I've had some experience driving quadrupole ion filters...

I would guess you would do best with a high Q tank circuit driving the
ion trap through as short a length of wire as possible. Keep the tank
and the wiring very well balanced with respect to stray capacitance to
ground, especially. You can use a link to couple RF energy into the
tank circuit. The capacitance of the trap is probably enough to
resonate the tank coil. You can split the coil into two halves, with
the "hot" end of each half going to the ion trap, and the "cold" ends
connected together with a bypass capacitor; then you can apply DC
across that bypass capacitor, and the DC will be carried directly to
the ion trap. I am assuming here that you want to drive your trap in
the normal way, with a combination of RF and DC. It is usual to use
feedback to maintain an accurate RF voltage. You can add small
capacitors, one from each "hot" lead to the trap, back to a rectifier
with low input impedance. The capacitor impedance then sets the
voltage to current conversion ratio, and will be frequency dependent of
course.

If you use a resonant tank circuit with a loaded Q (nearly equal to the
unloaded Q) of, say, 250, and you have net capacitance (ion trap plus
wiring plus coil plus voltage sampling capacitors) of, say, 32pF, and
the operating frequency is 1MHz, then the tank impedance across the
trap will be about 250*5k, or 1.25megohms. At 500V RMS, that's only
0.2 watts. It appears that your signal generator can put out 1 watt,
so it's quite possible that your signal generator can drive the tank
without an amplifier. You can experiment with links (one turn, or a
very few turns, over the area of the tank where it is split in two) to
see what gives best results. You can vary the coupling by adding a
variable capacitor in series with the link to the generator. Proper
adjustment will give you a 50 ohm load for the generator, at the tank's
resonant frequency. Winding a modest size tank coil with such a high Q
at 1MHz is not trivial, but certainly possible.

With such a resonant tank circuit, which is useful both in getting high
voltage and in reducing harmonics in the drive, it will be difficult to
operate over such a wide frequency range. Do you really need to do
that? If you do, I would advise you to add a medium-power amplifier
between the signal generator and the step-up transformer, and beware
about a lack of harmonic filtering as a result of your broad-band
circuit.

Further details about exactly what you are trying to accomplish might
be helpful.

Cheers,
Tom

Georg Holderied wrote:
I need to build a transformer that transforms the output of a signal
generator (0.5-2 Mhz 0-20Vpp into 50 ohms) into two opposite phases
at 0-250V.

Load is a few picofarads and a short length of shielded cable.

Where can I find formulas for winding such a transformer ?

Is it preferable to build it as an air or on a ferrite core ?

What ferrite material would I use ?

BTW the signal is used for driving multipole ion traps.

I am an electrical engineer but have not much experience with RF.

George