[Amps] FCC Denies Expert Linears' Request for Waiver of 15 dB Rule

[Manfred Mornhinweg]

Hi Jim, and all,

> When you see an SDR receiver selling for a few hundred bucks, it's
> pretty simple to add a low power transmitter.

> So adding a high gain amplifier makes for a cheap SDR radio capable
> of medium to high power a fairly cheap option.

Exactly!

At this moment I'm VERY tempted to brew a concoction from these ingredients:

- A Red Pitaya board, for around $300;

- A homemade board containing a low noise preamplifier, attenuator, and 
band pass filters for RX, along with a 10W or so driver amplifier for 
TX. The whole board should't cost more than $150 in parts;

- A high efficiency, high gain, deliberately non-linear, legal-limit 
final stage. This might take the shape of a simple overdriven broadband 
  class AB stage, or perhaps a class E stage with tuned, band-switched 
output tanks. It might use a single BLF188XR for about $150, or several 
small switchmode MOSFETs for far less money.

To these hardware ingredients I would add some freely available 
software, such as Pavel Denim's software for the Red Pitaya, plus 
PowerSDR, using closed-loop, RF-feedback amplitude and phase 
predistortion to linearize the amplifier. The software has this already 
implemented.

The result would be a legal-limit, high efficiency, high performance, 
compact, black-box HF transceiver to be used with a PC, for under $1000 
total cost, if you are good at ordering surplus power supplies on eBay, 
and such.

And those who prefer integrated transceivers, can of course add an 
embedded PC motherboard plus small monitor, for a few hundred dollars.

It would just take some tinkering with the final stage. The rest is 
pretty simple to do.

Some more development in the RF power device scenery would be welcome. 
The BLF188XR has the necessary capability at least for the saturated 
class AB version of this idea, in terms of gain, voltage, current and 
thermal characteristics), but it's quite hard (or more bluntly: close to 
impossible) to achieve the desirable 85% or higher efficiency on a 
broadband basis, due to the very low drain impedance and the ensuing 
trouble in properly coupling the drains together and to the output 
filters. Switchmode MOSFETs instead are available with voltage ratings 
that allow operation at convenient impedance levels, but they have less 
gain, higher capacitances, more variation between 1.8 and 30MHz, and 
require the use of several (or many) in parallel. It would be great to 
have true RF power devices that can operate from a few hundred volts, 
and have low enough output capacitance to allow broadband use to 30MHz. 
Along with good gain and convenient price, of course...

As usual, I'm asking too much.  But I'm confident enough in this concept 
that I intend to buy a Red Pitaya board and start experimenting in that 
direction, using a class-E output stage based on ARF449 MOSFETs, which I 
happen to have in my "selection junk box". That would make a few hundred 
watts only, but serve as proof of concept. I want to test, hands-on, if 
it is really possible to achieve adequate spectral purity by linearizing 
a deliberately non-linear amp via SDR. Normally the predistortion system 
included in SDR software is intended to make a run-of-the-mill linear 
amplifier extra clean, and not to make a dirty-and-ugly but highly 
efficient amplifier clean enough!

Manfred


========================
Visit my hobby homepage!
http://ludens.cl
========================