Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Deskamp power supply schematic.
New Posts  All Forums:Forum Nav:

Deskamp power supply schematic.

post #1 of 112
Thread Starter 

I threw together this schematic in Eagle in order to keep an organized idea of what I wanted to accomplish as I get ready to make a more complicated and better sounding desktop headphone amplifier.

 

I'm not interested in putting batteries in it at all, so I wanted to make sure that this was the proper way to use the LM317 to regulate a DC walwart and then split it into two rails.  I used the datasheet for the LM317 and Tangent's guide on splitting rails to draw this up.  It's basically just a copy, I'm sure, of what Tangent had on there with the DC jack and LM317 stuck on there.

 

By the looks of it, I may have gotten confused about what pin on the jack was positive in Eagle, so assume that the pin leading into the 317 is the positive pin.

 

I'm not looking to make this a perfect regulator, I just want a regulated supply voltage that is split in a reliable and clean enough way to provide an upgrade to two 9v batteries.  The goal is to supply the op-amps with 30v of regulated power.  I planned to pick up a 40v walwart.

 

I am most unsure of how the pinout on the DC jack works.  If I have that setup correctly for the rail split.  If anyone could tell me if there are any really glaring mistakes here, I would appreciate it.

 

Thanks in advance.

 

EDIT: I'm sorry for any issues with the readability of the schematic or if parts aren't necessarily what they should be... I did my best, but this was my first time with the program.

 

-Robert

 

psu.jpg


Edited by miskatcitnamor - 3/18/11 at 12:30am
post #2 of 112

- The biggest mistake is that the lm317 shouldn't be referred to the virtual ground but to the negative rail.

- It is good practice to put a 10 to 100uF cap in parallel with R2.

 

For what kind of headphones amplifier is it ?

post #3 of 112

For a Desktop Amp you should be using a dual rail power supply and

do away with the virtual ground setup.

 

Something like this:

 

317_337 Schematic.png

post #4 of 112

While I agree that dual rail from AC would be best, many use DC and virtual ground to avoid having to work with transformers and live voltage.  PPAv2 and M3 are examples of amps setup this way.

post #5 of 112

Here is an easy single rail supply.

Just add an AC wall wort.

 

 

317a Schematic.png

 

And a Board for it.

 

317a Board.png

post #6 of 112
Thread Starter 

I chose to do a single rail, rail-splitter, power supply in order to keep things within my realm of understanding for now.  I was pushing it with my application of the LM317.  Attempting to do a dual rail system from AC with a rectifier would have been slightly outside that envelope at the moment.  I'm taking baby steps up from the CMoy.

 

Heck, I was planning on connecting this to an amp barely more complicated than a CMoy.  I still have a lot of reading to do before I work on the amp, though.

 

Screw it, though.  I guess I will just go with the center-tapped AC route.  The rectifier is easy and all it really seems to be is the addition of an LM337 on the negative rail.

 

I'd be all for building the circuit you posted, Avro, just one question... what are those KK1 SK129-PADs?  How do they connect to the circuit?

 

The kind of walwart I will need to look for is a center-tapped AC, right?  Or would it just be called grounded AC?  Not 100% on it.

 

Thank you all for the posts and suggestions.

 

-Robert

 

EDIT: Looking at the datasheet for the LM337, I have another question.  In your dual rail schematic up there you put the LM337 right on the negative rail.  In reference to the positive rail, the negative is currently actually at 0v, right?  So if I have a 40v AC walwart then the positive rail would be 40v and the negative 0v in reference to each other.  However the positive becomes 20v in reference to the new ground between them and the negative becomes -20v in reference to the ground between them.  If I am correct in saying that, then this is also a "virtual ground" right?

 

What step is taken from there to make an active ground?  (Probably getting ahead of myself here.)


Edited by miskatcitnamor - 3/18/11 at 7:34am
post #7 of 112

The pads you mentioned are for the heat sink.

They do not connect to the circuit at all.

