Transistor Arrays - whats going on!
magnus at analogue.org
Wed Nov 19 01:29:59 CET 1997
>>>>> "RT" == Roy Tate <roytate at ionet.net> writes:
RT> I just talked to Mouser, and they told me the CA3046, CA3086
RT> and CA3096 are "non-catalog items." They offered to sell me
RT> 525 of the CA3046s (minimum order quantity), but I declined. :)
RT> This _apparently_ reflect Mouser's desire to stock only higher
RT> volume items, as I couldn't find anything at www.semi.harris.com
RT> mentioning the death of the CA3046.
Obviously isn't Mouser's interested in having you as a customer for
RT> So now would be a good time to figure out a scheme for matching
RT> transistors. I know Sean Costello has Moog's method on his web
RT> page (see my links.), but several members of Synth-DIY felt this
RT> matching was sub-optimal. What tests would be best?
RT> How about if I ...
RT> * Use a precision current source
RT> * Measure voltage at two current levels
RT> ... And independantly select for low noise in a noise circuit.
RT> (That last is just a wild idea ...)
That would be a good start...
A good matching would concentrate on some key parameters which is
important to a specific application. If it is antilog amplifiers you
want some parameters to be close, but are less importance in other cases.
One thing that you do miss with not having an array is that you loose
the relatively tigth termical connection between the transistors, this
can be more catastrofic than some other parameters.
More information about the Synth-diy