[sdiy] discrete SSM2018?

[jh.]

> Which is to say that I also see no reason why the gm of an OTA can't be
> spec'd somewhat accurately.

And what I don't understand at all, is that the gm *tracking* between
two OTAS could be so good nevertheless.

>From the LM13700 data sheet:

gm (@ a certain iAbc current): min: 6.7mS, typ: 9.6mS, max: 13mS

BUT:

Tracking of gm between two OTAs of one chip: typ: 0.3dB

So we have a variation of 2:1 of absolute gm, but a typical gm
tracking of only 3.5%.

All right, absolute gm is spec'ed min / max, while gm tracking
is only "typical". Strictly speaking, the tracking figure is worthless,
because it's not guaranteed.
If we want to make something useful of the data sheet nevertheless,
I can see two possibilities:

(1) gm tracking is typically better than absolute gm (the great
      mystery which started this branch of the current thread),

or

(2) absulute gm for a fixed temperature also varies typically
     around +/-3.5% (just as the typical tracking error), but
     the guaranteed worst case is much worse (2:1).

Which one is more likely ?

JH.