It's all about the Software

Hi Simon,

Now that you've done all the easy stuff (*grin*), how about Pure Signal?

Regards,
Steve, KW7F

Support for this request from me, too.

Thanks again for the s/ware!

Regards,

Charles.

Do you have the hardware to support it? It requires an unusually good directional coupler and a suitable attenuator, at least for something in the hands of hams.

{o.o}

I support this as well- it would be great to have Pure Signal available. However, there is some further work needed to tidy up TX issues for Anan. Several solid state amps now have an attenuated output suitable for Pure Signal and you can also make or buy a directional coupler.
73 Dave G3ZQH

Has anybody taken a pure signal output and applied an automatic antenna tuner to it and a spectrum analyzer on a stub antenna to see how much distortion overloaded toroid inductors in the tuners corrupt the signal? Heck, the toroid in the directional coupler could be overloaded and make distortion worse rather than better. Numbers would be a good thing. A resistive tap might be the best at the PA output. But, that still leaves what comes out of the tuner and coax as issues. (Coax is microphonic and can generate mighty spikes if you hammer it, even lightly. So a kW in a feed line could introduce some strange non-linear effects in the dielectric. Perhaps the best is open-wire feed line. But then you have the balun to worry about.)

I really do want to see the Pure Signal code implemented. But, be prepared to think the hardware end of Pure Signal through, carefully.

{o.o}

On my list for 2021.

jdow wrote: Fri Nov 06, 2020 9:17 am Do you have the hardware to support it? It requires an unusually good directional coupler and a suitable attenuator, at least for something in the hands of hams.

{o.o}

It doesn't require a directional coupler. I built my own very simple pure resistive adjustable non-directional coupler and put it after the output of my tuner, which is after the output of my AL-80B. So the PureSignal in my Anan 700DLE MKII is correcting for the nonlinearity in the Anan PA and the AL-80B and any distortion or phase error introduced by the MFJ-989D tuner.

So (all modesty aside) I've probably got what may be the cleanest signal on the air, with the PureSignal correcting the sum of all of the phase and linearity errors in my entire RF stream. And all it cost me was a few resistors and connectors from the parts bin and a $6 Hammond 1590B box.

I have the coupler adjusted (via vintage film rheostat) to provide a usable sample at any input power level between 15 watts (for digital modes) all the way to the 1KW limit of the AL-80B. The Anan automatically attenuates the sample across a 31dB range as needed during operation.

PureSignal on the Anan is Freaking Awesome!

Just noodling here 2 kW is about 316 volts RMS on the coax. The receiver input is "about" 50 ohms. If you want a -20 dBm, -50 dBw aka 0.0224 Vrms, going into the receiver the attenuator series resistor is about 700k ohms at 1/8th watt. Stray capacitance will be a small issue at 2kW continuous. Pick resistors appropriately. I admit to missing this part of the design process a few years ago when I thought to put a high value shunt resistor on an antenna that I expected to pick up gross lots of static electricity. A 1/4W 10 k ohm burned itself up on 100 W that was continuous a little too long.

{o.o}

My resistor pack is a 3x3 parallel-series arrangement of 2-watt carbon composition 18K ohm resistors (3 resistors in parallel per bundle, 3 bundles in series). Total dissipating capacity is 18 watts so power handling is certainly not a problem. And with that arrangement I could even have any given resistor go to an internal dead short failure mode (extremely unlikely with that resistor type) and not suffer anything more serious than a possible momentary ADC overload (not harmful) - the net result would only be a ~2dB increase in sample strength, and I'd immediately see the effect on the transmitted signal on the panadapter and stop transmitting.

The only real danger to a pure resistive sampler would be if I were to transmit large power (i.e., an amp) with no antenna connected to the coupler output at all, which would require a comedy of errors to successfully pull off.

E^2 for 2 kW at 50 ohms is 100,000V^2 (316 volts RMS). E^2/R is 5.5555W for each of the 18k resistors. You are at the "official" edge at 727 watts. If course, that poor thing WILL get HOT if it is "RF small".

{o.o}