Fourier modulation revealed

[John K5PRO]

I finally found time to research Broadcast Electronic's patented 4MX transmitter. Pretty clever, patent #7,092,26. Take a conventional class D H bridge power amplifier. Normally these things are set up so that driving the diagonal legs together, and alternating this at the RF frequency with the other pair of legs, forces current back and forth through a bridging transformer (the center bar in the H). What Jerry Westberg did was to now vary the phase relationship of the driving pulses to each transistor in a particular pair of diagonal legs. What i mean is that during any half cycle, the current through the transformer can be varied, by shifting the phase of BOTH drives to the right half of the H, for example. Looking at the H, the vertical pairs of transistors are still always driven 180 deg opposite, with a dead band to prevent shoot through. At zero carrier (-100%) the left top and right bottom legs are in phase so that there is no current through the primary. At >+100% peak, the phase is opposite, so maximum voltage is across the primary. And then it all reverses on the alternate half cycle of RF. Voila, the transmitter is all class D hard driven switches, no high level modulator needed, and it is truly an alternative design to the patented DX transmitter of Harris. +400 volts is used and in the patent, APT5010JFLL Fets are mentioned. It is essentially RF PWM. I didn't mentioned the inductors and capacitors required to keep current flowing in the bridge and to raise the efficiency.  Most of the work is done via DSP to generate the phase shifted RF square waves for the H bridges, at the incoming audio rate.

Now I understand how they packed a 50 kW rig into < 1000 lbs that could fit through a door.

[WA1GFZ]

H bridge running above the broadcast band would be quite the machine.
I think MOT had a push pull app note that did the same type of modulation but they varied the pulse width rather than the phase between drive signals.

[kb3ouk]

This caught my attention since I was interested in how they were modulating newer broadcast rigs, and getting the high efficiency they do. So it's basically a variation of outphasing modulation? Wonder how easy it would be to reproduce, but without the DSP?

[WA1GFZ]

Harris DX 50 does not need DSP. It selects the number of amplifiers to toggle on each zero cross of carrier.

[kb3ouk]

So it modulates by turning on or off the amplifiers based on what the modulation needs to be?

[WA1GFZ]

Yes A/D samples audio at carrier frequency sample rate and digital result goes into a rom address input. Data output goes to decoders to select the required number of amps to duplicate the level. All selection is made on the carrier zero cross so there is no distortion. Very simple, you just need 128 amplifiers into a combiner.
DX 50 used IRF 350s or IRF450 (IIR) in a bridge configuration for each amp

[kb3ouk]

So with no modulator it should've been pretty efficient, because each amplifier would be running wide open, right?What voltage would they have used on them? IRF350s are only 400v 15A devices, and IRF450s are 500v 13A. Both are in a TO-204AA package

[DMOD]

I seem to recall the B+ was about 200Volts, so each transistor in the H briidge was seeing about 100Volts.

I suspect they used a Flash Converter with dithering to do the analog to digital conversion.

Phil - AC0OB

[DMOD]

I seem to recall the B+ was about 200Volts, so each transistor in the H briidge was seeing about 100Volts plus any voltages due to inductive kicks or ringing.

I suspect they used a Flash Converter with dithering to do the analog to digital conversion.

Phil - AC0OB

[WA1GFZ]

Each amp was either full on or full off. Three phase was rectified to get DC. The ac line to RF out spec was about 85%. I don't think there was a power transformer. The only way to improve efficiency would be to go to active rectification. 250KW through a three phase bridge having a volt or two drop is a lot of juice. I think there were some 1/2 bridges to get finer resolution.
Distortion was claimed to be better than 1%. Very interesting how the combiner worked. All "off" modules were set to short their combiner input inductor to kill leakage inductance from an undriven input.

[kb3ouk]

That's a lower voltage than I expected. I know that there is a 50 kw Nautel transmitter that uses APT5010JFLL's for amplifiers and those are a 500v 41A device and they are running 400v on them in an H bridge. That seemed to be pushing it from my point of view.

[WA1GFZ]

I think the DX50 came out in the '80s before the high performance FETs hit the street.

[kb3ouk]

First one was built in 1988, went to KFBK, which was also where the predecessor to the first Ampliphase was built.