ARRL Clean Signal Initiative
Posted: Tue May 16, 2023 8:47 pm
Not strictly related to Apache hardware or openHPSDR software, but perhaps of interest to those of us who already have the cleanest signals on the HF spectrum. See the article in the June 2023 edition of QST, page 55.
The only mention of PureSignal, and not by name, is this one sentence in a three page article:
"Predistortion, a design technique known to help with signal purity and address IMD, is a currently niche feature available from only a few select manufacturers."
Obviously the software and firmware doesn't come from a "manufacturer". Supporting hardware includes Apache, Red Pitaya, Hermes Lite 2, and TRX Duo--am I missing anyone?
In fact, it seems that the author of the article, Mike Ritz, W7VO, and his reviewers and editors, went out of the way to discuss everything under the sun except for predistortion linearization.
Of course the single most important thing that anyone can do to clean up their act is to simply avoid pushing their amp past the 1dB compression point. Doing so is probably the cause of most band splatter. Not doing so is a simple idea that the article did not talk about much. But that will never happen because even if you are using a high end, (almost) peak reading RF power meter like an LP-700 it's still going to tell you that peaks are only at 1200W when you reach the 1dB compression point of your favorite, legal limit amp (give or take a dB). And since nobody is going to accept that ("I paid for 1500W, damn it!") that approach will never work.
The cool part about PureSignal (or any well-engineered predistortion linearization scheme) is that it acts like an automatic limiter for this sort of behavior. Keep turning the drive up and linearization will fight you tooth and nail, effectively turning it down as required to prevent putting the amp very far into compression.
Since building an HF amplifier in the kilowatt range with -50dBc spurious output is technically demanding and probably costly, linearization is almost certainly a much less expensive way to accomplish the same thing. Especially since even great amp's can be driven to be bad amps!
Both Elecraft and Icom are showing flashes of brilliance. The Elecraft K4 has a feedback input on the back, but no schedule for firmware to use it has been published. The yet-to-ship Icom PW2 amplifier is claimed to have the ability for "digital predistortion" when used with the IC-7610, but nobody has seen it yet.
73,
Scott
The only mention of PureSignal, and not by name, is this one sentence in a three page article:
"Predistortion, a design technique known to help with signal purity and address IMD, is a currently niche feature available from only a few select manufacturers."
Obviously the software and firmware doesn't come from a "manufacturer". Supporting hardware includes Apache, Red Pitaya, Hermes Lite 2, and TRX Duo--am I missing anyone?
In fact, it seems that the author of the article, Mike Ritz, W7VO, and his reviewers and editors, went out of the way to discuss everything under the sun except for predistortion linearization.
Of course the single most important thing that anyone can do to clean up their act is to simply avoid pushing their amp past the 1dB compression point. Doing so is probably the cause of most band splatter. Not doing so is a simple idea that the article did not talk about much. But that will never happen because even if you are using a high end, (almost) peak reading RF power meter like an LP-700 it's still going to tell you that peaks are only at 1200W when you reach the 1dB compression point of your favorite, legal limit amp (give or take a dB). And since nobody is going to accept that ("I paid for 1500W, damn it!") that approach will never work.
The cool part about PureSignal (or any well-engineered predistortion linearization scheme) is that it acts like an automatic limiter for this sort of behavior. Keep turning the drive up and linearization will fight you tooth and nail, effectively turning it down as required to prevent putting the amp very far into compression.
Since building an HF amplifier in the kilowatt range with -50dBc spurious output is technically demanding and probably costly, linearization is almost certainly a much less expensive way to accomplish the same thing. Especially since even great amp's can be driven to be bad amps!
Both Elecraft and Icom are showing flashes of brilliance. The Elecraft K4 has a feedback input on the back, but no schedule for firmware to use it has been published. The yet-to-ship Icom PW2 amplifier is claimed to have the ability for "digital predistortion" when used with the IC-7610, but nobody has seen it yet.
73,
Scott