matching diodes

[Christophe Stoelinga]

> I guess the current varries in the different Ohm ranges, indeed diode is
> the same as 2kOhm range in most meters.  Measuring in one point is good,
> measuring in two points is maybe better, ie. use another range.
> It is possible that two diodes have similar voltages for one current,
> but have considerable error at some other current, e.g. if a higher knee
> voltage is combined with lower parasitic resistance.

I once tried to find a couple of diodes this way (checked one point). 
I found 4 that perfectly mached ( as good as my DMM could measure) then I
heated one. Of course voltage dropped. But it never returned to the original
voltage. It were old germanium types in glass housings. this glass may have
insulated the heat, wich then stayed longer than I cared to wait. 
What i mean to say is I don't think you will ever match diodes (or
transistors) this way very good. 
I would prefer diodes that are placed on one piece of Si. (perfectly mached
and on the same temperature).
Do they exist? I know of diode arrays to protect digital circuits but
suspect them of just being multiple diodes in one (sil) housing.
otherwise I would use transistors as a diode. 
Another question about ring-mods and carrier bleed trough:
would it be possible to reduce (eliminate?) carrier bleed trough by changing
the DC offset of the carrier? if the ringmod is loaded with a not too high
impedance DC offset would mean the current trough the diodes varies with the
carrier amplitude. DC offset would mean vari the current trough one pair of
the diodes. this could compensate for diode voltage mismatch, or am I wrong?
maybe I should buy two transformers and try it. 


Christophe.