AC adapter for symmetric power supply

[JanJansen]

Hi, i like to make a +15 and -15 power supply for my drummachine & synthesizer audio project.
I,m getting a 15v ~ AC/AC walladapter.

Ive been told with 15v going tru diodes (1n4003) : it will be 1.4 times higher voltage,
15 volt will be 21 volt.

I wanto select good condensators/capacitors,
is a condensator voltage rating of 35v to low to flatten 21 volts ?
What would be the right capacity value to choose for this ?, the walladapter is rated 1A.

Maybe i should have asked this in the beginner section ?
Thanks in advance, greetings

[Zero999]

My advice would be don't.

As you're building this from scratch, avoid having to use a dual power supply in the first place, so a single, rail, efficient switched mode power supply can be used.

Do you already have the schematic? If it states it needs a dual supply, it's very likely it can be converted to work off a single supply with minimal effort.

[donmr]

If you want any kind of quality audio (not just buzz and hiss), then you will need a well filtered and well regulated power supply.
You can either learn about power supply design (an interesting and useful subject), or you can buy a module that does it for you.

There are modules for every combination of inputs and outputs you might want.  They will be happy to take your rectified and crudely filtered power from a bridge rectifier and capacitor and produce good supplies for your circuits.

[Kleinstein]

To get a +-15 V supply one usually wants a transformer with 2 windings of about 15-18 V each. Not many wall adapters offer that.

35 V rated capacitors so should be OK. The typical rule of thumb is something like 2000-5000 µF for 1 A of DC current. As 15 V AC is more on the low side the caps would need to be more on the high side. For 1 A of AC current one will get something like 0.6 A of DC - this is due to the pulsed current.

Using only on winding and half way rectifier would need considerable larger (e.g. 2-2.5 times) caps. Uneven load to both sides of the +-15 V would result in DC current seen by the transformer and thus could cause some noise.

[The Soulman]

I've seen this arrangement being used in several commercial audio products, but only for low power applications (sub 100mA).
What are your power requirements?
 

[westfw]

With a 15VAC transformer, you can easily get inputs for +/-15V regulators using a voltage doubler rectifier configuration, although you'll need additional filtering capacitors since it's effectively half-wave rectification.   Also, you get half the transformer's rated current...
Hmm.  Like this, approximately:



(From http://www.subatomicglue.com/202l0g/ )

[Ian.M]

It doesn't double the voltage (each side never has a greater DC voltage than the AC peak), and D3,D4 never conduct.  OTOH if you remove the excess diodes, you are simply left with the dual half-wave rectifier circuit Kleinstein mentioned in the final paragraph of reply#3.

Also, according to Hammond's Design Guide for Rectifier Use, the derating factor for HALFWAVE Capacitor Input Load is:  I D.C. = 0.28 X Sec. I A.C.
Its fairly easy to knock up a SPICE sim and verify the derating factor. You can also check the minimum transformer output voltage required to provide enough headroom for the regulators without having to do a lot of gnarly maths.

LTSPICE sim attached.  Note the simulation is parameterised - edit the transformer parameters to match your proposed one.  lf sets the AC line voltage as a proportion of its nominal value - its currently set to 0.9 (10% low) to see if the PSU will cope during low line conditions.  If you are checking dissipations, RMS current limits etc. set it to 1.1 (10% high).  For an explanation of the diode stuff, see https://www.eevblog.com/forum/beginners/help-me-debug-my-transformer-sizing-for-my-lab-psu/msg1145623/#msg1145623

[Zero999]

If you want any kind of quality audio (not just buzz and hiss), then you will need a well filtered and well regulated power supply.
You can either learn about power supply design (an interesting and useful subject), or you can buy a module that does it for you.

Not true. A well designed audio circuit can run off an unregulated power supply. The power supply rejection of most op-amps at double the mains frequency (100Hz or 120Hz) is high enough that ripple on the supply isn't a problem, as long as the minimum valley voltage is high enough to power the circuit. Higher frequencies present a greater challenge but a linear regulator won't do much to attenuate them. Ferrite beads, RF chokes and capacitors are better suited to eliminating RFI.

