Since I have a weakness for Intel 430HX mainboards I recently acquired a Micronics D5CUB.
It was in a bad condition with missing capacitors, dirt all over it, missing (= torn out) 2nd voltage regulator and of course not working.
So I cleaned everything carefully, flashed the latest BIOS (Oct. 1996), added the missing jumpers, some capacitors and the voltage regulator. While at it I also replaced the heatsink of the core-voltage regulator by a slightly larger one.
Now everything looked quite fine:

This mainboard was available with 256 or 512KB L2 cache and it´s obvious the 512 KB version was absolutely pointless due to the missing socket for a second TAG RAM.
O.K. then 64MB cacheable area it is.

Needless to say the mainboard was dead - no signs of life except for the chips getting warm. A POST-card showed all voltages and clock signal present but no code at all.
A closer inspection revealed 8 pins of the southbridge at one corner had lost contact. After resoldering the board finally POSTed.

Since this is a split voltage model, I tried with a Pentium MMX (and K6-2 later) that gets identified as "Pentium-S" due to the 1996 BIOS but with the correct frequency (233 MHz).
I then desperately tried to set the IO-voltage and core-voltage. There´s a jumper block for each voltage but they interfered with each other. When core voltage was set correctly (2.8V) the whole 3.3V rail of the mainboard (for chache chips, super-IO etc.) had 2.8V. The board wouldn´t POST sometimes and I couldn´t boot into DOS at this far too low voltage. Having IO set to 3.3V made the CPU overheat.
First I suspected Micronics had mixed up the jumper positions in the manual, but flipping the jumper positions didn´t help either.
As we know, mainboards of this era - between single voltage and dual voltage CPUs - often have some jumpers to bridge or seperate the two voltage rails. So does the D5CUB.
The two jumpers JP10 were in position 2-3 which is absolutely correct.
Measuring the resistance between the 2 regulator outputs prooved that depending on the jumpers the voltages are either bridged or seperated.
BUT as soon as a CPU was inserted the voltages were bridged!
This explained everything, but what´s the reason?
I then checked the connectivity of all the Vcore and Vio pins at the cpu socket and...snare roll... one connection was wrong.

Pin A17 that is supposed to be Vcc2 (core-voltage) is connected to Vcc3 (IO voltage). Since inside the CPU all the pins for Vcc2 and Vcc3 are bridged (at least in this column) inserting a dual voltage CPU will instantly short IO-voltage and core-voltage!

The connection is made in some in between PCB-layer so cutting a trace was not an option. I opened the CPU socket instead, desoldered the pin and pulled the whole contact out.

After this fix the voltages could finally be set correctly.
It´s still annoying that the Vcore heatsink that dissipates most of the heat has been put in the position at the corner of the mainboard where the airstream of the CPU cooler won´t reach it.