Doubt on Power Supply

Hi,
I have 4 loads at different points. 24V Supply is given to them from 4 different 7824 regulator ICs. I want all of them to be driven by same FET as shown in the diagram.
I think I have to connect all the 24Vs together. Anything wrong in connecting like that?

No, keep it like that. No 7824 is exactly alike so connecting them in parallel would cause the one with the highest voltage to try and supply everything while the other ones do nothing. As long as they are sharing the ground, it's all cool.

It's like connecting a bunch of batteries of varying voltages in parallel. The higher voltage batteries most cancel out the voltage of the lower voltage batteries, creating the equivelant of a tiny voltage source that is short-circuited producing lots of current. THe result is that a LOT of current flows from the highest voltage batteries into the lower voltage ones trying to charge them to make all their voltages equal, and leaving only the highest battery to power the load.

It's also inversely similar to the case where you connect a bunch of diodes in parallel. THey will all have slightly different voltage drops, and the one with the lowest voltage drop will conduct everything (pretty much) while the rest will not turn on because their required voltage drops are higher but the voltage drop across them is that of the diode with lowest voltage drop that is conducting since they are all in parallel.

You need to combine all the power supply grounds together.

dknguyen said:

No, keep it like that. No 7824 is exactly alike so connecting them in parallel would cause the one with the highest voltage to try and supply everything while the other ones do nothing. As long as they are sharing the ground, it's all cool.

It's like connecting a bunch of batteries of varying voltages in parallel. The higher voltage batteries most cancel out the voltage of the lower voltage batteries, creating the equivelant of a tiny voltage source that is short-circuited producing lots of current. THe result is that a LOT of current flows from the highest voltage batteries into the lower voltage ones trying to charge them to make all their voltages equal, and leaving only the highest battery to power the load.

It's also inversely similar to the case where you connect a bunch of diodes in parallel. THey will all have slightly different voltage drops, and the one with the lowest voltage drop will conduct everything (pretty much) while the rest will not turn on because their required voltage drops are higher but the voltage drop across them is that of the diode with lowest voltage drop that is conducting since they are all in parallel.

Click to expand...

True. Thanks.

dknguyen said:

It's also inversely similar to the case where you connect a bunch of diodes in parallel. THey will all have slightly different voltage drops, and the one with the lowest voltage drop will conduct everything (pretty much) while the rest will not turn on because their required voltage drops are higher but the voltage drop across them is that of the diode with lowest voltage drop that is conducting since they are all in parallel.

Click to expand...

An additional factor in diode current hogging is that Si diodes have a negative temperature coefficient of the forward voltage. Thus the diode with the lowest initial forward drop will be carrying the most current which will cause it to become hotter than the other diodes. This further reduces it's forward voltage drop and it carries even more current. This positive feedback effect continues until the one diode is carrying pretty much all the current.

Is it necessary to connect an EL capacitor at the O/P pin of IC? How that value is selected? Should I connect CD capacitors in parallel with EL capacitors?

EL capacitors are only required if your load has large "momentary" current draws that are too large for the CDs to supress.

One last case is some output CD capacitors are almost too good in their low ESR and that causes some linear regulators to not be stable because of how their feedback loop works- but that should all be in the regulator's datasheet about capacitor selection.

CD capacitors in parallel with EL capacitors are sometimes good if your the EL capacitors are too crappy for the high frequency stuff but the CD capacitors are too small to supply all the energy requirements. One reacts fast but the other can provide more energy. But (and there always is one), the non-ideal characterstics of capacitors means that paralleling different sizes (and kinds of capacitors) makes them interact with each other resulting in reduced performance the frequencies each capacitor was best for, than if they were just used alone. YOu don't "get the best of both worlds". WHat you are doing is sacrificing wider frequency range for quality. Really technical, probably doesn't make a difference in your case, but it does matter for things like the CPU in your computer and RF circuits.

So basically, do what you feel like. As long as the linear regulator is stable you are good. It really depends on what you are powering.