[sdiy] LM13600 Gain Question

[Tim Stinchcombe]

> I'm looking at the LM13600 data sheet and fig. 4 (stereo volume 
> control) has a formula to calculate the gain given Iabc as:
> 
> Vout
> ---- = 940 X Iabc
> Vin
> 
> Why 940?

I too had been troubled by this expression a few years ago: I have more
resources at my disposal now, and (hopefully!) a better understanding of
such things, so I decided to have another look at it.

Running a few simulations was enough to convince me that their main
expression Iout=Is*2*Iabc/Id is out by a factor of about 2. The culprit
appears to be the assumption that the voltage splits in the diodes as
Id/2+/-Is: in fact Id/2+/-Is/2 would be much closer to reality (making
Iout=Is*Iabc/Id). I then wrangled a few expressions and ended up with a
nasty non-linear one relating Is to the +/-'delta' current flowing in the
diodes. Using Mathematica to solve the expression gave me very similar
figures to the simulation, i.e. that the 'delta' is about half Is, which
acted as independent confirmation I'm on the right track. Conclusion: the
940 was either obtained a) empirically, b) by simulation, or c) by some
heavy maths using goodness-knows-what approximations.

As already noted, using it as a starting point and experimenting is going to
be the order of the day. Incidentally, the 0.94 does seem to arise from the
component values in Fig 2 (with Vs at 15V); using Fig 4 values, the
simulation gives the factor around 1.05 or so (Iabc in amps).

Tim

__________________________________________________________
Tim Stinchcombe 

Cheltenham, Glos, UK
email: tim102 at tstinchcombe.freeserve.co.uk