Switching dual supply rails with a single switch

[SirNickity]

I'm just getting started with designing and building circuits. One of my first projects is a CMoy headphone amp. The amp itself is simple enough, and I'm completely comfortable with it. But, I have a single pole power switch I'd like to use (for aesthetic reasons), and I'm stuck on how I can use this to switch both the + and - rails.

My topology and self-imposed design constraints:

- There will be a small toroidal transformer providing 9-0-9 AC. This will be rectified and feed 7806 and 7906 linear regulators. (+/-6 because I already have a pair of them and when will I ever need 6V?)

- I could just switch the transformer primary, but the switch will be on the front panel next to the volume knob, and I'd like to keep the mains and audio as far apart as possible.

- The actual circuit of course is pretty much an OPA2134 and an LED, plus the loss in the regulators. 100mA is probably a generous target. The whole thing could just run continuously, but as both a learning exercise and as good form, I'd like to only power the load while in use. Preferably, switching before the regulators.

- I don't want to use relays. It's going in a small case (of course), and I'd like to use this as an opportunity to learn more about transistors.

Ideally, I would have the rectified +9 run through the switch, and have it saturate a pair of MOSFETs -- one for each rail. Something like this, where the 1k resistors represent a load:



I have a couple spare MTP3055s (NPN) on-hand, so switching the positive rail is easy. What I don't know is how best I could use that same +9 to turn on the negative rail.

In the Falstad simulator (from which the image was taken), current runs through the negative MOSFET regardless of the switch. I confess, I don't really know why -- maybe just a fluke in the simulation of ground?

I'm sure this is a really basic problem that has been covered a million times before. I just don't know where to look. I could use an opamp as a comparator, but it seems a bit foolish to use an opamp to switch on another opamp! There's gotta be an easier way...

I appreciate any ideas or suggestions.

[sofaspud]

I'll make no claims regarding this circuit (I'm pretty much noob myself) but offer it to try in your simulator software. I got about +/-5.25V across the load. Pay no attention to the device numbers - I just grabbed from the library.

[preiter]

If you don't switch the transformer main, then you have live mains-level power in your case as long as this thing is plugged in. This could be a safety concern.

[raudio1969]

Try this circuit.

[SirNickity]

Thanks for the replies, everyone.

I've tried putting SofaSpud's idea in my simulator, and I can't get it to work right -- yet. I'm not sure if I'm doing something wrong, or running into bugs or limitations in the simulator. (It's designed as an educational tool to tinker with -- not a design tool -- and I've found at least one confirmed bug so far.) I'm still plugging away at it, and getting a feel for how to use the various parts (N- vs P-channel, etc.) along the way. If anyone has suggestions for free or cheap sims that are (relatively) friendly to beginners, I'd love to hear about them.

I'll try the second idea, too. I had a fleeting thought to try switching the ground, but had not quite come up with the hows and wheres yet. Thanks for that.

@preiter: Concerns duly noted. Even if I do switch the incoming AC, I have live mains all the way to the switch at least -- which will be on the opposite end of the case from the inlet. Instead, if I have the mains side of the transformer unswitched, but completely within the rear few inches, I can shield it with a box-within-a-box to keep that portion isolated -- for both noise control and safety.

[sofaspud]

I can pass along that I used CircuitMaker6 (an old obsolete program, but easy to use and probably available somewhere on the web) and it would not simulate when I put an open switch on the 9V line.

[raudio1969]

There is a free version of Tina SPICE-Based Analog Simulation Program available on the TI website.

SPICE-Based Analog Simulation Program - TINA-TI - TI Tool Folder

[Elvee]

With two small, sensitive gate triacs, you can switch at the AC level.

You could use the same trick after the rectifier with NPN and PNP transistors.

[SirNickity]

I think raudio's design will do the trick. I need to order some P-channel MOSFETs before I can breadboard it, but the simulation is working great. (I've got to get up to speed on the SPICE modeler, so I'm still using the Falstad sim for now.)

I'm not sure what VF is supposed to stand for, so I'm guessing R5 and R6 are still representative of the load, and VF1 and 2 are probe points? Is that right, or do R5 and R6 serve some other subtle purpose, while VF1 and 2 are outputs? The rest of it makes perfect sense to me.

[KatieandDad]

Originally Posted by Elvee

With two small, sensitive gate triacs, you can switch at the AC level.

You could use the same trick after the rectifier with NPN and PNP transistors.

There is no gate drive current in this circuit ?