# +/- 60V Sawtooth/Triangle/Sine Wave

I'm trying to create a power supply for biasing a Langmuir probe. It needs to run some sort of predictable wave (I need to compute an I-V curve for a plasma, so I need to vary the voltage in a predictable way) at about 1000 Hz. It needs to be +/- 60 V. I don't really have time to order any new major parts so I've been wondering if there's a purely-analog solution to this problem. A friend recommended using a capacitor circuit to generate a triangle wave but this presents the problems of:

1. Having a 120 VDC supply (perhaps daisychaining 9 V batteries), and
2. Biasing such a wave -60VDC (it should run at +/- 60 V, not 0-120 V).

Anyone have any ideas?

• and the output current will be ? Mar 7, 2014 at 19:16
• Quick : lab signal generator + high power audio amplifier (300Wrms into 8 ohms) gives you approx the voltage you need)
– user16324
Mar 7, 2014 at 19:35

The current for such a probe seems to be in the low mA at most. I'll assume that you have a suitable function generator that you could just transform up perhaps 5:1.

A good quality 24V:120V mains frequency transformer may do the trick, or you could wind one on a pot core or ferrite or toroid core (perhaps salvaged out of a SMPS).

If the function generator can supply 8.5V RMS into 50 Ohms, that would be the equivalent of +/-60V peak into 1250 ($50\cdot 5^2$) ohms or 34mA RMS.

You will likely get some rounding of the triangle wave with a mains transformer, especially at 1kHz, since the first harmonic is at 3kHz, but perhaps that doesn't matter or perhaps you could do your tests at a lower frequency.

If not, the ferrite-core transformer should do the trick. If your function generator can't supply the required current, as Brian Drummond commented, an audio amplifier would make up the difference easily, even without a transformer if it's a high power one.

I think the hardest part is to generate the high voltage so I'm focussing on this first. Below is a circuit showing how to connect an op-amp to a bunch of transistors. This can deliver over 200 volts peak-to-peak: -

It has a voltage gain defined by the 100kohm (bottom left in // with a 20pF capacitor) and the 10kohm input resistor; Gain = 10. Note the use of positive and negative voltage supplies. These can be lowered to something like + and - 70V to suit the output voltage required of 120Vp-p. If using batteries then half way up the stack of batteries is ground.

To generate a triangular wave of 12Vp-p is relatively easy using an op-amp fed from +/-15V rails (like the one above) and configured as an RC relaxation oscillator - it won't be perfectly linear AND if this is a definite must a 555 and constant current transistor can achieve this. Show below are a few different ideas that should all prove OK: -

If the current needed by the output was fairly low you could create the +/-70V supplies using a boost regulator chip attached to a 12V battery.