Audio Quality: The most obvious thing is, the K3's audio is much more crisp than that of the JST-245. By "crisp" I mean it has a bit more high frequency response, while the '245 has more punch at the low frequency end. Which is better is more a matter of preference than science. My ears generally find the weakest signals more easily readable with the K3, while signals S9 or better tend to have a more natural tone with the '245. But the differences are subtle, neither radio sounds dramatically better than the other when using a common audio system (in this case, the Multi-RX feeding a pair of small Sony speakers left over from my home theater surround sound installation, as well as a pair of JRC ST-3 communications headphones). The K3's stereo audio effects (AFX) settings can tremendously improve the quality of reception and reduce listening fatigue, and works equally well with speakers and headphones; the AFX feature gives the K3 a slight edge.
Noise Reduction: Most notable is the DSP Noise Reduction (NR). Even at it's lowest setting, the NR removes so much background hash with the push of a single magic button that I find myself turning it on and off just to come to grips with the fact that it actually works. There are several NR settings selectable by holding the NR button and dialing in the level. Lyle Johnson KK7P, Elecraft's DSP genius, says:
NR is more properly Signal Enhancement rather than Noise Reduction. The distinction is subtle but important if you wish to understand how it works and how to best apply it. If ti were noise reduction, the implication is that it passes everything until it figures out what is more likely to be noise, which it then attempts to suppress. Since it is signal enhancement, it tends to pass nothing until it figures out that which is more likely signal, which it then attempts to pass.Noise Blanker: The K3 Noise Blanker (NB) is also a masterpiece - between the separate IF and DSP blanker settings I've been able to remove all sorts of crap that would otherwise make it impossible to hear any but the big gun signals. It's as simple as this: When I hear any sort of electrical noise, I first crank the DSP NB up to see if it has any effect. If not, I try the IF DSP. Usually one or the other does the trick, I haven't found any noise that requires both at the same time. The '245 NB is useless by comparison; if it works at all it must be against a very specific type of impulse noise that I've never encountered, because I've never found it to be of any use at all (unless distorting a desired signal is of use to anyone).
With this understanding there is one more concept I must explain: correlation. This is the degree to which a signal is similar to itself; conversely, it can be used to figure out how noise is dis-similar to itself (more random). This is the basis on which NR works.
In the current implementation of the K3, we have four (4) basic NR filters. These are displayed as F1..F4. F1 is the gentlest, F4 the most aggressive. Recalling that the filters tend to suppress everything until they can sort out what is probably a signal, there is some time delay involved in their application. Further, especially with voice signals, some components of the signal may not be recognized as well as others. THis gives rise to distortion.
So, we added a second field to the NR. This is the -1..-4 which specifies a certain amount of "bleed through" of the original signal. This reduces the apparent distortion and delay, but limits the ultimate S/N improvement. -1 provides the least distortion but limits the S/N improvement on weak to moderate signals to about 6 dB. -4 provides no "mixing" whatsoever and can result in dramatic S/N improvements, but at a cost of slight time delays and increased distortion.
NR is not appropriate for very weak signals, so a -1 or -2 is best if you wish to hear them.
Having said all this, the correct way to use the NR is to listen and adjust it for the best compromise between noise reduction and distortion. Everyone is different in this regard. Some tolerate noise better than others, while some don't tolerate distortion very well at all.
I live in a quiet location, and use NR1-2 as my most common setting. This is low in distortion, allows me to hear weak signals, and shows no apparent delay as I tune through the bands. On the rare occasions when things get noisy here, I crank it to higher values.
Notch Filter: With the '245 I almost always use the manual notch filter as a mid-cut to tailor the receive audio to my liking. I find this technique to be a bit more effective with the '245 than the K3 as the latter takes an awful lot of dial turning to adjust the notch range from 200 Hz to 3.92kHz in 20 Hz steps. Also, the K3 notch width is much sharper than the '245 so the effect is more subtle. Of course the K3 has built-in graphic equalization for the purpose of shaping the audio response, so this issue is pretty much moot and leaves the notch filter to be used for its intended purpose - to remove heterodyne interference. To this end, the K3's Auto Notch is superb. The manual says it will, in some cases, remove multiple carriers although I've yet to find a situation that will confirm this. The '245 does not have an auto notch function, but it has Notch Tracking which will offset the manually adjusted notch filter when the VFO is tuned slightly; this is useful but not nearly as effective as a fully automatic system. Notch depth is adequate on both rigs.
