In article d3fUf.1230$K11.421@clgrps12, "Dave"
wrote:

"G" wrote in message
...
Dave you wouldn't put the diodes from the 7915 output to ground but from
the output to your load in series......being a negative polarity output
you'd put the cathode towards the 7915 (and any additional diodes as
well).

The trick to bumping up the voltage of a 3 terminal fixed regulator is
to insert a diode between the ground pin and ground (keep in mind to
insulate the tab if you do this!). The regulated voltage will now be
pushed up by the junction voltage of the diode or diodes (Diodes placed
as such in a 7912 would have their cathode face the ground pin.....anode
to ground).


Makes sense. One question: why do you say insulate the tab? If the tab is
already connected to ground, and you are just adding a diode, wouldn't you
only have the 1V of potential difference between ground and the pin with a
diode? Would I want to use a mica insulator between the tab and heatsink?

Dave-

I agree with G for the most part. I prefer the 7912 approach rather then
using the 7915. As someone else suggested, a capacitor from the output to
ground would be a good idea. I think you mentioned having one in the
original circuit.

Another alternative to the 7912 approach would be to use the negative
version of an LM317 with pair of resistors instead of the one diode. One
resistor would be connected from the regulator's output to its common
("adj") terminal. The other would be connected from the common terminal
to ground. Values would be chosen to produce the required output
voltage. The LM317 regulator actually has about 1.2 volts output plus the
voltage dropped by the resistor to ground. I assume there is a negative
version of the LM317, if you wanted to look further.

Of course an exact replacement transistor is the most sure-fire solution.
If you were to use an alternative part, I suggest using one rated for at
least the minimum hfe of the original part, or at least one having an
actual (measured) hfe of that value or higher.

Fred