Dan, here are two simulations. One has a diode across L14, as Paul
described. One does not. You can see one leg of the diode dangling.
The plots show the drain voltage at the BS170 transistors on the left
of the schematic. Note the big spike at 9 uSec when I end the
transmission burst. The spike maxes out at 60 volts because the
transistor model includes a 60V breakdown spec.
Thanks to Evan AC9TU and Tony S, for assistance in evolving this model.
JZ KJ4A
toggle quoted message
Show quoted text
On Mon, Jul 17, 2023 at 11:54?AM John Zbrozek <jdzbrozek@...> wrote:
Dan,
The QMX modulator can gradually increase or decrease transmit power in response to its DAC input.
One possible use might be to recognize a "tune" mode and preemptively reduce power to a safe value. Response to high SWR is another use. Yes, timing might be critical, but in the case where high SWR is the result of a low load impedance, the transistor failure mode is over-current and over-heating.
You have a little time then to respond in that case. The highly reactive load over-voltage case could be a different matter.
The modulator can also be used for wave shaping a CW
dot or dash, or it can be used to impress audio modulation onto a carrier. It can be used to reduce the di/dt portion of the equation V = L di/dt at the end of a transmission, thereby killing off the dangerous inductive spike. Think of it as applying a little wave shaping at the end of an FT8 burst.
On your SSB question;
Yes SSB is very different from the digital modes.
FT8 for example transmits at a constant amplitude but with a varying tone frequency.
SSB will be generated in QMX by modulating the angle/frequency of the carrier signal, amplifying it in an efficient but nonlinear power stage, and then impressing an audio amplitude modulation signal onto the final amplifiers power source rather much like an old style AM transmitter.
The resulting varying angle/magnitude vector simulates an SSB signal
The rub here is that at the lower output levels at the finals, the amplitude you command with the DAC is not necessarily what you actually get.
As the drain voltage of the final amplifier decreases, the drain capacitance of the transistors increases sharply. Zeners would do the same thing.
This results in knocking the output filtering around, changing the output power in ways which adds audio distortion to the transmitted signal.
The generally accepted solution is to pre-distort the audio in a compensating way.
It's tempting at that point to say " Easy! The audio is digitized. It's just a game of numbers!" Sure, but which numbers do you use? That depends on how much pre-distortion is required and the answers could be different with or without Zeners.
Regards and 73,
JZ KJ4A
On Mon, Jul 17, 2023, 11:00 AM Daniel Walter via groups.io <nm3a@...> wrote:
Jz,
re: "Regarding QMX, yes the pads for Zeners exist but are currently unpopulated.
Hans has a truly elegant solution to the above problem in that the PMOS
modulator that is present in QMX and is under DAC control can be used to
gracefully wind down final amplifier current, avoiding the Ldi/dt problem
altogether. Hopefully it is to be used in all transmission modes."
I don't quite understand how Hans fix will work. But if it simply reduces power in response to high reflected power, then that will not really help as it will probably be too late and the damage is already done.
I suppose it could start with low power and ramp up if no significant reflections but that seems cumbersome. Won't the Zeners along with Hans fix take care of it more easily?
As far as SSB is concerned, I can't imagine any different conditions there from digital. After all it still will be a very similar signal The finals won't really see any difference, will they?
I already populated mine with 51 V, 1 W Zeners, but I am eager to hear from those who can do quantitative analysis on this.
--
73, Dan NM3A
