power supply: 7809 or LM317

[Lizard King]

I built my first power supply with an LM317 and adjustable Vout.  I gave it to a friend so I'm back to the drawing board.  From looking around the web I have three options:
1) Adjustable 317
2) Fixed 317
3) Fixed 7809

I found schematics for each.  Is one way to go more suitable to driving 9V effect boxes or are they all about the same in the end?  I'm leaning towards a fixed 317:
http://www.bing.com/images/search?q=9v+dc+power+supply+schematic&qpvt=9v+dc+power+supply+schematic&FORM=IGRE#view=detail&id=BFB38E4830DC26321F700FA20C98110B0AC01E4B&selectedIndex=0

Thanks.

[amptramp]

The schematic as shown for the LM317 will give better 120 Hz suppression than the LM7809 at the cost of having more components.  If you are driving transistor circuitry, this may be important.  For op amp circuitry, not so much, since op amps have a decent power supply rejection ratio.

[R.G.]

That's true, the LM317 ...can... be set up to have better ripple rejection than a 7809. However, the simple two-resistor circuit is about the same.

In actual practice, having any regulator is miles (or kilometers) ahead of having nothing. Once you have a regulator, you're at the point of diminishing returns for most pedals.

There are always special cases, though.

[Lizard King]

I assume both the 7809 and 317 handle about the same current draw - somewhere around 1 amp or so?

[R.G.]

They are both specified to provide at least 1A before going into current limit, I think. The typical current limit value is higher. On the LM317, I think it may be as much as 1.5A. I haven't looked up the datasheets this morning.

[mth5044]

Are the 1N4002 significant? The one wrapped around the 317 (for input short circuit protection?) and the one later down the line (output short circuit/cap discharge?), not the rectification part. Would any 1N400x work? Or even 1N4148?

[R.G.]

The diodes *have* to be there to protect the regulator against input shorts.

A 4000 series is surely enough, a 4148 is probably enough. The diodes have to last for a countable-few input short circuits, possibly as few as one. The surge current is what matters, and a 4148 will live through a couple of these, probably, depending on how big the output capacitance is that it has to discharge. I can mentally concoct a setup that would kill a 4148, but in pedals a scenario like that would be inconceivably rare.

[armdnrdy]

Are the 1N4002 significant? The one wrapped around the 317 (for input short circuit protection?) and the one later down the line (output short circuit/cap discharge?), not the rectification part. Would any 1N400x work? Or even 1N4148?

The only difference between the 1N400X series is the voltage rating.
The 1N4001 is rated for 50 volts and should be good for any pedal power that I've seen.
Some pedal designs with 9-18 volt power supplies call for 1N4002's or even 4004's....that may be because they got a "deal" on a large quantity of that component.

Also:
About the power supply link that you posted:
It shows a transformer with a 25 volt secondary. This is overkill for a 9 volt supply. By the time the 25 volts passes through the full bridge rectifier and the 1000 volt smoothing cap, you're looking at around 35 volts at the input of the regulator. If you are not making this an adjustable supply, and it will be fixed at 9 volts, may I suggest a 12 volt transformer. A 12 volt transformer will give you about 16+ volts at the regulator input.

[Lizard King]

Another question.  Does one generally ground the metal case for their power supplies?

[Lizard King]

Also:
About the power supply link that you posted:
It shows a transformer with a 25 volt secondary. This is overkill for a 9 volt supply. By the time the 25 volts passes through the full bridge rectifier and the 1000 volt smoothing cap, you're looking at around 35 volts at the input of the regulator. If you are not making this an adjustable supply, and it will be fixed at 9 volts, may I suggest a 12 volt transformer. A 12 volt transformer will give you about 16+ volts at the regulator input.

That was just a circuit I found with a search.  I'd probably get a 12V transformer.  They're hard to find in a small form factor at 1A.

[armdnrdy]

You can get a low profile board mounted transformer from Digikey, Mouser, Jameco, Ebay, ect. or better yet...feed the input with a 12VAC wall transformer and make the actual power supply a smaller size.

If you look at the design of the DC Brick, although it is a DC transformer, that's what MXR did.

[Lizard King]

You can get a low profile board mounted transformer from Digikey, Mouser, Jameco, Ebay, ect. or better yet...feed the input with a 12VAC wall transformer and make the actual power supply a smaller size.

If you look at the design of the DC Brick, although it is a DC transformer, that's what MXR did.

I've looked at all those possibilities.  I like the idea of a real transformer but a wall transformer would probably be the most cost effective.  If I wanted to get real fancy I'd get the multicore weber transformer but that is a whole different price point.

[gjcamann]

The problem with the LM317's is that you have current flowing through the voltage adjust resistors, so your wasting current,  generating heat and have a more parts in your circuit. I think the extra ripple in the 7809 is insignificant if you put  power cap in your pedal.

Small bear has a nice tiny and cheap 12V transformer (but only 125mA ).
http://www.smallbearelec.com/servlet/Detail?no=1012
If you need 1 amp, then you've probably got something digital and don't care about power filtering, just use a 7809 (or 1spot) and be done with it.
For most analog effects, 125mA could easily power several pedals.

You may want to look at the TL317 from Linear, they're like the LM317 but do not require the protection diode.

Also, you may want to look at something like the linear LT3080 or LT3083
http://www.linear.com/product/LT3080
http://www.linear.com/product/LT3083

These are large improvements over the LM317, have simpler circuits with reduced part counts.

 

[armdnrdy]

Also, you may want to look at something like the linear LT3080 or LT3083
http://www.linear.com/product/LT3080
http://www.linear.com/product/LT3083

These are large improvements over the LM317, have simpler circuits with reduced part counts.

 

These regulators are a large improvement.....but at a LARGE price! and they don't seem to be readily available.

Digikey LT3080 $4.40 USD
Digikey LT3083 $8.82 USD

I think I'll stick to LM317's and LM78XX

[gjcamann]

Well nothings perfect 

[Henry89789]

Does the adjustable LM 317 allow one to adjust the output voltage to get 9, 12 or 18 volts?   For example a Fulltone GT-500 can run on from 9-18 volts. Would a power supply made from a LM317 allow one to adjust the voltage going into the GT-500 to explore all the tone possibilities at every voltage level from 9 to 18 volts?  Thanks.

[armdnrdy]

Does the adjustable LM 317 allow one to adjust the output voltage to get 9, 12 or 18 volts?   For example a Fulltone GT-500 can run on from 9-18 volts. Would a power supply made from a LM317 allow one to adjust the voltage going into the GT-500 to explore all the tone possibilities at every voltage level from 9 to 18 volts?  Thanks.

Yes.

If you make it adjustable with trim pots or a standard pot to adjust the voltage. The 317 is often used for adjustable power supplies.