Cheap, simple bi-polar supply from common parts (DC->DC)

[ExpAnonColin]

CORRECTION-When I said 7660 I meant 1044 :oops:

Thanks to Hal for trying out the first circuit (I've misplaced my 386's )

It's DC to DC and provides +/- 9v with not a whole lot of current drain.  Does anyone want to construct the whole thing and test it out for me?  If so, could you get some mA measurements, so we can compare to other methods?  I might get around to a PCB layout soon.  I've just ordered some 386's so I can check it out in it's entirety myself.

I didn't include schematics for the first 3 voltage doublers because they're commonly available on the net:
1) http://geofex.com/circuits/voltpmp2.gif
2) http://www.kronjaeger.com/hv/hv/src/mul/greinacher.png
3) http://ww1.microchip.com/downloads/en/DeviceDoc/21453b.pdf

-Colin

[R.G.]

The LM386 idea does definitely work - I posted it here quite a ways back as a way to get an automagic bias voltage. Check the archives. It's good up to more ma than any effect has a reason to need.

The 555 used as a doubler is OK, but it suffers from the other problems that a 555 has - mainly current spiking on the power supply line when the output is changing. It's really hard to keep those spikes from coupling to other circuits on the +9V line.

A possibly better and cheaper alternative is a CD4049 hex inverter set up to run two sections as an oscillator and the other four as inverting buffers to give you a 9V high current square wave that can be attached to the same diode/cap network as the 555 for -9V.

Another possibility that I posted here was to use a MOSFET gate buffer driven by an oscillator to make complementary +/- square waves that are coupled through capacitors to schottky rectifiers to give an *isolated* 9V supply of either sex. I have the TI MOSFET driver chips in the garage, just haven't had a chance to hook them up.

Any source of AC can be run through a capacitor/inverting doubler to give -X volts.

You might want to investigate Schottky diodes to keep from losing so much voltage to silicon diode junction drops.

[Brian Marshall]

arent there some opamps that self bias to half of the supply.  ive never used one, but thought they were available.  they make their own 0v internally.  

I dont know.  i have a bunch of 386's and have never even used them yet.

[ExpAnonColin]

I dont know.  i have a bunch of 386's and have never even used them yet.

Then here's a perfect use... if you need any 555's, I managed to get about 50 of them in a misc. lot.  I can pop one in the mail for you if youwant.

-Colin

[Hal]

I hate to be the one to ask this...but what schematics need bipolar voltage?

[ExpAnonColin]

I hate to be the one to ask this...but what schematics need bipolar voltage?

Well, it gives you more headroom in op amps, allows you to properly power AD633JN's, 3005's, and 1496's, all of which I use pretty often...  I just got sick of not having enough volts.

[petemoore]

I've been using the 386 method for numerous Opamp builds that require 1/2 voltage.
 I find this way simpler and  easier to wire, and quite dependably get very close to 1/2 Supply with them.
 At RS at least, the 386 is less expensive than buying the caps to do conventional Vb configurations.