 

You might be able to find a center tap transformer, but

it is more likely you will want a center tap transformer.

 

Something like this if you were mounting it on the board.

That would be the biggest you would need.

You would probably use something smaller.

 

Something like this for mounting off the board.

Again, this is the biggest you would need.

post #8 of 112
Thread Starter 

OK, I hadn't really considered putting the transformer in the case itself... might need to invest in a new case.  haha  Anyway, what kind of power cord would I need to get to get wall power to those transformers?  Would it use a standard 2 or 3 pronged power cord like what you might see from a computer power supply or monitor power supply?

post #9 of 112

Here is the board for the dual rail supply.

 

317_337 Board.png

post #10 of 112

yes

 

Something like this.

Then you can plug in a standard power cord


 

Quote:
Originally Posted by miskatcitnamor View Post

OK, I hadn't really considered putting the transformer in the case itself... might need to invest in a new case.  haha  Anyway, what kind of power cord would I need to get to get wall power to those transformers?  Would it use a standard 2 or 3 pronged power cord like what you might see from a computer power supply or monitor power supply?



 

post #11 of 112
Thread Starter 

AH!  Ok, I see.  Now using a board mounted transformer seems to be a bit of a waste of board space, so using a chassis mounted one seems more economical.  Now you posted that particular one and suggested it would be enough, I had been aiming for a 40v so that I would have enough to supply the rail splitters to ensure that I could get a +15v rail and a -15v rail so that I would be in a good healthy upper range for the op-amp down the road in the circuit.

 

The one you suggested is 25v total, so 12.5v split between each rail, right?

 

Is there an advantage to staying at a lower voltage?  (Coming from computer tweaking, I had thought more voltage without going overboard would aid in stability and performance.)  I have heard that the idea performance voltage will differ from op-amp to op-amp, so not really sure where I should take it.

post #12 of 112

You also have to consider things like power supply cap voltage ratings.

 

I pick the 25 volt center tap so I can use 35 volt caps.

The unloaded voltage on the transformer is more like +-18 volts

and you want a two times comfort zone for your filter caps.

 

The maximum voltage for most op amps is 32 volts or +-16 volts.

Some are less, some are more.

 

Unless you have 600 ohm headphones, you don't need anywhere

near the rated voltage of the op amp so it is fine to use less voltage.

 

By keeping closer to the voltage you need you keep the overall cost

down and reduce the heat generated by your components.

 

If your amp does not load the power supply that much you will get

more than the +-12.5 volts out of it. How much more is just a guess.

 

There are quite a few amp designs on here that specify +-12 rails.

Some want as much as +-15 volts.

 

I actually change transformers depending on how much voltage

I want from the power supply as I use mine for varied projects.

post #13 of 112

In response to your "Edit" question...the transformer acts a the Virtual Ground Splitter.

The center tap of the transformer is at 0 volts.

post #14 of 112
Thread Starter 

I had planned to get the 40v from the transformer and use resisters on the regulators to reduce voltage down to desired levels (taking the 40v to the +-15v I desired.)  But I am using 600ohm headphones, so if higher voltage might be needed for that, then it's something to consider.

 

So if I am using a total of 40v before the regulator, then I need caps that are rated for 80v, to be safe?

 

I'm not too worried about heat, per se.  These heatsinks seem pretty beefy and should be able to handle whatever is needed.  Worst comes to worst I can always wire a fan into the chassis.  (thinking about using an old PS2 case for this.)

 

The caps I have already picked are only 50v but if I wire them up past the regulators, shouldn't that be ok?  I could keep a couple lesser quality electrolytics before the regulators and then have the bulk of my capacitance on the output?

post #15 of 112

Yes, whatever your transformer secondary rating is, you

want at least two times the voltage headroom for your filter caps.

 

After the regulator, you can relax a bit and use the the next voltage

rating higher than the highest voltage you expect from the regulator.

 

You want the bulk of your filtering before the regulator.

The output will be quite clean after the regulator.

New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Deskamp power supply schematic.