Power Supply dropping resistor placement..

[antonis]

Quick querry:

Suppose we want to drop a certain amount of voltage between an unregulated supply and a voltage regulator using a power resistor..
(I know it's a brute force implementation but let it be..)
Conventional implementation is resistor placement between reservoir capacitor and regulator input..

What if that resistor is placed between rectifier output and reservoir capacitor, to form a LPF with respective cap..??
(of course, its value should be recalculated  due to voltage drop participation in secondary winding V_RMS..)

In other words, could it be faced up as a rectifier of much higher forward voltage drop or I definately need a big cap of coffee..??

[PRR]

It will significantly drop your DC volts. You need a much smaller resistor, maybe 1/6X.

It may reduce the pip at the top of the ripple wave. But the regulator can eat that. It will worsen the dip at the bottom of the wave, and the regulator can't fill that in.

And-- I hate to refer to "everybody does it", but.... nobody does this. Not in products, not in references like TI Voltage Regulator Handbook (37 MEG PDF) or Analysis of Rectifier Operation, SCHADE (17 MEG PDF).

[antonis]

Thanx Paul..

>It will significantly drop your DC volts<
That's exactly what I want..

To be more specific:
It's all about a 9V/400mA regulated PS, using a TO-220 LM7809 regulator, inside a sealed plastic enclosure..
(actually, it's a 4 isolated ouputs supply fed from 2 X 2 X 12VAC secondaries..)
Due to minimized ventilation conditions, regulators dropout should be 2V (min Vin-Vout) + 650mV (reservoir cap estimated ripple) + 5-10% for mains fluctuations..
That results into about 3.5V drop between reservoir cap and regulator input, obtained via a 8R2/2W series resistor..
(which resistor, in case of been placed before the cap should serve both for LPF item and inrush current limiter..)

But, as I realize (thanx again) , it might worsen overall regulation so I have to face it up in another way..
(maybe a series pass transistor capable to dissipate those extra 1.4Watts should be a convenient way..)

[PRR]

> minimized ventilation conditions

The total heat to be taken-away does not change, chip or resistor.

There is a potential advantage in that a dumb resistor "can" run red-hot and a chip can't. But a red-hot resistor must be placed AWAY from chips. It is unlikely to improve the situation unless you can fit a chimney (and know which way is "up").

[Rob Strand]

>It will significantly drop your DC volts<
That's exactly what I want..

It's only the value of the resistor.  I suspect in practice the series resistor before the cap will be about 1/4th to 1/3rd the value of a resistor after the cap.   The rectifier only conducts for part of the cycle so more current needs to flow through the resistor before the cap, and hence the value needs to be smaller for the same voltage drop.   In most configurations, there's already some series resistance before the cap due to the transformer winding resistance - that contributes voltage drop as well - normally put under the umbrella of regulation.

I will tell you a resistor before the cap doesn't help filtering!  When the rectifier is off the resistor cannot help filtering as the ripple is just the cap discharging.  When you have a cap after the rectifier and a second cap after the series resistor, the first cap supplies current through the resistor while the rectifiers are off and that's is what helps the filtering.

[antonis]

I think I'll lay down all the regulators on a aluminium sheet (190 x 113 x 2,5 mm) on the outer surface of the lid of the enclosure and use it only for winter outdoor gigs, hopping for a windy day..

Or just go for a metal enclosure with regulators mounted on it..

[antonis]

Thanx Rob..

I was considering about 3ms for diodes conduction angle (capacitor charging and load current time for 50Hz mains) which time lies between 1/4 and 1/3 of full wave rectified frequency..

[Rob Strand]

I think I'll lay down all the regulators on a aluminium sheet (190 x 113 x 2,5 mm) on the outer surface of the lid of the enclosure and use it only for winter outdoor gigs, hopping for a windy day..

Or just go for a metal enclosure with regulators mounted on it..

Getting the heat outside of the box is always a good move.   Hot things locked inside of a plastic box will eventually heat up.   I'm sure a lot of routers fail because of this, since they are left on permanently and eventually heat up and cook for years.

The three terminal regulators are cheap so it might be possible to split the load into two regulators but that can be risky for digital circuits.

It's not a crazy idea to use series resistor to offload some power dissipation.  Dissipating heat in resistors can be more reliable than semiconductors.   Where to draw the line comes down to a lot of specifics.   

If the load current is constant then a series resistor can work.  If the load current varies then you have to make sure the resistor is small enough not to let the regulator drop out (sometimes a few diode drops are a better choice for varying load current).    If the input voltage is already close to regulator point the series resistors can be risky since the regulator can dropout when the mains voltage is low.

In the old TV days they use to design their own discrete regulators and to off-load power dissipation they would put a resistor from input to output to reduce the current through the pass transistor.  The idea works well provided the load current doesn't drop.     It's not recommended at all for three terminal regulators since the bypass resistor can blow up the output zener which is inside those regulators.  Also possible to get some weird behaviour if the regulator is current limiting.

[antonis]

To be honest, 400mA should be the *ideal* maximum current draw from a 12V/5VA secondary winding..
I do know than anything over 300mA, or so, will kneel down the transformer but I want to take care of possible snag, without extra current limiting circuitry..

On a second thought, why not..??
(a 2R2 resistor across B-E of a BJT should make the job - and a couple of resistors should upgrade it to foldback current limiter..)

P.S.
If I dig into more, I might result into a power supply more expensive than the powered effect..

edit: In case of using two regulators, I prefer to share voltage drop instead of current..
e.g. in 15V -> 78012 -> 7809 -> out..