What would you recommend for supplying power for multiple voltages?

[acehobojoe]

I don't want to waste power, and I need to power a couple of 4.5v op amps in my circuit and don't know how to do it properly.

[Mustachio]

Check out voltage dividers.




::edit::
I think two 10k resistors should do it, if your V in is 9v.

[MaxPower]

You might need to take into account the current required to figure out the exact size of resistors needed. Maybe not so much with op amps...

[GibsonGM]

You might need to take into account the current required to figure out the exact size of resistors needed. Maybe not so much with op amps...

^^  Yup.

Do some thinking about how much power you need...~10mA?   Do some quick easy math:  9/YOUR RESISTOR = current.
Apply that to the divider.

2, 10K resistors will be 9/20,000 = only 450uA!  That's a bias voltage, different application.    You might need something more like two 330 ohm resistors in series...9/660= .014A, or 14mA.   Pretty close to 10mA, gives you some 'stiffness' there.   You regulated better with a bit more current available (actually, much more would be better, but then we need to dissipate the 'extra'...).   

Then check power consumption:  9v * .014A = .126W, so you are ok with 1/4W resistors...don't forget a bypass cap, 10u-100u....

Another way to do it: use a 4.7V zener diode and series resistor to set up the voltage you need....use 1/2W rated part.

ANOTHER way to do it: get an LM7805 regulator, and at the output put in a diode to drop the voltage to ~4.3V.

[R.G.]

LM317

[merlinb]

LM317

Or LM7805. No brain required.

[acehobojoe]

but then if it is regulated... doesn't it use up pretty much all of the available power?

[Jdansti]

I might be wrong, but it seems that whether you use an IC regulator, a Zener diode, or a vintage voltage divider, you're going to have to dissipate excess power.

[GibsonGM]

I might be wrong, but it seems that whether you use an IC regulator, a Zener diode, or a vintage divider, you're going to have to dissipate excess power.

Yup, especially for the resistive divider.  To regulate well, it has to have about 10x the current drawn in it!!   Don't know what the others require, but surely you should try to 'pad it', to allow your circuit room to 'breathe'...

Ace:  "Regulated" just means that if the input voltage increases or output voltage decreases, the regulator will adjust itself to keep you at the nominal voltage, pretty tightly.   It won't "use up all available power" by any means.  If you DRAW more than your source, then yes you'll lose voltage pre-regulator and it will drop out.       The other types will just stop regulating and voltage will drop. 
I guess the easy answer for that is that your 9V can run a 9V circuit AND the regulator, as long as your main power supply can deliver the current you are asking it to. Won't hurt to try it.    Note that the IC regulators require a couple of volts to operate, so a 5V 7805 might need 7, 8 volts input before it will do the job.    You're cool at 9V.

Whatever you pic, watch out that you don't exceed power dissipation ratings!    Not too likely (esp with IC reg...) but hey, worth noting, someday you probably will do that and get to see some smoke

[Jdansti]

I don't want to waste power, and I need to power a couple of 4.5v op amps in my circuit and don't know how to do it properly.

You're concerns of wasting power might not be warranted. Can you give us some specifics about your power source?

[R.G.]

LM317

Or LM7805. No brain required.

... or output voltage adjustment possible.

Then there's the LM386, which auto-adjusts to half its power supply and has low idle current, too.

I might be wrong, but it seems that whether you use an IC regulator, a Zener diode, or a vintage voltage divider, you're going to have to dissipate excess power.

Yes. If you do **anything** you're going to have to dissipate power. The question is entirely "how much?"

Resistor strings are the worst, as they have to let about 10x the net current in or out of the reference point through to remain about right in voltage. Shunt retgulators - zeners and a few ICs - can be better or worse than resistors depending on how accurately you know or guess the current in/out of the reference node. Series regulators are much better if they are biased for low standing currents. Push-pull series regulators with low standing currents, like the LM386 or an opamp with a low-current reference itself and class AB output stage are about the same.

And as the man says:

You're concerns of wasting power might not be warranted. Can you give us some specifics about your power source?

Without more info about what you're trying to do, all you're going to get are random answers.

A really important thing is why power conservation is important, as the amount of power even in a resistor string is pretty trivial compared to what's wasted in an LED plus current limit to tell people the thing's on.

[petemoore]

 For stable voltage, a power supply [AC/DC conversion + regulator] doesn't voltage drop out like battery, the heat dissipation factor now looks small and seems insignifigant because AC power is way cheaper than batteries, and hasn't required a trip to the store to get more when they're closed.
The best way I can help explain it is to ask you to checkout, while considering a Spyder [type] Power Supply, Power Supplies for Effects at GEO.
Anything less than you need can be met or exceeded by individual regulated power supplies [basically lifts all grounds w/regulated supply circuits], otherwise a simpler or more immediate solution may work just as well in all or certain situations.

