The K3 and contesting – when ignorance is bliss

The running thread on the Elecraft reflector at the moment concerns a number of complaints regarding the K3’s performance during certain conditions during CQWW (CW). Specifically at issue is the complaint by some (quite a few?) that when 4 or 5 weakish signals…say, S3, S4 and S5 – are within the passband of the filter in use, the receiver turns them all into “mush” making each signal indistinguishable from the others.

Said one:

I personally know at least two top-tier contesters that are seriously
considering replacing their K3s with a different rig because of this
single issue, and I personally know two others who did not buy a K3
because of it.  If I knew then what I know now I would not have bought
my K3 either.  I spent a lot of money (tower, antennas, amplifiier) to
be able to put a strong signal into various DX locations, and it is
simply embarrassing to be working a pileup with a half dozen weak
callers that obviously can hear me fine and not be able to pick their
callsigns out of the mush.

The same problem is mentioned by other experienced contesters/DXpeditioners including ZK2V’s comments of the same problem on SSB from Nov 4 here.

Turning off the AGC alleviates the problem thus allowing each of the signals to be heard individually. This of course presents the potential problem of overload in the event a strong station begins transmitting in the passband.

Elecraft is already on the issue, attempting a fix and offering additional capacitors to be placed in parallel with one currently being looked at as the possible culprit.

As may be seen from my previous post, I don’t hesitate to give negative virtual ink to a product when it rates such criticism, but in this case, I have to say that I was impressed with the K3 during the contest. No, there were no pile-ups calling me, but I was in many pile-ups of a type that are unique to CQWW:

Contest pile-ups are unique in that the split difference between transmit & receive is minimal. Non-contest DX splits are almost always in excess of 1 kHz…in CQWW, I noticed they were often 150-400 Hz (if 150 Hz even constitutes split). This puts several stations within my passband as other contesters (and me) listened to the DX on one VFO and transmitted on the other VFO just a tad higher.

If I’d been aware of the issue now being discussed on the Elecraft reflector, maybe I would have noticed it under those circumstances. But  the fact is, I was too busy being impressed at the ability to operate with such a minimal split and get the DX logged quickly. I never could have done that with any other rig I’ve owned. Rather than hearing 3 or 4 stations within my passband, I would have heard 3 times that amount and their combined influence on those rig’s AGCs would have made the DX unreadable.

What makes CW fun in the DX game and separates it from all the other modes is the ability to inject strategy rather than random luck or brute force into split operation. The K3, more than any other rig I know of, allows this ability to be exploited.

In February’s ARRL DX Contest, I’ll listen for the symptom described on the reflector and, if Elecraft has engineered a fix by then, I’ll implement it, even if the problem is beyond my own meager threshold of detection.



  7 comments for “The K3 and contesting – when ignorance is bliss

  1. Jim K6OK
    December 5, 2011 at 11:51 am

    I’m wondering if we’re seeing the limits of analog filtering and SDR rigs will emerge as the top contest rigs. Analog filters, by their very nature, will always create some degree of ringing, mush, and insertion loss. With SDR’s you can filter by simply telling the computer to change 1’s to 0’s outside of any chosen passband. I don’t pretend to have much knowledge of A/D conversion and Fourier transforms and the limits they bring to the table; however, I’ll speculate anyway and venture a guess that the next generation of rigs will be SDRs-with-a-knob because of superior performance and lower cost to manufacture. The era of the superhet may be over…

    • December 5, 2011 at 12:43 pm

      Isn’t the K3’s DSP filtering equivelent to SDR in that it’s digital, ie controlled by firmware algorithmns (which can be – and frequently are, updated)?

      As I continue to follow the thread on the reflector, I’m a bit surprised at the expectations of some…that several S3, S4 signals within a very narrow passband should be individually discernable. I wonder what rigs these guys used prior to their K3’s which lead them to have these expectations.

      • Doug
        December 10, 2011 at 5:15 pm

        “I’m a bit surprised at the expectations of some…that several S3, S4 signals within a very narrow passband should be individually discernable”

        The effect I have heard is the following:
        Numerous US stations calling me all about s5(noise floor s4) on 80m cw. The pile up grew until they all “vanished” i.e. I was suddenly listening to a clear frequency. I thought condx had changed for the worse, then realised I could still hear perfectly discernable letters from the end of the someones call. I worked another 20 stations or so by repeating the only 2 letters or so I could hear. I realised that the K3 was mushing the pile together and presenting nothing to the speaker, then when only the longer calls were still sending and most others had stopped sending, the K3 recovered audio to the speaker. In 30 years as a ham operator I have never heard this effect before. The problem with this discussion is that until you hear the K3 doing it, you will put it down to operator finger trouble or unrealistic expectations. When you hear it, you will know something is broken. Lets hope for everyones sake that Elecraft will keep searching for a solution to this rare and hard to replicate problem.

        • December 12, 2011 at 3:07 pm

          The Index QRP+ is a great concept, but the AGC problem makes it virtually useless in a pile up when a strong adjacent freq. signal (i.e., within ~ 1.5kHz) transmits. Another factor is the shape of the audio frequencies of signals overlapping at roughly similar frequencies (i.e., 420Hz, 430Hz, 440Hz. A razor sharp filter could slice the single signal out for Q5 copy. But at the overlap of two audio signals (see on a scope) where the two are of roughly equal magnitude, what occurs is an artificial “noise level” (SNR etc) which decreases the audible magnitude of the peak of each signal. It is the same effect when an intermittent power line noise shows up. Some signal, but mostly hash. Given the difference in timing of the c.w. characters, what would result logically is noise — destructive interference of the two beating signals. Ever listen to the violin section of a middle-school orchestra tune up? The “recovered audio” is undoubtedly recognizable as a violin sound, but what note frequency? Vibrato helps. What we need is a vibrato/chirp circuit on weak signals. The two beating signals will then always be changing relative to each other, hence, differentiation! More than interesting subject to me because I’ve held off getting a K2 since I still love the Wilderness Sierra. Then the K3, and hopefully the KX3-to-come would lead me to reconsider the Sierra’s priority.
          72, Ade W0RSP

          • December 12, 2011 at 3:28 pm

            Ade, I really enjoy the QRP+ although of course it’s not in the same league as anything from Elecraft. One of its receiver’s significant shortcomings is the asymmetry of its passband. Check it out here and note where the center freq is:

            I’m a bit surprised at your comments on using the Sierra due to it not having split capability.



  2. Ed
    December 5, 2011 at 2:53 pm

    I think you hit on the $64K question, John. Were they even hearing those S3/S4 signals before at a point where they were half-way intelligible before? Or has the K3 simply stretched the playing field?

    I will say that my ‘590 took a lot of getting used to and I still don’t care much for how it can make a bunch of signals all sound like they’re at the same strength. It does make identification and copy more difficult. With the K2, I can immediately recognize who is slightly weaker and stronger, even within a narrow bandwidth. On the other hand, my interest in contesting has pretty much dwindled to the non-existent level, so the ‘590 does a bunch of things good for DXing. I’ll be curious to see how this develops and what type of solution Elecraft comes up with, very interesting.

  3. Jim K6OK
    December 5, 2011 at 3:06 pm

    I don’t believe so. DSP following a crystal filter can’t completely undo the damage a crystal filter does to a signal. In my mind pure SDR should provide cleaner filtration than DSP-after-crystal filters.

Leave a Reply

Your email address will not be published.