Power Supply design

I have been working on this power supply design. I already have the parts, so I want to keep the values if possible. I based it on other designs and some figuaring but I am not sure I did it right. I originally desgned to be used with a 16V ac wall wart, which I have. I probably would use a CT supply now, but this is what I got.

Anyway, I think I am pretty close but because there is no CT on the transformer, I am concerned I may have done the voltage tripler wrong. Should the first diode be hooked up to the input I am using for the ground, or the other one - the way I am doing now?

Any other issues? This is my first attempt at a power supply so I am sure I could have done better, but I am learning.

I just linked to this in another thread, this is what you're trying to do.

http://www.jlmaudio.com/ACDCVer2sch.pdf

Your tripler ground should not be connected to AC.

Here's some good reading on PSU design..

http://www.groupdiy.com/index.php?topic=31858.0

gemini86 said:

Your tripler ground should not be connected to AC.

Click to expand...

Thank for the reply. I have some questions.
Not quite sure what I would hook the ground to. I am not using a center tapped transformer. I was originally using something like this -
http://sound.westhost.com/project05.htm
I also found this -
http://freecircuitdiagram.com/2008/08/14/regulated-15-15-volts-symmetric-power-supply-from-12-volt-transformer/

I already breadboarded the main power supply part, it seems to work fine. I was going to do a home made board and add phantom power. That is where the question comes up on the tripler.

If I don't tie it to one of the AC inputs, what do I use for a ground?

ddt said:

Here's some good reading on PSU design..

http://www.groupdiy.com/index.php?topic=31858.0

Click to expand...

Good stuff. I didn't see this before but there is SO much info here, you could read for a lifetime.
Thanks.

ripgtr said:

gemini86 said:

Your tripler ground should not be connected to AC.

Click to expand...

Thank for the reply. I have some questions.
Not quite sure what I would hook the ground to.

Click to expand...

Your connection is "correct", meaning, it will work somehow.

I am not using a center tapped transformer.

Click to expand...

The problem is that the rectification is single-wave instead of full-wave; the consequence is ripple is much higher, so high that the regulators will have a hard time. I think you can expect about 1/4th the power you would have with a transformer rated at the same power but center-tapped.

That is where the question comes up on the tripler.
If I don't tie it to one of the AC inputs, what do I use for a ground?

Click to expand...

You don't have an option here, you must connect one side of the secondary to ground.

availible power might be your concern, as abbey already said.
Just some edits to your circuit.

Harpo said:

availible power might be your concern, as abbey already said.
Just some edits to your circuit.

Click to expand...

Oh, excellent! Thanks.
The resistor was just a mistake, I have read the data sheet and should have caught that. The trim pot too, as, the way I had it hooked up there wouldn't do anything. Got that, should have double checked.

On the rectifier, it looks like you are taking it from the full wave rectified side going into he 7815 and then using a voltage doubler to get the same effect. Without the ripple of the single wave. Wow, very cool idea. Am I getting that?

I did read the artcle ddt linked. Really good info on the traces being wide for the power and keeping seperate traces. I had to reread it several times and got a lot of it. Some of it I have to digest more.

Thanks, all. Great help and I do appreciate it. I will be building this out in the next couple of weeks and will report back. Any other suggestions are welcome, of course.

abbey road d enfer said:

ripgtr said:

gemini86 said:

Your tripler ground should not be connected to AC.

Click to expand...

Thank for the reply. I have some questions.
Not quite sure what I would hook the ground to.

Click to expand...

Your connection is "correct", meaning, it will work somehow.

I am not using a center tapped transformer.

Click to expand...

The problem is that the rectification is single-wave instead of full-wave; the consequence is ripple is much higher, so high that the regulators will have a hard time. I think you can expect about 1/4th the power you would have with a transformer rated at the same power but center-tapped.

That is where the question comes up on the tripler.
If I don't tie it to one of the AC inputs, what do I use for a ground?

Click to expand...

You don't have an option here, you must connect one side of the secondary to ground.

Click to expand...

The schemo I linked to shows how one would use a non-ct wallwart. I would go fill wave doubler, rather than half wave. But whatever work.

.

gemini86 said:

The schemo I linked to shows how one would use a non-ct wallwart. I would go full wave doubler, rather than half wave. But whatever work.

Click to expand...

Agree. And apart from the resistor mistake, the original post schemo is correct.

For a setup like this, I've once had a lot of problems with 78/7912 regulators "latching up" when powering up - if one regulator sees power a half-cycle before the other one, it can draw the other's output beyond zero for a moment (through whatever consumer it's connected to), which blocks one output voltage from showing up.

See the old thread http://recording.org/diy-pro-audio-forum/42410-need-help-simple-psu-question.html for details

Jakob E.

Thanks, everyone.
A lot here to digest. I have some questions but I want to research a bit before I ask.

So, the more I look at this, the more I seem to be thinking I just need to get a CT transformer and go with that. I can get a Toroidal one at mouser for $20. It is the same one I used in the SB4000 and would probably make this easier.

Am I kind of wasting my effort on using the wall wart, or is this doable?

ripgtr said:

So, the more I look at this, the more I seem to be thinking I just need to get a CT transformer and go with that. I can get a Toroidal one at mouser for $20. It is the same one I used in the SB4000 and would probably make this easier.

Am I kind of wasting my effort on using the wall wart, or is this doable?

Click to expand...

