What happens when you have two voltage regulators in series?

[Mbira]

Does it just halve the amount of amps that each pulls? Can they then recombine as a voltage doubler?

 

[bcarso]

In series the current is the same except for what's used to power the error amp and reference. Do you perhaps mean to ask "in parallel?" If the two were absolutely identical they would share the net current from the unregulated source, and the load current, equally. But they never are, and when regulators are paralleled it's usually a situation in which one is a master and the others slaves.

An exception would be current regulators, which are generally happy having outputs paralleled.

Think of most regulators as putting a variable resistance between the unregulated source and the load. Then their error amp compares the load voltage, or more likely a precise fraction of it, with the reference voltage, and adjusts that resistance accordingly. They aren't batteries or generators---all they do is impede current flow to varying degrees, to cause the output to remain stable despite input voltage changes and output current changes. As they do this they burn up energy, converting electrical energy to heat.

 

[Svart]

you can do it to spread the wasted heat but you better be sure that ONE regulator can safely handle the WHOLE load if the others were to fail. This is the same with MOSFETs, you see them parallelled all the time but the failure mode of one would put the whole load on whatever is left.

Usually you see them going through low value Rs to balance the load a little better.

I would likely use the vreg as a master and a BJT as the actual power device. I think the LM317 datasheet shows this.. But it would probably just be as easy to create a zener-follower Vreg with better specs all around.

 

[Mbira]

Doh, yes, I meant parallel...

My intention was to use two 7815's in an application that uses more amps than one could handle, but I guess that's playing with fire?

 

[alk509]

Yeah, that won't work... Your best bet is probably to use some hefty transistor on the output, like shown on oh-so-many datasheets. How much current you need?

Peace,
Al.

 

[bcarso]

[quote author="Mbira"]Doh, yes, I meant parallel...

My intention was to use two 7815's in an application that uses more amps than one could handle, but I guess that's playing with fire?[/quote]

Don't tie them together directly---it's an inherently unstable situation. The reference voltages are never identical, so the one that is slightly higher will dominate. But even if they were precisely identical the feedback loop stability is unpredictable. It's like tying the outputs of opamps together---generally a sure-fire recipe for oscillation, at the least.* It's one of the downfalls of the otherwise useful MFB inverting highpass active-filter topology.

There's probably a way, with a few external components, to use two with one slaved to the other. Of course two or more could be isolated with output diodes, but besides dropping the voltage the regulation will suck. Better to use an external power device as Al says, as shown in the datasheet app notes, or as Svart says, roll a discrete regulator.


*unless you want an oscillation to occur---remember the old adage, Amplifiers oscillate, and oscillators don't. :razz:

 

[Mbira]

Thanks guys. I'll just use a higher current raiting regulator. :wink: