Since I do read German, here is my quick attempt to translate the section ("Fazit") where the author summarizes his findings about testing methods and receiver performance:
"No IP3 value can be derived from the results of this measurement technique using broadband pulses with a crystal-based notch filter. However, the results do show the resistance to interference generated from summation of a broadband interfering spectrum, which seems crucial to me when it comes to evaluating the interference handling of an SDR with the ADC at the front. In any case, the listed signal levels (for +10 dB (S+N)/N) provides a very good way to compare the relative performance of the tested receivers; the higher the listed signal level (dBm), the better the better the performance in handling strong signals. Roofing filter quality can be observed by proper selection of pulse frequency for close-in measurement (use e.g. 2 kHz). This is in contrast to what can be done with two-signal testing (even using many different frequencies). In particular, the pulsing measurement technique illustrates how the use of a preselector influences receive performance. Earlier one could only rely on subjective impresisons in practical use. Our results show that use of a preselector always makes sense, even when used with the very best receivers!!! Even with the RX-OE3HKL receiver design which by itself can handle levels that are up to 26 dB higher, improvements up to 9 dB from the preselector are seen. Notable is the performance of the IC-751A with preselector and roofing filter: Results are almost as good as those of my own receiver design. However, in contrast to the situation when using the 2-signal method of testing, when using the pulse signal, without the preselector, the roofing filter doesn't yield much improvement. Apparently the reason is that the broadband signal quickly overloads the front end circuits. Without using the broadband pulse signal it would not be possible to reconcile lab measurements with real-world reception results seen when using the 2x120m V-antenna! However, the improvement from using the preamp shows up only if the preselector handles large signals at least as well as the rx itself. The K3 is by far the best of the commercial units, and without having to be modified. This result is confirmed by practical experience in reception testing using the 120m-V-beam during a contest. Results improved by another 2-4 dB after a firmware update was performed (see the red corrections). The IC-7300 clearly performs more poorly than the analog radios with roofing filters, in practical reception as well as in lab measurement. When the preselector is added, it partly outperforms the FT-1000 Field with preselector and roofing filter. Based on this result, I conclude that an SDR with the ADC in the front end, when used with good antennas and in the presence of sftong signals, should if possible only be used in combination with a preselector. I would like to establish that the here described test setup with its broadband discrete signal is a better proxy for real-world interference signals than is the stochastic noise signal used in Noise Power Ratio testing. Furthermore, by varying the pulse frequency, one can vary the intensity of the interference in a way that is clearly quantified by observing spectrum analyzer output. In contrast, the noise level in NPR measurement is defined by the high ratio of peak value to rms (10 to 11). In this scenario t would likely take very expensive equipment to perform measurements with sufficient accuracy for comparison purposes. I would also like to mention that NPR measurements at 3dB (S+N)/N will not be sufficient to take into account the effect of higher-order IM products. To achieve that, one must increase input levels to the point that a further increase of 1 dB will lead to a 3 dB increase at the output (3rd order). This is what it takes to get results from the NPR method to agree with results from the pulse signal method." My apologies in advance to OE3HKL for any misinterpretation I might have made. 73, Erik K7TV -----Original Message----- From: elecraft-boun...@mailman.qth.net <elecraft-boun...@mailman.qth.net> On Behalf Of Erik Basilier Sent: Friday, April 06, 2018 5:06 PM To: 'Bob DeHaney' <bobdeha...@gmx.net>; elecraft@mailman.qth.net Subject: Re: [Elecraft] OE3HKL's Measurements The German website describes how this om has developed his capabilities over the years, based on his specific circumstances and needs. This has resulted in a measurement technique with arguable advantages, achieved with a homebrew, very specialized test setup, implemented for the 40m band. While the author apparently had access to some lab-quality test equipment for testing his home-brew test equipment, it appears to me that it would be much more expensive to implement his receiver testing setup using only a combination of off-the-shelf lab test equipment, if that is even practically doable. I haven't read everything on the website, but for those who find the German confusing, the following is my summary overview of some of the site content. He put up a huge horizontal V antenna that picks up high signal voltages from BC signals etc, so rx overload and even damage is a concern. His latest home-built rx is good, but he still looks for improvements, and he focuses on preselectors as a possible way to improve performance. Preselectors don't seem very popular these days, but the author makes a case that at least for his extreme conditions, they can contribute in a meaningful way to how strong, nearby interfering signals can be handled. This turns out to be the case for his homebrew rx and also several commercial rigs, including the K3. The K3 of course has its roofing filter as standard equipment. The IC-751 and an FT-1000 Field have both been modified to add roofing filters that are not part of the original design. The IC-7300 which places its ADC at the front end, cannot be so modified due to its basic architecture, and is tested as manufactured. To test for unwanted response to nearby strong signals the author used 3 types of test setups: IP3 (two input frequencies), continuous noise spectrum (measuring Noise Power Ratio) , and the broadband spectrum carefully generated by pulses, which varies with pulse frequency. The latter method is considered the best, and most similar to real band conditions, and applicable to all the tested receivers (the IP3 method makes no sense for the IC-7300 type architecture). The pulse generator setup has a notch filter that keeps the pulse spectrum out of the rx passband, which is set to 500 Hz. When the power of the pulses is cranked up, the rx does respond eventually, but the higher the pulse power tolerated, the better the rx. The table presented shows "tolerated" (in some sense) pulse power in dBm for the different receivers with and without use of the homebrew preselector, which always helps. 73, Erik K7TV -----Original Message----- From: elecraft-boun...@mailman.qth.net <elecraft-boun...@mailman.qth.net> On Behalf Of Bob DeHaney Sent: Friday, April 06, 2018 12:52 PM To: elecraft@mailman.qth.net Subject: [Elecraft] OE3HKL's Measurements You are correct Breitband means Wide Band in English. Only his home-built receiver is better than the K3 measurements Vy 73 de Bob DJ0RD/WU5T --- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus ______________________________________________________________ Elecraft mailing list Home: http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/mmfaq.htm Post: mailto:Elecraft@mailman.qth.net This list hosted by: http://www.qsl.net Please help support this email list: http://www.qsl.net/donate.html Message delivered to ebasil...@cox.net ______________________________________________________________ Elecraft mailing list Home: http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/mmfaq.htm Post: mailto:Elecraft@mailman.qth.net This list hosted by: http://www.qsl.net Please help support this email list: http://www.qsl.net/donate.html Message delivered to ebasil...@cox.net ______________________________________________________________ Elecraft mailing list Home: http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/mmfaq.htm Post: mailto:Elecraft@mailman.qth.net This list hosted by: http://www.qsl.net Please help support this email list: http://www.qsl.net/donate.html Message delivered to arch...@mail-archive.com