With a 15VAC transformer, you can easily get inputs for +/-15V regulators using a voltage doubler rectifier configuration, although you'll need additional filtering capacitors since it's effectively half-wave rectification.   Also, you get half the transformer's rated current...
Hmm.  Like this, approximately:



(From http://www.subatomicglue.com/202l0g/ )

D3 & D4 don't do anything.

How much current is that supposed to provide? As drawn you'd be lucky to get 50mA of ripple free 15VDC, when the mains voltage is on the low side.

I suspect the filter capacitors are undersized. It seems odd that output capacitance is so high. It would be better if C7 and C8 were moved to the input side. I'd use Schottky diodes, make C1 & C2 2200uF 35V and remove C5 to C8 to reliably give 100mA.

It doesn't double the voltage (each side never has a greater DC voltage than the AC peak), and D3,D4 never conduct.  OTOH if you remove the excess diodes, you are simply left with the dual half-wave rectifier circuit Kleinstein mentioned in the final paragraph of reply#3.

Also, according to Hammond's Design Guide for Rectifier Use, the derating factor for HALFWAVE Capacitor Input Load is:  I D.C. = 0.28 X Sec. I A.C.
Its fairly easy to knock up a SPICE sim and verify the derating factor. You can also check the minimum transformer output voltage required to provide enough headroom for the regulators without having to do a lot of gnarly maths.

LTSPICE sim attached.  Note the simulation is parameterised - edit the transformer parameters to match your proposed one.  lf sets the AC line voltage as a proportion of its nominal value - its currently set to 0.9 (10% low) to see if the PSU will cope during low line conditions.  If you are checking dissipations, RMS current limits etc. set it to 1.1 (10% high).  For an explanation of the diode stuff, see https://www.eevblog.com/forum/beginners/help-me-debug-my-transformer-sizing-for-my-lab-psu/msg1145623/#msg1145623

Technically it is a voltage doubler, since the total output voltage (across C1 & C2) is double the peak AC input voltage.

[JanJansen]

If you want any kind of quality audio (not just buzz and hiss), then you will need a well filtered and well regulated power supply.
You can either learn about power supply design (an interesting and useful subject), or you can buy a module that does it for you.

I have this Recom 1215D DC/DC converter, that ripple from 50mv is to much for me, i dont want HF ripple at all.
I also going to avoid mains voltage, thats why i do this with wall adapter.
If i connect the Recom 1215D Vin- to COM, the ripple is gone, only slow peaks that go higher then the ripple voltage.

I've seen this arrangement being used in several commercial audio products, but only for low power applications (sub 100mA).
What are your power requirements?

Not so much, i build a few drumsounds + maybe mixer with eq, about 100mAh each + & - 15v.

Thats about what i trying to do for the power supply, without those extra diodes and bigger caps.

My advice would be don't.

As you're building this from scratch, avoid having to use a dual power supply in the first place, so a single, rail, efficient switched mode power supply can be used.

Do you already have the schematic? If it states it needs a dual supply, it's very likely it can be converted to work off a single supply with minimal effort.

I,m making this schematic ( maybe twice if it works ), and another drumsound, it cannot be done without symmetrical power supply,
if you wanto figure it out how it will cost you more then a week, your welcome to make this schematic single supply if you like, or tell me why it is not possible :

[Zero999]

I,m making this schematic ( maybe twice if it works ), and another drumsound, it cannot be done without symmetrical power supply,
if you wanto figure it out how it will cost you more then a week, your welcome to make this schematic single supply if you like, or tell me why it is not possible :

It's fairly simple. You need a power supply with double the voltage (30V for a +/-15V supply) and the TLE2426. Unfortunately 30V isn't quite so widely available but 24V will probably do, giving you +/-12V.

[westfw]

How much current is that supposed to provide? As drawn you'd be lucky to get 50mA of ripple free 15VDC, when the mains voltage is on the low side.

Why do you say that?   Just the cap size?  (yeah, it looks pretty bad.  For a half-wave rectifier, I found: "For 10 % ripple, C = (5 * Io)/(Vs* f)"  (Vs is peak input voltage.)  For 15V RMS transformer, providing 200mA, that comes out to about 8000uF !)

