Power Supplies

[DaKraut]

I am planning on building the ultra-clean power supply from ggg:

http://www.generalguitargadgets.com/diagrams/ultra_clean_ps_sc.gif

But I am also wanting a way to switch between 9 and 18 volts, because the transistor is a variable voltage transistor I know all you have to do is change out the resistors to control it, but i cannot figure out which ones to modify, if someone could please tell me exactly which resistors to modify and their modified value I would be greatly appreciative, here is the data sheet for the transistor I am going to use:

http://www.fairchildsemi.com/ds/LM/LM317M.pdf

[JKowalski]

I am planning on building the ultra-clean power supply from ggg:

http://www.generalguitargadgets.com/diagrams/ultra_clean_ps_sc.gif

But I am also wanting a way to switch between 9 and 18 volts, because the transistor is a variable voltage transistor I know all you have to do is change out the resistors to control it, but i cannot figure out which ones to modify, if someone could please tell me exactly which resistors to modify and their modified value I would be greatly appreciative, here is the data sheet for the transistor I am going to use:

http://www.fairchildsemi.com/ds/LM/LM317M.pdf

It's not transistor - in fact, it is around 26 transistors with a bunch of resistors and a couple diodes in a transistor-type package. Take a look at figure thirteen in the datasheet here to see a complete schematic for the part. This is a three terminal regulator integrated circuit, a whole circuit in itself.

This is a good exercise in reading datasheets to design circuit snippets - the datasheet gives you the formula you need to find the output voltage, plainly stated. Find it in the datasheet and read up on the lengthy explanation of it to figure out where you need to go. After comprehension and some simple algebra you'll be where you want to be.

V_out = 1.25V * ( 1 + ( R2 / R1 ) ) + I_adj(R2)

R2 and R1 are shown in the schematic next to the equation.

If you don't feel like bothering with it... Google a LM317 calculator (there are some) and plug in numbers. But I recommend actually going through the mathematical process, learning is always nice.

[PRR]

> the datasheet gives you the formula you need to find the output voltage, plainly stated.

Bizarrely, the Fairchild sheet doesn't.

Try the National sheet: http://www.national.com/ds/LM/LM317.pdf



In the formula, ignore the "+I_adj(R2)" term. Instead, pick R1 lower than 500 ohms; then "I_adj" becomes relatively too small to matter.

Pick R1 higher than 100 ohms to avoid waste power.

The '317 tries to force 1.2V between OUT and ADJ pins. If you ground ADJ it puts out 1.2V. If you have R1 flowing current and stand the ADJ pin up on a resistor R2, then R2 adds some voltage to the output. If R2 were equal to R1 the output would be 2.4V. If R2 were ten times large then the output is 11 times 1.2V or about 13V.

Since the reference voltage is 1.2V, the math looks simpler if you pick R1 as 120 ohms or 240 ohms.

It may also help to think of your output in terms of 1.2V units. If you want 18V, that is fifteen 1.2V units. One of these is across OUT and ADJ, the other fourteen must be ADJ to ground. Therefore R2 is fourteen times larger than R1. It may be hard to find "exact" resistor ratios. But the '317 is not a precision part, and most audio circuits are VERY tolerant.It "works" on 9V or 18V, it will work on 17V or 19V nearly the same as 18V. 150+2K2 is very near 18V.

[DaKraut]

Ok, I think I can figure this out, but ggg's schematics seem weird to me, as there are two resistors in parallel for R1 and the same for R2, if i switched these to just one resistor would it have any adverse effect on the circuit, because it would be much simpler to calculate and to build, so would it make the power any less regulated?

[PRR]

> ggg's schematics seem weird to me, as there are two resistors

The designer was attempting to get "exact" values using standard resistors. Note that 274(!) and 2000 in parallel is 240.985 ohms, very close to the 240 on the suggested plan. Likewise 6810 and 2000 makes 1546 ohms. If you assume the '317 makes 1.26V, this gives 9.35V as the plan says.

No such precision is necessary. R1 likes to be 100 to 500 ohms, but the exact value is not critical. The actual reference voltage may be 1.20V to 1.30V. Your resistors are usually +/-5%. Your load (pedal etc) will usually tolerate +/-10% or +/-20% variations.

[DaKraut]

Strangely enough, this is not going to power a pedal but instead be used for phantom power to a set of active pickups. It will be supplied through the ring of a TRS cable, put into a box to power the pickups, which will have a normal TS jack for out to the amp. This is why I want such a clean power source, because the pickups are the source of the sound, so it would be bad to have a problem start there. I want it to be 9 and 18 volt switchable which is why I am asking this question, would you guys also think that a buffer for the signal would be a good idea? especially if long lengths of cable will be used?

[PRR]

Not technically "Phantom", since power isn't on the signal line. Whatever.

Active pickups should be able to drive a reasonable cable.

Active pickup designed for battery may lack additional filtering.

Breadboard it first.

[G. Hoffman]

Anyone else concerned about dumping something like 12 volts across that regulator with no heat sink?  (If you are using it for 9V.)  Obviously, its going to depend on how much current you are actually drawing, but even if you get up to 500 mA, you are needing to get rid of 5-6 watts, maybe a bit more.  Doesn't seem the best transformer to use, to me.

Myself, I'd go look at R.G.'s Spyder, and the Weber WPDLXFMR-1 transformer.  You can tie two of the 11V windings in series to drive your 18V DC circuit (just make sure you don't get them backwards!!), and by keeping your effects on different windings you can keep things as quiet as possible. 


Gabriel

[DaKraut]

That transformer would be pointless for my project in using it to power pickups,

the pickups themselves draw less than 100 microamps

and i would only need to have 1 output, not 8, so i dont need 8 secondary windings

if need be i could always throw on heatsinks where required, one is already planned for the lm317

[PRR]

> concerned about dumping something like 12 volts across that regulator with no heat sink?

No. It gets hot, it shuts-down. A wide-awake tech can figure it out: it works for a second, then quits, a smell of hot fresh plastic, and when you put your finger in the '317 it sizzles.

And we learned that this is replacing a battery. Unless it was a lantern or motocycle battery, the current will be several mA at most. If you have 20V drop at 10mA that's 0.2 Watts.... the TO-92 package will handle that e-z. Now Nick reports 100uA 0.1mA demand, reasonable for nice long life on compact battery, which makes '317 total over-kill (but cheap-enough over-kill that why not?).

Me, with ~~18V 0.1mA known up front, I'd just wire 12VAC to a diode and 470uFd cap, 5K resistor, another 470uFd cap, and out to the axe. It will be close-enough to 18V, the buffer is either 9V-picky or it will stand any reasonable voltage IF it is steady and clean. The ripple will be a fraction of a mV at first cap and microVolts at the second cap. If 18V proves to be too much, 50K or 100K instead will drop toward 9V (although it may take a minute to come up to voltage; 22uFd would be a better second cap).