[R-390] OT: Pass Transistor Question

[2002tii]

Jon wrote:

>I purchased 10 new transistors and after sorting ended up with a 
>measured Beta (current gain) RANGE of about 2:1.  I now have two 
>well matched sets installed and two lesser well matched sets as 
>spares.  The two unused outliers  have the 2:1 Beta ratio

Beta does not really matter much for emitter follower applications 
like the NPN pass transistors on a positive voltage regulator or the 
"top" output transistors in the Crown amp.  It has some limited 
relevance to the "bottom" transistors in the old Crown 
quasi-complementary output stage.  By far the greatest worry as far 
as current-sharing is concerned is the threshold voltage of the 
parallel transistors (for power transistors, this is usually measured 
as the base-emitter voltage required to produce a specified small 
collector current, generally in the 1-10 mA range).  [Note that this 
parameter is rarely specified on data sheets.]  The problem is that 
the increase in base-emitter voltage required to increase the 
collector current from 1 mA to 1 A (a factor of 1000:1) is very small 
-- on the order of the normal variability of the threshold voltage 
from one transistor to the next.  So, one transistor tends to hog all 
the collector current before the other one even turns on.  Then, as I 
mentioned, even if they are reasonably well matched and the other one 
carries some current, the one that is carrying more current gets 
warmer, which lowers its threshold voltage, which makes it carry an 
even greater portion of the current, which makes it even warmer, etc., etc.

Best regards,

Don