual image also matches the pattern of the table you posted. And that 5th "Reserved" jumper is probably the way to go below 2V, again matching the table.

It even says in the table: All jumpers removed = shutdown. So it suggests that the lowest working setting without touching the reserved jumper is 2,1V.

( I have a similar thing with the Chaintech 5RSA0 and 5RSA2 boards. Their manual suggest the minimum is 1,7V, but the VRM datasheet explains the pattern to go as low as 1,3V. Verified with a multimeter and now using 1,5V for my K6-III+.)

shamino wrote on 2023-07-27, 18:39:

The regulator is made to be compatible with the 5-bit VID setting that is provided by later Intel P2/P3 CPUs. It covers the ran […]
Show full quote

gerwin wrote on 2023-07-27, 17:07:

shamino wrote on 2023-07-27, 16:10:

The P5M4-M+ is posted on theretroweb:
https://theretroweb.com/motherboards/s/tekram-p5m4-m-2

I looked at the pdf manual there earlier.
The voltage jumpering is a bit primitive. It seem like it starts with 2,0V as a baseline. Then the four jumpers allow for adding 0,1V / 0,2V / 0,4V / 0,8V. Or combinations of these four values. Adding all four makes 2 + 0,1 + 0,2 + 0,4 + 0,8 = 3,5V.

The regulator is made to be compatible with the 5-bit VID setting that is provided by later Intel P2/P3 CPUs. It covers the range from 1.3-3.5V. The jumpers just make that setting manual, since socket-7 CPUs don't have pins for automatic VID.

The way Intel defined the VID values, there is an odd shift in the scheme that happens at 2.0V or 2.05V, so the pattern changes when going above or below that value. There's actually some overlap - there's 2 ways to declare 2.0V but some regulators only allow 1 of them (including this LM2637 - that's why it went to no output with all jumpers removed).
On my super-7, I use a red jumper cap for the high order VID jumper. It serves as a reminder to not move it. This ensures the voltage can't go any higher than 2.05V.

In the picture it looked like there are 5 jumpers. If there are only 4, then it means the high order bit is soldered, so then the minimum will be 2.0V. But if all 5 jumpers are present then it can go below 2.0V.
On my board (different brand/model), early revisions only had 4 jumpers but later revisions added the 5th.

Very nice explanation,

If memory serves, the way some of these work is as follows

The one you call high order basically cuts all values in half.

Goes from 2.0 base voltage (not working) to 1.2 or 1.3v but that isn’t all

The jumpers that were once +0.1, + 0.2, + 0.4, +0.8v become +.05, +0.1, + 0.2, + 0.4
Giving finer precision in the lower range.

So 1.3v plus +.05, 0.1, 0.2, 0.4 equals 2.05 if you remove the .05 you get 2.0.

As you said, two ways to reach 2.0. 😀

Florian J wrote on 2023-07-26, 20:38:

Actually the P5TX-Bpro just overvolts by 0.1 volt, while the Tekram P5M4-M+ overvolts it by 0.2 since it can't go lower than 2.2 volt and the AMD K6-2+ is rated for 2.0 volts.

What is the voltage value on P5TX-Bpro you were referring to, 2.1 volt? What is the jumper setting for this, please?
I just bought K6-2 that needs 2.1 V Core, and I have P5TX-Bpro v2.0.