This design: http://www.jlmaudio.com/ACDCVer2sch.pdf works, and will do what you are looking for. I have used it in about 4 projects so far with 16v and 18v wall warts and have had great results. On the most recent two I used a CRC filter on the input between the wall wart and the JLM circuit. This reduces the ripple seen by the regulators.

Wall warts have the great advantage of meaning you don't have to deal with mains voltages inside your box, or magnetic fields wreaking havoc with your sensitive mic preamp circuit. A disadvantage is that this field is then outside the box, possibly interfering with another peice of equipment!

MikeClev said:

This design: http://www.jlmaudio.com/ACDCVer2sch.pdf works, and will do what you are looking for.

Click to expand...

As PRR and I have said, it would work. But you need to be aware that the smoothing caps must be made much larger (3-4x) for single-wave rectification than for full-wave, for a given current.

OK, I have taken the JLM schematic and played with. I cut out all the regulators in the schematic at this point, I am just trying to get the rectifiers correct. I used only half the diodes as the other side of the bridge rectifiers did not seem to be hooked up on the wall wart part, if I am reading this correctly. The top diagram of the power input shows "AC not connected" which it wouldn't be on a non CT transformer. OK, I went through and redrew this.

I have some questions -

PRR shows in the diagram on reply #8, the 2nd set of diodes being useless. I disconnected them and see no voltage difference. Is this because they are referencing the same voltage as the ground and thus there is no difference (thus no voltage)?

The phantom power, after the diodes, where the resistors and caps are. Is this a voltage divider? I have googled this and I can't figuare out what this is doing. What is the advantage over a single large cap?

On the phantom part, the bottom diode is hooked up AFTER the recifiers for the 7815/7915. This is what I am seeing in the JLM schematic and what Harpo drew. PRR marked on his reply "no AC, thus no doubling". I kind of understand this but am a bit confused which is right.


I knew there would be a lot to this and I would have to do some learning but there seems to be a lot going on. I was reading about ripple and cap size, and you have to know the draw, which seems like a lot of work/figuaring and acceptable ripple - how do you decide what is acceptable? I read the 317 data sheet and I don't quite get it. I get the theory on what the cap does and why it has to be big enough (enough of a reserve to not drop too much between cycles) and not too big (draws too much to charge). Seems like the math is pretty complicated. But I am going to keep on slugging at this, I like learning this stuff and have always wanted to get a better understanding of it. So, comments are most welcome and thanks to those who have replied.

ripgtr said:

I used only half the diodes as the other side of the bridge rectifiers did not seem to be hooked up on the wall wart part, if I am reading this correctly. The top diagram of the power input shows "AC not connected" which it wouldn't be on a non CT transformer.

Click to expand...

Right!

ripgtr said:

PRR shows in the diagram on reply #8, the 2nd set of diodes being useless. I disconnected them and see no voltage difference. Is this because they are referencing the same voltage as the ground and thus there is no difference (thus no voltage)?

Click to expand...

Right again!

ripgtr said:

The phantom power, after the diodes, where the resistors and caps are. Is this a voltage divider? I have googled this and I can't figuare out what this is doing. What is the advantage over a single large cap?

Click to expand...

This is a voltage divider, correct. JLM use this arrangement so that you can use the same 1000uF 50volt caps all over the circuit. (Easier and cheaper to source all the same type). The voltage divider makes sure that each cap only sees half the voltage, which will be peaking at double or triple the voltage of the transformer...  You could replace this with a single 100v cap of half the capacity if you felt so inclined.

ripgtr said:

On the phantom part, the bottom diode is hooked up AFTER the recifiers for the 7815/7915. This is what I am seeing in the JLM schematic and what Harpo drew. PRR marked on his reply "no AC, thus no doubling". I kind of understand this but am a bit confused which is right.

Click to expand...

Read this page on voltage doublers/multipliers: http://www.play-hookey.com/ac_theory/ps_v_multipliers.html

The voltage doubler in the JLM circuit can be either a voltage doubler, or tripler, depending on how that jumper is set. Connect the bottom of the diodes to ground and you have a doubler, connected to before the regulator (where you have whatever your transformer outputs) and you have a tripler.

I wouldn't worry too much about specific current draw... the JLM circuit along with a 500ma or so wall wart will power at least a pair of preamps, probably 4 without trouble unless they are real power hogs...

That's right, you can stack two caps on top of eachother to get double their voltage rating, and half their capacitance, but you MUST divide the voltage across both caps.

MikeClev said:

Right!

Click to expand...

OK, thank you. So I guess I do understand this a bit.

Yea, the comment about No AC thus no doubling confused me. I did read the play hokey page (a great resourse - I can actaully understand what they explain and have learned a LOT there), so I thought it should work.

And I got the part about the doubler right, that is incouraging. Yea, I can see it when you explain it. Not sure I wouldn't just get one bigger cap instead of 2 caps and 2 extra resistors, but I guess I can see the advantage of smaller caps and using the same ones. Yea, 50 volt would not be enough there, 100V makes more sense, we are talking probably 65 - 70V there I am roughly fiquaring.

OK, I am going to breadboard this, see what happens. I have the supply side already working just have to add the phantom side. It is just a That 1510 pre with no balanced out, so I don't think it is drawing all that much. I was hoping to get two channels out of it, that is more than I need usually.

the point of using two caps and two extra resistors is that 50v parts are cheaper, and you're buying a bunch, so that gives you a discount as well (usually).