VCF matched transistor pair workings

[VA1]

Hi, i build many transistor/diode ladder filters,
each time it will cost me 4 matched NPN transistor pairs :

For the exponential converter :
This is important for the keyfollow so the notes wont sound false.

Then the input/reso input pair :
Why is this important ?

Then the top of the ladder :
Why is this important ?

Then the differential amplifier pair :
Why is this important ?

Are these pairs to level the reso to exact selfresonating for example ?,
or how does it change sound wise ?

I measure these pairs, it cost a lot of time, so i like to eliminate the not so important ones that dont change the sound, only the volume.

Can you really hear a difference ?
thanks

[AlanP]

I heard somewhere that matching the capacitors in the ladder is more important.

I socketed all the transistors in my first couple of Minimoog ladder VCFs. First go around, I just put random transistors in every place, out of sheer laziness.

Then I spent a couple hours measuring a bunch of transistors, and replaced the lot with matched pairs.

Couldn't hear that much of a difference at all.

[Grumble]

I think that is because the fabrication process is so refined and controlled that transistors from the same batch barely have any differences, especially when compared with transistors fabricated in the 70’s

[ixtern]

There is a document: "Analysis of the Moog Transistor Ladder and Derivative Filters" by Timothy E. Stinchcombe on:
http://www.timstinchcombe.co.uk/synth/Moog_ladder_tf.pdf
but is packed with math. Some decades ago I could do such analysis but not today

[VA1]

Its a good document, i seen it before.
No farmer knowledge, i just look at the pictures.

When measuring transistors you could find the oddest pair to test in your sockets.
Or taking 2 different transistors.