Re: VLF: DK7FC's "big" VLF PA

[Andy Talbot <andy.g4jnt@gmail.com>]

Or you could use opto couplers in the input.   There are plenty with built in logic level converters

 

Andy

www.g4jnt.com

http://www.g4jnt.com/137tx.pdf

2011/1/3 Stefan Schäfer <Stefan.Schaefer@iup.uni-heidelberg.de>

Dear group,

Since a long time i wanted to publish the circuit of my "big" VLF PA. It is titled as a 600W PA since it has never handled more power so far, due to the power limit of my /p generator. In fact i think it will handle at least 1 kW, probably even 2 kW. As announced i will try a 2 kW generator in my next VLF experiment and i am not afraid to use this PA for the first test.

The circuit is nothing new or very special. The mains voltage is rectified and stabilized by a 2.2 mF capacitor. A H bridge using 4x IRFP 460 FETs is switching at the desired frequency. To achieve an accurate 50% duty cicle, a CMOS4013 is used in front of the MOSFET driver. It is switched as a frequency divider (/2), thus the input frequency has to be 2x the actually wanted frequency (17.94 kHz > 8.97 kHz). The driver(s) are 2x IR2110. The operating frequency is about 0.1...20 kHz (320 VDC) or 0.1...150 kHz (100 VDC). It is designed to work properly below 9 kHz. The input waveform has to be a 12Vp rectangular waveform, ideally. The waveform is not very critical, e.g. a sinusoidal or triangular waveform is no problem as well.
A small transformer is used to generate a local 15V supply for the drivers, ICs and LEDs.
There are several methods how the input signal can be applied. Either directly (please use an audio transformer to decouple the PC from the circuits ground which has mains potential!), as shown in the circuit or via a fiber optic cable, as i do it. The fiber optic solution is a special and (for some) strange method so i prefer to show the usual way to do it...

If the mains is used to supply the PA (much lower output impedance), a resistor has to be used to charge the 2.2 mF capacitor. A 100 W/230V lamp might be used. If the capacitor is charged, this resistor should be bypassed by a relay.

One has to take care about the output: There is no isolating transformer used so the output has main potential! Thus, a galvanic isolated coupling winding has to be used for applying the RF to the loading coil. No ferrite output transformer is needed!

This PA is still using the first pair of FETs, there was no fault so far. It was received in several countries that it is no problem to transmit a one hour carrier at P > 500 W. Here, the heat sink gets not even hand warm! So the efficiency can be assumed to be almost 100%.
The design can handle pretty much reactive power. It was no problem to drive an antenna that is totally out of resonancy (kite lying on the ground or wire suddenly disconnected due to strong pull).

The actual (active) output power can be regulated continuously by adjusting the resonance.

So, if one of the new VLF TX stations want to change from an audio PA to a high power switching mode PA, this well proven design may give some ideas. Just ask if something is not understood.

See once again one of my very amateur like ugly hand drawings ;-) at http://www.iup.uni-heidelberg.de/schaefer_vlf/pic/DK7FC_600W_VLF_PA.pdf

Vy 73, Stefan/DK7FC

PS: Once again i explicit want to say that one has to take care about the parts that are on mains potential! If you are not sure about the handling, add an isolating transformer at the 230V AC input!