On 23-May-10 3:49 PM, Joe Subich, W4TV wrote:
>
> I don't care what kind of transition Joe has programmed into the
> DSP. The rise time/wave shaping makes little difference with a
> class C amplifier because the sharp turn on/turn off characteristic
> and saturated operation will always square up the pulses.
There is no turn on/turn off. That's what I'm trying to tell you. The
smooth phase transition that occurs as the tones change from one to the
other is exactly the same as slowly turning the knob on a linear (as
opposed to a PLL digital) VFO. NO CLIX!!!
>
> A Class C amplifier requires a certain level of drive to "turn on"
> and operates in saturation ... the turn on threshold will remove
> much of the low level ramp up and the saturation will create sharp
> "corners" at the top of each pulse. Even the most basic measurement
> with a dual trace oscilloscope will show the difference between a
> linear amplifier and one run in class C. Run a string of dits
> through a class C amplifier and compare the difference between
> the drive signal and the output.
Any time you turn on and off a "signal" you generate sidbands (sinX/X).
You can modify the wave shape and reduce the sidebands (also known as
clix) but you cannot eliminate them entirely. At least not until the
signal is either on all the time or off all the time or looks like a
sine wave. If the keyed wave form looks like a sine wave there will be
no sidebands generated.
>
> The clicks are not receiver generated ... they are generated in
> the transmitter/amplifier by pulse squaring and eliminating the
> wave shaping created in the exciter. W7AY has done extensive
> work with wave shaping and can show the "damage" caused by non-
> linear amplifiers in FSK and AFSK systems.
Strong RTTY does indeed generate clix in some receivers, at least they
do in my "economy" DSP if filterd radios. NB makes it worse in those
radios.
If the RTTY tones are "keyed" as you imply above, then yes I agree they
can and will make clix. However if the tones phase is seamlessly
transitioned at the zero crossing, there will be no clix. It all
depends on who did the DSP programming on a SSB type setup or who
designed the radio internals on a direct keyed setup.
An amplifier that is only conditionally stable "in band" will also
generate clix as it tries to self oscillate over a small portion of the
RF cycle during phase transitions. Not common, but I've seen it happen.
>
> > Last I looked Duty Cycle was defined by the ratio between ON time and
> > OFF time. No time limit specified.
>
> Any calculation of "duty cycle" must include a time limit ...
Why "must" it? Duty cycle specified on transformers, motors, is not time
limited. Duty Cycle implies unlimited time operating normally. Normal
operation of a ham amplifier includes RX periods as well as TX periods.
By your definition, CW has a 100% duty cycle since you only want to
count the time when it's actually transmitting. I think we are confusing
duty cycle of the modulation envelope verses duty cycle of the SYSTEM.
>
> Simply put, when one limits the calculation to transmission
> periods, AM, FSK/AFSK, WSJT, etc. are 100% duty cycle modes.
> CW is a 50% - 60% duty cycle mode (depending on message
> content) and SSB is up to 60% (depending on compression or
> clipping levels). If you start adding "receive time" all
> of those numbers will decrease but it still does not eliminate
> the fact that AM/RTTY/WSJT/etc. are as much as four times as
> taxing on an amplifier than CW at a similar peak power level.
By the above numbers, FM/RTTY/WSJT is still only 2x taxing. 50% CW vs.x
100% WSJT/FM/RTTY. AM is not even close. 100 watt carrier is 25% of the
400 watt PEP of an AM signal. With modulation at full PEP assuming tones
not voice the average is still only about 50%, so same same as CW.
>
> Good for you ... however, the Eimac data sheets clearly show 400 mA
> single tone (CW) plate current in AB2 service at all plate voltages
> from 1500 - 3500 (http://www.g8wrb.org/data/Eimac/3-500Z.pdf).
I stand corrected, that spec sheet does indeed show 400 ma. As much as
those things cost these days even 350 makes me nervous!
73, Larry
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|