> If what Joe claims is true then this is a more serious issue > than even I imagined, since it affects SSB operation as well > as data modes. It would be so serious that I suspect he is > wrong about it. > > If I dial up 100W in SSB or data I expect the power to be > limited to 100W PEP, not 100W average power.
The K3, like any other peak reading wattmeter uses a time constant appropriate for "normal" voice. As such in SSB the K3 adjusts for the typical peak to average of voice - one of the reasons for TXG VCE. Serious investigation will show that TXG VCE will vary with the particular voice and the level of compression (clipping) prior to the power control circuitry. Watching the output with a scope or fast peak reading wattmeter like the LP-100 will also show the differences (and deviations) with voice characteristics and compressions levels. However with digital modes like PSK31 the peak to average ratio varies with modulation and compression/clipping can not be tolerated if IMD is to be avoided. To be accurate with arbitrary digital modulation, the wattmeter must be able to respond to peaks in the fraction of a millisecond range. To adequately control the power level and must be able to predict the required peak to average ratio (headroom). In commercial applications (cellular telephone, digital TV, etc.) the power amplifiers are regularly operated at average power levels 10 to 20 dB OR MORE below saturation (the 1 dB compression point) with long time constant average power control due to the high peak to average ratios. > One reason I suspect he is wrong is that the observation of > getting around twice the requested power applies to any data > mode, not just PSK31. I can feed a pure sine wave in, request > a power of 5W, and get 10W out. If the K3 power meter was > indeed average level then it should still limit the power to > 5W as it does in SSB. No. If you were operating in SSB (and DATA A is SSB), the internal "adjustment" for peak to average in a voice signal would cause the gain in the amplifier chain to be increased to provide approximately twice the requested power output. I can create almost any arbitrary level of "overpower" for a single tone in DATA A simply by adjusting TXG VCE. If it is set (as I normally do) for +1.5 dB, I will see just about 3dB of "overpower" when driving the k3 with a single tone and adjusting the line input for five bars of ALC. If, on the other hand, I reduce TXG VCE to -1.5 dB, the power level will be almost exactly the requested level (15.8 for 15, 98 for 100 on 20 meters). This has nothing to do with the peak to average ratio since this is with tone ... only with the way the K3 controls power in SSB modes. Again, the K3 wattmeter is essentially average responding (as you would expect for CW and to be used with the tuner). The MCU then makes an "educated" guess bout the peak to average power and duty cycle in SSB. TXG VCE then allows the user to "tune" that adjustment for their own voice and level of compression. TXG VCE does not seem to effect the power level for a single tone in SSB (as expected) but it seems to have a significant effect in DATA A. That appears to be as expected if the power control is based on an average reading wattmeter with a compensation factor. Perhaps the MCU needs another control - TXG DTA anyone? By the way, AFSK A behaves exactly like SSB and does not show the "overpower" with tone. 73, ... Joe, W4TV > -----Original Message----- > From: [email protected] > [mailto:[email protected]] On Behalf Of Julian, G4ILO > Sent: Sunday, January 04, 2009 5:50 AM > To: [email protected] > Subject: RE: [Elecraft] K3: Still unable to set power > correctly in DATA A(sigh) > > > > > > AD6XY wrote: > > > > > > > > Joe Subich, W4TV-4 wrote: > >> > >> > >> Since the K3 power control is based on average power and not > >> instantaneous (absolute) peak power, the power level (both > >> requested and reported) can not be relied on as a gauge of > >> proper operation unless the user knows with certainty how much > >> headroom is required to handle the peak power without compression > >> or clipping. > >> > >> > > > > Unfortunately, the life of amplifier power transistors and > to a lesser > > extent valves is based on the peak power. Most license > conditions are > > also based on peak power. > > > > A power control based on average power is therefore, > irrespective of > > any technical merit, inappropriate. It is theoretically illegal in > > many countries to use with a following amplifier as well as being > > likely to result in damage to that PA. > > > > This is why it is such a serious issue and why we keep banging on > > about it. > > > > Mike > > > > > > If what Joe claims is true then this is a more serious issue > than even I imagined, since it affects SSB operation as well > as data modes. It would be so serious that I suspect he is > wrong about it. > > If I dial up 100W in SSB or data I expect the power to be > limited to 100W PEP, not 100W average power. > > One reason I suspect he is wrong is that the observation of > getting around twice the requested power applies to any data > mode, not just PSK31. I can feed a pure sine wave in, request > a power of 5W, and get 10W out. If the K3 power meter was > indeed average level then it should still limit the power to > 5W as it does in SSB. > > The issue here is not how the K3 power control works. It is > that it is not working in DATA A as it does in SSB. DATA A > ought simply to be treated as another SSB mode that just > allows a different choice of audio input and filter settings. > I believe that it did originally work like that in early > versions of the firmware. > > Sorry to continue the topic after Wayne requested we call a > halt on it, but the assertion made by Joe opens up a whole > new can of worms if he is indeed correct about it. > > ----- > Julian, G4ILO. K2 #392 K3 #222. > http://www.g4ilo.com/ G4ILO's Shack > http://www.ham-directory.com/ Ham > Directory http://www.g4ilo.com/kcomm.html KComm for > Elecraft K2 and K3 > -- > View this message in context: > http://n2.nabble.com/K3%3A-Still-unable-to-set-power-correctly -in-DATA-A-%28sigh%29-tp2106260p2108786.html Sent from the Elecraft mailing list archive at Nabble.com. _______________________________________________ Elecraft mailing list Post to: [email protected] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com _______________________________________________ Elecraft mailing list Post to: [email protected] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com
