Hello Alain

> I've simply disconnected one wire (pin 1) on
> connector 18 of KLM-355C (filter board) that was bringing the
> keyboard CV coming from KLM-356D to the filter.

Attention:
The pinout has inverted between old and new version of KLM355 (You must
read this with a monospace font like courier):

PinOLD PinNew
KEYBD-TRACK 1 3
SINGLE-TRIG 2 2
MULTI-TRIG 3 1

> So the glide times are identical to that of VCO 1.

This will be useful only in Unisono and Chord-Memory modes!!! In
Unisono/Share and Poly modes the VCF should follow the CV of the highest
VCO (that is what pin20-1 is delivering) - which is not necessarily
VCO1. In the opposite: VCO1 is typically the lowest note (in Unisono/Share).

> Korg should have added a portamento feature to the filter cutoff
> in the first place (as on many Moog synths).

I think Korg did it because of the mentioned dependency. And also they
used a fourfolded potentiometer, which is already difficult to get, but
a fivefolded seems close to impossible

The better solution would have been (or would be) to establishe five
voltage controlled portamentos, where the fifth portamento affects the
KBD voltage of pin20-1/3

Florian