[Zero999]

How much current is that supposed to provide? As drawn you'd be lucky to get 50mA of ripple free 15VDC, when the mains voltage is on the low side.

Why do you say that?   Just the cap size?  (yeah, it looks pretty bad.  For a half-wave rectifier, I found: "For 10 % ripple, C = (5 * Io)/(Vs* f)"  (Vs is peak input voltage.)  For 15V RMS transformer, providing 200mA, that comes out to about 8000uF !)

Size of the capacitor, regulation of the mains and the drop-out voltage of the LM7815.

[JanJansen]

It's fairly simple. You need a power supply with double the voltage (30V for a +/-15V supply) and the TLE2426. Unfortunately 30V isn't quite so widely available but 24V will probably do, giving you +/-12V.

Thanks for the advice, i will try that part, thanks.
Will this have a problem with unregulated powersupply`s ?, they dont give stable voltage, and adjusts the sound volume & frequency ?
So i need a 30v regulated wall adapter, nice.

Wait, its not for much milliampere, i need something better maybe ?
Since the 5v regulator needs to be on the same ground to make it usable, i think you cannot connect Vin- to -12 volt.

[Zero999]

It's fairly simple. You need a power supply with double the voltage (30V for a +/-15V supply) and the TLE2426. Unfortunately 30V isn't quite so widely available but 24V will probably do, giving you +/-12V.

Thanks for the advice, i will try that part, thanks.
Will this have a problem with unregulated powersupply`s ?, they dont give stable voltage, and adjusts the sound volume & frequency ?
So i need a 30v regulated wall adapter, nice.

Yes, a rail splitter doesn't regulate the voltage, so you need a regulated power supply.

Wait, its not for much milliampere, i need something better maybe ?

How much current do you need?

Don't forget that with the TLE2426, it's only the current through 0V, i.e. the difference in current taken from +V and -V which counts. For example, if 100mA is taken from +V and 110mA from -V, then the TLE2426 is only supplying -10mA.

Since the 5v regulator needs to be on the same ground to make it usable, i think you cannot connect Vin- to -12 volt.

You didn't say anything about a +V power supply before.

The 5V regulator's 0V needs to go to the output of the TLE2426 but check the current is within its maximum rating. If the you need more current than the TLE2426 can supply then you could try adding pass transistors to it. Failing that there are other possibilities such as a, potential divider with unity gain op-amp, driving a couple of transistors.

[JanJansen]

Yes i also need 5v for the MCU, with LED and stuff.
Arent there bigger parts for more milliamps ?
Dont forget the elkos need to be filled also.

[Zero999]

Yes i also need 5v for the MCU, with LED and stuff.
Arent there bigger parts for more milliamps ?

That's the biggest you're going to get. Fortunately a couple of power transistors don't cost much.

Anyway, you still haven't said how much current you need?

Another alternative is to use a single power supply audio amplifier IC which biases its output to half the supply rail.


http://powersupply33.com/simple-split-rail-power-supply-lm380.html

I probably wouldn't bother with the 1M pot. as it should be close enough for your application without any adjustment.

Dont forget the elkos need to be filled also.

What's an elkos?

[JanJansen]

Hi, a elko is a elektrolithic condensator.

I still dont get it how you use the transistors with the 1 ohm resistor.
I dont know how much amps, it also need to charge resistors, that takes the maximum.

[Zero999]

Hi, a elko is a elektrolithic condensator.

Oh, you mean electrolytic capacitors. Well that's not a problem. It's only a surge. It doesn't have to provide the current for long time.

I still dont get it how you use the transistors with the 1 ohm resistor.

To understand how this works you need to look at the internal schematic for the TLE2426, which consists of an op-amp buffer and potential divider.

The external transistors start to turn on and divert current around the TLE2426, when the current exceeds a certain level determined by V_BE/R_BE, where V_BE is about 0.6V. 4k7 was too high. It should have been 47R for 0.6/47 = 12.8mA. R_E reduces the gain of the transistors slightly, making the circuit more stable and less likely to oscillate.

I dont know how much amps, it also need to charge resistors, that takes the maximum.

Resistors don't hold any charge. The part you posted should work fine off the TLE2426, without any external transistors. but you need to look at the whole schematic to determine the current requirements.