Power supply problems

[MohiZ]

I built this power supply with my own vero layout:

http://www.generalguitargadgets.com/pdf/ggg_bipolar_ps.pdf

I'm using a 12V AC wall wart and 9V regulators. The -9V voltage is a steady -9.13 volts but the +9V output voltage is less than 9 volts and fluctuates about. Also, the +9V regulator gets VERY hot, while the -9V regulator remains cool. The regulator input voltages are about -19V and +19V with respect to ground. This problem is present even though the supply isn't connected to any load.

What could cause this? Should I use a bridge rectifier instead of just two diodes? One of the AC adapter's leads is labeled (-) so I connected that to ground.

[km-r]

7809 and 7909 have DIFFERENT pinouts... noted this? 

[MohiZ]

yes, I checked the datasheets. could the ceramic caps have failed or something? I'm hoping someone could point me in the right direction with this.

[MohiZ]

Hmm, it seems the +9V voltage regulator was busted ... I wonder why... I replaced it and the voltages come out ok and, but now I notice that the case of the -9V regulator 7909 holds the same potential as its input pin (-19 volts), while the 7809 case is at ground potential. This means that I can't screw them to the same heat sink. Is THIS normal?

It seems although they have a different pinout, both of their case is connected to the middle pin (ground on the 7809, input on the 7909).

EDIT: According to the datasheet, this is indeed the case. I just find it a little weird.

[Ice-9]

yes thats correct the cases are connected differently, so if your going to use heatsinks or bolt then to a metal enclosure make sure there isolated otherwise you will short them together. The positive regulator can be bolted directly to the case as long as your case is ground.

[MohiZ]

Ok, I still can't get this thing to work. I'm using this PSU to power an LXH2 Marshall Amp&Cab sim/preamp which consumes about 130mA at +/-9V. When I plug the PSU into the preamp, the +9V regulator's output voltage becomes slightly negative (!) and the -9V regulator's output fluctuates between -3 and -9 volts at a frequency of over 100kHz.

I'm starting to think it's just not possible to power this thing up. Let it be known that I did get it to work with a single power supply of 18 volts, using an op-amp generated virtual ground at 9V. This was a bit noisy so I thought I'd build a proper dual power supply. Should I just give up? 

It's totally possible that I have a stupid mistake somewhere but I just don't know where to start looking.

[head_spaz]

With a single 12 volt AC supply, you're gonna have problems if you try to make a common ground between the two regs. Separate the grounds and see if each circuit works independantly. Of course separating the grounds will render it useless to you... but it will verify the point.
That's why the virtual ground setup worked, and this one does not.

Virtual ground will work, but you'll end up with a higher impedance ground, which is probably where your noise is coming from.
You'd be better off using either two separate windings, or a center tapped winding. Or just use two wall warts.

[MohiZ]

Is there any way to make this work with just the one AC wall wart? I've seen supply designs that use four diodes as a bridge rectifier, but they seem to use a center-tapped transformer.

[head_spaz]

Is there any way to make this work with just the one AC wall wart? I've seen supply designs that use four diodes as a bridge rectifier, but they seem to use a center-tapped transformer.

See for yourself...

[slacker]

That power supply should work fine with a 12 volt AC wallwart, I've built something very similar.

One of the AC adapter's leads is labeled (-) so I connected that to ground.

Stupid question but are you sure it's an AC wallwart. AC as no polarity so it's a bit strange that one of the leads is labelled as negative.

[Gus]

As posted above make sure you know the pinout of the regulators. 

You will need isolated thermal pads or wafers and thermal compound and isolated mounting kits.  Make sure the heatsink(s) is large enough.

If you power the supply up you need a small load on the regs to let them regulate correctly about 10ma should be good.  9VDC/.01amps= 900 ohms, DC  power = E x I =.09watts.  load with a 1K on each side.

[MohiZ]

Stupid question but are you sure it's an AC wallwart. AC as no polarity so it's a bit strange that one of the leads is labelled as negative.

Yes, that one I'm sure about. Even checked it with a scope. Nevertheless, the other lead is labeled. I was thinking maybe that lead stays at a constant voltage while the other fluctuates, but then again I don't really know that much about power supplies.

About the heatsinks.... the regulators don't seem to be getting too hot right now.

So if I conclude that I can't make this supply work, is there ANY other way to create a bipolar supply out of a 12V AC wall wart, or an 18 volt DC wall wart. I've tried a voltage multiplier chip but it fell short in terms of current sourcing capabilities.

I know there exists a lot of info on this, but seeing that I've tried quite a few things this is getting frustrating. Thanks for your patience! 

[MohiZ]

N E 1?  What I'm looking for is another power supply design, one that can source at least 150mA of current.

[Tubebass]

The GGG circuit will work if built as shown. That means you'll need to ditch the bridge rectifiers and go back to single diodes.

[MohiZ]

That's what I'm using right now, but it still doesn't work. Only with two wall warts can I make the preamp work properly. I wonder if the GGG power supply can supply enough current... 

[Nasse]

Perhaps you downloaded the data sheets but did not read those? Perhaps your wall warts are giving too much voltage at idle and regulators lose most of their capacity putting that down. You have something connected wrong and perhaps too much input voltage. The circuit works if done properly.

I have a simple working vero layout for this, (look mother, no drilling needed) a bit dizzy photo somewhere...

[MohiZ]

The high input voltage might be a viable explanation to this, although I'm using a wall wart that is recommended in the GGG project file for a +-9V power supply (12VAC unregulated). After rectification, it's about 40V difference between the - rail and + rail. That's quite a lot for the 9V regulators to handle.

What I meant with the two wall wart setup was that the preamp I'm trying to power up works when I just use two 9V DC wall warts connected together in a way that provides +9V and -9V voltages. It's just a bit clumsy and it would be better if I could get it to work with a single wall wart.