19 Jan
2010
19 Jan
'10
2:48 p.m.
Sridhar Ayengar wrote:
Chuck Guzis wrote:
But I thought part of the problem was that the regulator with the
slightly higher output voltage would drive the voltage difference
against the output of the lower-voltage regulator? Wouldn't isolation
diodes prevent that from happening?
Diodes will effectively leave you with only one regulator supplying the
current, as only one diode will switch on.
Using Chuck's numbers as an example, the one reg supplies 5.05V into one diode
anode, the output at the cathode is 5.05-0.7=4.35V. The diode from the other
reg (cathodes connected together) now has 4.95-4.35=0.6V dropped across it,
which isn't enough to even send it into conduction, that reg supplies nothing
to the load (..until the first reg over-temps and burns up or shuts down).
If one looks at high-current regulators they have multiple pass transistors
connected 'almost' in parallel, the catch is a small resistor is placed in the
emitter lead of each so the CE & BE circuits take on some linear
characteristics and distributes the current between the transistors. The
regulation sense point is then placed after all the resistors join together for
the output, so the resistors are inside the feedback loop and hence accounted
for in the regulation.
Couldn't you get around at least part of the problem with isolation
diodes?
No sir. Consider that the output of one 7805 might be 5.05 volts and
of the other, 4.95 (manufacturing tolerances being what they are,
that's not unreasonable). The bulk of the current will be conducted
by the 7805 with the higher voltage output. One might put a resistor
in series with the output of each one to balance the load a bit, but
that would play hob with the voltage regulation. (On the other hand,
this technique is used when paralleling rectifiers to balance the
current.)