So, I recently got my hands on a few MAX038 ICs (function-generator on a chip), and they require +/- 5Vdc in order to operate.

So far, I've put together a 555-based charge-pump power supply (based on this). I take the -V output and put it through an LM7905 to get ~-5Vdc, along-side a LM7805 to get +5Vdc. This sort've works but has the following issues:

• I need to drive the circuit at about +13Vdc to get -5V out of the 7905 regulator. At this point, the 7805 (which receives the full +13Vdc) is starting to get fairly warm, which is not ideal. (P.S. the MAX038 seems to draw ~60-80mA -- the datasheet suggests it draws 35-45mA on +V and 45-55mA on -V)

• The output voltages are "messy" - According to my scope, I'm seeing fluctions of over 100mVp-p, which seems to be a bad thing for the MAX038. Putting smallish (Mylar 100nF) caps between the regulators outputs and GND helps a bit, but it's not perfect.

The net result of this imperfect power-supply is that I get a really irregular sine-wave out of the MAX038 (the output of the MAX038 is all over the place, creeping up and down by several kHz, maybe by +-5% of the expected frequency). I suspect there are a few reasons for this (it's built on a breadboard, so parasitic capacitance probably isn't helping) but I primarily suspect the poor power-supply.

So I'm wondering if anyone has any suggestions for a better way to get a stable +/-5Vdc power-supply? Any particular power-supplies (i.e. kits) that have been found to have a very clean DC output? Currently I'm driving the 555-based power-supply from a bench-top power supply, but would be quite happy to build/buy something that plugs straight into the wall, so the bench-top power supply is freed up for other uses.

Take a look at this.

-Raja

Also take a look at this.

-Raja

Thanks Raja. The ST735 looks like a good option - probably a bit of a more robust solution than I've come up with so far. Might be a bit interesting to source, however!

OK, so I chose a different route and built up a +/-5V powersupply using a 9-0-9V toroidal transformer. I joined one side of each coil to give a common ground, and passed the other side of each coil through a rectifier (well, 4 x 1N4001 diodes that were lying around, but the effect is the same).

Unfortunately, this gives approx. +/- 13.5V DC output, which wastes energy (heat!) from the 7805 and 7905 regulators, which I don't think is ideal. Having said that, after ~10 minutes running they're only luke-warm to touch so I guess it's not a big issue. I didn't think to measure how much current the circuit is drawing - probably the same 80-90mA as before, so I guess each regulator is dissipating approx. 13.5 - 5 * .045 = .3825W or 382mW of heat, for this circuit.

I didn't think to take any pictures of the power-supply circuit board, but here's a pic of the power supply almost fully assembled (I went crazy with the glue gun :)

And the finished product

Here's a picture of the power supply running the basic, bread-boarded MAX038 circuit at just over 6MHz. Note, I found that I had to put additional 1000uF caps between GND and the +/-5V rails, as the 470uF caps internal to the power-supply don't seem to smooth the voltage out enough (producing some really nasty distortions on the MAX038 output).

And running over 22MHz - the output starts to get pretty distorted at this point (as the MAX038 datasheet suggests it will).

Next up will be figuring out how I can start controlling at least some parts of the circuit using an ATmega. Sine/Square/Triangle output is selected by setting pins A0, A1 high or low, so that's easy enough to do with a micro-controller.. I can probably come up with a way to provide a voltage between -2.35 - 2.35V to control the duty-cycle adjustment.

I think I will have to use a physical switch to select the value of the capacitor that the MAX038 continuously charges / uncharges, which is connected between the "COSC" pin and GND.