[Amps] Harris DX transmitters

[John Lyles]

The Harris DX design, first released in the DX10 in the late 1980s, used 
inexpensive plastic switching FETs and was essentially a giant digital 
to analog converter. The audio is first converted into digital, some DSP 
magic applied, and then a whole lot of series connected modules generate 
the carrier directly by turning on and off to create steps. More steps 
are on for the positive peaks of carrier frequency and so forth. 
Filtering into RF waveform is in the output. The audio is applied to the 
gating of these switches, so as to create amplitude modulation, not just 
steady carrier. Somewhere I had papers on this;
A basic 1500 watt block had 8 FETS, so at 200 kW they used about 1800 
FETS. Combining five of these they made a 1 MW rig and combined two of 
these for 2 MW. But the bread and butter design was the 50 kW DX50 which 
has morphed into a more recent version as others stated here.
Nearly 100 MW of long wave and medium wave RF power has been supplied 
with DX-type transmitters by Harris, according to Tom Yingst, former CEO 
of the division. However, this technology has not been extended up into 
HF range, as the transistors are not capable of switching at those 
rates, at least not economically.

Harris has the honor of owning the DX design; their competitors use PDM 
in their SOTA MW transmitters.
John
K5PRO


On 12/18/16 12:28 AM, amps-request at contesting.com wrote:
> Message: 2 Date: Sat, 17 Dec 2016 22:57:44 +0000 From: Steve Thompson

 > I believe, in transmitters of this type, that the modules aren't 
combined in the same way we're used
> to in the likes of Quadra, PW-1 or SPE type amps - in the AM txs I think it's more like the module outputs are connected in series akin to
> stacking PSUs to get a higher voltage. Steve

 >> That is becoming common practice. Modular transmitters. Just
>> plug in the modules you want to get the power you want and I
>> believe they have been hot swappable for some time. .
>>
>> I had never thought of switching the modules in and out, but at
>> 250 W per step it'd be a pretty smooth transition, albeit they
>> may do some filtering. I wonder about switching pulses?

>> 73
>> Roger (K8RI)