New way to make AM waves

[N9NEO]

There is a bunch of information at a yahoo group called phaseX. The site has just been made public for anyone who might be interested. Some patent issues kept site from being made wide open before now. I think you can just log onto site. If not just get hold of Art - K3XF and he will tell you how to locate site.

A rig was built that does 160m and modulates by using a phase shifting technique. The RF deck and modulator have been combined into one stage.

I have gotten a 'best I ever heard' audio report while talking across town here on 1w using a D104 wide open. Very poor audio report on 4 watts of carrier from WA2CYT & Frank a few weeks ago. I expect the problem has been taken care of with a grunge plane and some minor tweaking. Did get a few hundred watts out of thing before I finally smoked it. A wire fell off the mosfet and left the gate hanging in the breeze.

Anyway, this is just FYI, and very experimental. I have left some readme files in the neomod folder in the files section. There is also a pic in the photo section. There's actually enough information to build one of these things including artworks - but you are on your own. You can build for your own ham radio use if you want.

NEO-later
Bob

[Rob K2CU]

This is not really a "new" technology. It was originally called Ampliphase and was invented in 1935. See :

http://www.rossrevenge.co.uk/tx/ampli.htm

It wouod seem that this technique might be suitable for Class E transmitters. Simply put, you would just use two Class E bricks and a combiner, as  might be done to achieve hgiher power in the normal sense. But, instead of PWM modulating the high voltage power supply, you run the transmitters at full bore, but phase modulate the input drives in opposite directions (+/45 degrees) instead.  You need wide band circuitry to ensure that all the PM sidebands can cancel.  The big issue has to do with how the combiner reflects the adding and opposing RF voltages back to the two Class E bricks.  Might just go "PZZZST".....

The above article and others discuss the non linearity between the phase shift and output amplitude.  There was also the need for multiple pahse shift stages to accomplish the desired amount of phase shift.

For anyone interested, you can use a PLL to make a phase mdulator as I did some 30 years ago in a homebrew 2 mtr FM transmitter.  A PLL is a closed loop system that locks a voltage copntrolled oscillator onto an incoming signal, usually at 90 degrees. Any shift in the pahes of the input signal produces a correspoding error volatge form the phase detector, and that is used to alter the phase of the VCO to trask the input signal.  Well, if you introduce an additional voltage (audio) to the error voltage, the VCO will be shifted in phase to an amount needed to conteract the added error voltage.  

So, you start with a carrier oscillator, feed it to two PLLs.  In the above article, they mention having the two drive signals at 135 degrees with no modulation.  Youo could use a simple pair of RC and RL networks to make the two PLL inputs be out of phase by 135 degrees, or even use a  4x carrier clock and dual flipflop circuit to get 90 for starters and add a DC bias along with the audio to set the resting phase to 135 ( +17.5  on one and -17.4 on the other).  Audio would then be added to swing the +/- 45 degrees with  the two audio signals at 180 degrees.  

You still have to deal with the non linearity of the phase combining technique.  And, there is alos the issu of what happens with over modulation....well, at least there is no flat topping or cut off splatter.

[N9NEO]

Hi Rob,

Yes, using phase shift to do the modulation seems to have been around for quite some time now.   Not sure if there are any transmitters in existance still using the outphaseing technique.

I don't think class E can be used to do the phase shift modulation.  Success with Class E switching depends to a large degree on the load impedance remaining constant.  As the phase is moving about, the load impedance can theoretically move from infinity when the poles are directly in phase, to some low value when the poles are directly out of phase.   I don't think you can maintain turning on into both zero voltage and zero current with the phase shifting going on.  I have heard claims of somebody knowing how to do it class E, but haven't heard the details yet.

The modulator I put together uses a pair of mosfets and a comparator to develop the phase shift.   It is very simple and no need for any esoteric pll or dsp.  Phase A is the master and phase B is just slaved to that.   In this way only one phase needs to move relative to the other.

The old ampliphase system used sinewaves where the new way to make AM waves uses squarewaves.  The old sinewave method I have been told has some very peculiar transfer function between phase angle and output power - Some hoops to jump through in order to get good sound. The new way to make AM waves has basically an exponential relationship between the two.  I've exploited this fact with an exponential ramp (a simple cap & resistor) to generate the phase delays.  As I said, FB audio reports, so I'm sure the ramp is doing what it's supposed to.  A feedback loop would be an improvement, but for the relatively low powers encountered with ham radio it probably isn't worth the effort.

The output section is actually quite simple as well.  Don't need two complete tank circuits and a combiner in order to get the waves.  Just one tank circuit, and you can actually put a secondary winding on the tank coil core and couple that to the antenna - through an appropriate filter.  Wire the series tank circuit between the  two poles and away ya go.

The difficulties with the design are getting the mosfets on and off in the 10ns - 20ns timeframe.  Using squarewave drive with plus/minus 10v, and a few amps of gate drive current to slam them fast.  I've gotten a few hundred watts out.  My present efforts are to do full pep output with 4 Fairchild FQA16N25C Qfets.

I am hopefull that with some emerging fet technologies multi-kilowatt transmitters using this strategy will become popular.  www.cree.com

73
NEO

[NE4AM]

The new Broadcast Electroncs AM broadcast transmitters use a variant of outphasing to modulate their power amplifiers.  Ad Bob pointed out, the problem indeed lies in maintaining efficient operation of the power amplifiers under all modulation phase angles.  BE's transmitter uses Class-D-E power amplifiers, and some cool fiddle-jiggery to keep the MOSFETs happy.

The circuit I've developed is an improvement on BE's - with more efficient operation (1.5 KW from four TO-220 devices, @ 97% efficiency) and has a linear phase-modulation-to-envelope-modulation function, and works with either Class-E or Class-D-E power amps.  Details to follow.....

[John K5PRO]

5 years ago I built an extremely well behaved phase shifter which worked up to dozens of MHz. I was thinking about using it in a class E/F mode outphasing transmitter, very similar to the ideas here. The phase shifter could go 360 degrees with a control voltage (or audio) and a pair of them would easily drive the "amplifuzz". However, it was an all analog implementation based on the Analog Devices AD639 trig function chip to compute a sine of the control voltage and a pair of their high speed AD834 multipliers, in a standard sideband modulator design. It would be much wiser to do it digitally today.

Chiriex outphasing modulation has experienced a revival in the wireless comm business. There has been a lot of effort to take the power wasted and feed it back into the power supply to boost efficiency. There are lots of papers on this at various IEEE conferences in the past 5 years.

[Ian VK3KRI]

This is not really a "new" technology. It was originally called Ampliphase and was invented in 1935.

I believe ithe 'Ampliphase' term was coined by RCA. In the older references I have it is called Chiriex system. It was actually deveolped before the classic Class B high level modulation. The out of phase sine waves were generated by Chireix with some sneaky stuff with a low level AM signal and  another quadrature carrier This gave the transfer function directly.

I think class E migh be a bit dodgy due to the changing impedance seen each PA as the phase difference changes

                                                                      Ian VK3KRI