[acehobojoe]

Ok, here is a specific example of a circuit I'm working with.


http://i.imgur.com/EOUSEzM.png

It has 3 IC's. How on earth do I power that many things with just 9v??

and another example would be this big muff I'm working on, using the MPQ3904 Quad transistor. It needs about 4v, but then how would I run an op amp along with that, or split the available power?

[bluebunny]

You should have no problem powering three ICs from 9V, even from a battery (though it may not last very long).  The picture you posted also appears to be incomplete (unless I'm very tired and not seeing for looking...): there are references to 5V and 4.5V, but it's not shown where these come from.

The 5V is doubtless being supplied by a 7805 or 78L05 regulator from the V_CC supply (which is 9V), and is the power for PT2399.  And the 4.5V is a V_REF (half-supply "bias") being generated by a resistor-divider network from the 9V supply.

[R.G.]

It has 3 IC's. How on earth do I power that many things with just 9v??

and another example would be this big muff I'm working on, using the MPQ3904 Quad transistor. It needs about 4v, but then how would I run an op amp along with that, or split the available power?

Ah, I see the problem.  You don't understand the relationship between voltage, current, and power as it's applied in circuits.

Neither voltage nor current are "used up". Power is "used up". All things connected in parallel to a voltage source get the same voltage, and each pulls the amount of current it would independently pull if it were the only thing attached to the voltage source. All things connected in series with a current source will have the same current through them, and the same voltage across them that they would have if they were the only part with that current through them

What gets unsatisfactory is when you supply the wrong voltage to something, and it cannot get the current it needs, or too much current flows because the voltage is too high. In that case, the power supplied is either too low and it won't work right, or it's too high and it melts or fries.

ICs are generally particular about the applied voltage. Some of them work OK at well under their maximum rated voltage, like opamps. But nearly all of them die if you force too much voltage into them. If you're using ICs, plan on being able to supply them with enough but not too much voltage, per their datasheets.

The schematic you show has a variety of ICs which may have different "absolute maximum" voltages listed in their datasheets. For instance, the PT2399 will die if you put 9V across it. So to use that chip at all, you need a voltage source to supply it that is what the chip needs. If what you have is 9V, then you're going to have to do something to make the right voltage for it. To make a long story shorter, you're going to use a regulator IC to run that from 9V as a raw power supply. The other options are not practical. Same for the "Taptation", which is actually a preprogrammed PIC microcontroller. Those will die if fed more than about 5.5V on their power pin. But chances are they can work from the same regulator that feeds the PT2399. The TL072s will run from 9V directly, or up to about 30V just fine.

So you need to look up the datasheet for each IC, find out what it's necessary voltage supply limits are, what current it needs to pull, and then design how you will get that voltage at at least that much current to each one. As a practical matter, you're going to use IC regulators for the lower power supply voltages, and either ICs or resistor dividers for the bias voltages that don't have to supply much current.

I am synopsizing a moderately complex subject heavily. Ask for clarification where you need it.

[acehobojoe]

wow, thank you. That makes so much sense now.


So, after the 4.5 or 5 supply to the pt2399, I can take the power from the vref and apply it to the op amps? or from the VCO?

[PRR]

> It needs about 4v, but then how would I run an op amp along with that, or split the available power?

I have a lamp that needs 120V. How would I run a TV along with that?

Everybody has one of those 6-way wall-power strips so we can distribute some 120V to the TV, the lamp, the cable-box, the PlayStation, etc. Inside, all the outlets are just wired together.

Apologies if some of this is repetitious....

That drawing is, as Marc suggests, incomplete? I see only Vcc, 5V, and +4.5V. There is no clue where the Vcc comes from, though there is a "9V" hangin'out doing nothing, and IMHO 9V might be a good vcc supply for most of this.

So 9V to all the Vcc points. Just like your 6-way power strip, except just the wires, you don't need outlets.

9V also to a LM7805 voltage regulator, because the '2399 stupidly does not eat 9V.

The 4.5V is surely low-low current, and nominally "half of Vcc" (even if Vcc wanders to 7V or 10V), so is probably the usual voltage-divider.

There IS another way to wire things. I have ten 12V lamps. I want to run them on 120V. Or I have 12V 12V 12V 50V 35V tube heaters. In this case you can wire in *series*. 12V+12V+12V... +12V=120V. However note that both my examples just glow, they do not have signal or control connections one to another. I can think of stacked-up audio electronics but they are freaks. Mostly you run parts on the same (or related) power wires so they are all on the same level for audio/control interconnections.

[acehobojoe]

Thanks man