Ergonomics: Both of these radios are a joy to use. The '245 has a solid feel and a heavy, balanced main VFO knob that spins like a top. All of it's buttons have a nice positive 'click' when pushed, and the smaller knobs and controls feel good and not overly 'plastic'. And the color LCD! - it's about the finest looking display of it's generation and, in my opinion, better than anything even today short of the newer Icom color TFT displays of the IC-7800, IC-7700, etc. Because JRC took the time and effort to find out what amateur radio operators really wanted and where things were supposed to go on the front panel, all of the most commonly used controls are assigned to individual buttons or knobs with very few alternate functions or hidden menus to toggle through; outstanding ergonomics was a chief goal during the JST-245 design review, especially after the JST-135 got raked over the coals by Dave Newkirk in his QST review (and deservedly so). So although I may be biased as a member of aforementioned JST-245 design review team, I find the JST-245 to be a more enjoyable radio to sit in front of and operate.
The K3, however, is no slouch. Elecraft had different design criteria, one of which was to make the K3 compact and easily transportable. Neither of these adjectives can be applied to the '245, therefore JRC had a whole lot more front panel real estate to work with than did Elecraft. Consequently many of the K3's controls are doubled up, several often-accessed settings are menu-driven, and some things which take a single button push with the '245 take two or more with the K3. But by and large I think Elecraft got it right, the K3 has just the right amount of compromise between keeping the size of the radio down, making it attractive to the field user and people with limited space, and including all the high-performance features demanded by discriminating users such as contesters and DXers. So while I may have to toggle through different modes to set the K3 instead of pressing a single button, it's not a big deal to me. The K3's buttons have a soft feel to them which reminds me of the Drake R8 receiver's front panel buttons. While the main tuning dial does not have the same flywheel feel of the JRC it still spins nicely. The main liquid crystal display is very sharp and has a wide range of backlighting levels (including OFF), though it must be viewed at a fairly straight-ahead angle. All in all, from a usability perspective, the K3 is a nicely designed little rig.
The one area in which I find the K3 to be deficient is in its selection of memory channels - unlike the '245 which has a dedicated Memory Channel knob that lets the user dial through channels one by one, the K3 has a more convoluted process of pressing the M>V button, dialing up the memory channel with the VFO knob - without being able to hear what's going on at that channel's particular frequency! - then pressing M>V again to set the rig to the stored frequency, mode and other parameters. If I could change one thing with the K3 it would be to let me hear what's on each channel as I dial through the memories, before I hit the M>VFO button. This should be an easy firmware fix.
And the winner is...: In the tweakability department the K3 has a few bells and whistles that are missing (or ineffective) in the '245. In practice I find that I am able to get similar reception with both rigs by using each radio's tools to maximize the desired signal and minimize the junk. I am still amazed at how well the JST-245 holds up to the latest and greatest in DSP technology, and it is such a pleasure to use. Therefore I've abandoned any thoughts of selling the '245 to finance the acquisition of a second K3; it is simply too good a rig to let go of.
That said, the K3 is clearly the future of HF transceivers, and it looks to be a pretty bright future at that. While the above observations are subjective and based narrowly on SSB reception under less than perfect antenna and propagation conditions I think it's pretty clear that the K3 is a gem of a rig. Of course many of the design elements that on paper put the K3 in the league of the mega-buck rigs from JA-land do not come into play with my limited antenna system - for example, I rarely see a signal stronger than S9+10dB, and have yet to find two such signals close together as in a contest situation, therefore the benefit of the narrow roofing filters and strong front-end performance remain untested by me in any real way. In any case, others (ARRL, Sherwood*, etc.) will confirm or deny Elecraft's published dynamic range and IP3 specs soon enough, so I'll leave it to the pros to do the scientific evaluations.
If held at gunpoint and told to choose one of these two radios, subjectively, emotionally, and based solely upon my ears, fingers and eyes... it would have to be the K3. But barely.
* Update 09-Feb-08: Serendipitously, Rob Sherwood has released some of his K3 test results which were promptly passed on to the Elecraft reflector by Wayne Burdick:
Rob Sherwood, NC0B, an independent and well-known receiver performance specialist, has completed his K3 receiver tests. We're pleased with the results, which will place the K3 at the top of his comparison chart.
Rob will be updating his web site in a few days. For now, I'll just mention a couple of his numbers (with his permission).
Elecraft K3, S/N 00149, 20 meters, preamp off:
Dynamic Range 20 kHz 104 dB
Dynamic Range 5 kHz 102
Dynamic Range 2 kHz 101*
* with 200 Hz 5-pole filter
Blocking above noise floor at 100 kHz spacing, AGC On: 140 dB
Phase noise (normalized) at 10 kHz spacing: 138 dBc/Hz
Rob performs some tests differently that we (and the ARRL) do, but in general we're all in close agreement. Note that the unit under test had only Elecraft 5-pole filters. Our tests show the 8-pole filters to be as good or better, and we sent some of them to Rob to test when he gets a chance.
73,
Wayne
N6KR
(From 1978 WRTH)
Posts
