To give an illustration of the Sherwood Performance Index (SPI) and it’s relationship to Rob Sherwood’s Narrow Dynamic Range measurement, consider this scatterplot of the two. The SPI is the first Principal Component (linear composite) of ALL of Rob’s nine bench measurements. I view it as a broader index of the bench tests whereas Rob, who is an ardent CW contest operator, ranks his famous Table on the single measurement of “close in” or narrow dynamic range in reception. So, if you are not a CW contest aficionado, the SFI might tell you more about the fuller receive performance of a particular rig.
The scatterplot below is responsive so you can hover over each rig’s point and identify that rig. The R2 for the equation is .768, suggesting that about three-fourths of the SPI is shared with the narrow dynamic range. The Mean Square Error (MSE) is 52.28 and it’s square root is the standard deviation of the rig points round the linear regression list (or how variable they are from the predicted value). This value is 7.23 around the SPI. So there is a close but not perfect association between the two indices. But this chart reveals other things upon interacting with the data points.
If you consider that the linear line is a demarcation of what the full set of Sherwood’s bench tests versus what only the narrow bandwidth dynamic range tell us about each receiver, notice what systematically clicking on each data point above and below that line reveals.
Let’s start at the low end of each test range. Both the Kenwood R-2000 and the Ten Tec 340 rigs have the worst narrow dynamic range (< 50dB). But the Ten Tec has an SPI of about 85. The Ten Tech 340 is a commercial-grade Rx with extensive digital signal processing whereas the venerable R-2000 was innovative for it’s day but has far fewer design features and amenities than does the RX-340.
Moving to the mid-points on this graph, the Icom 7800 has a narrow dynamic range of about 80 with an SPI of 117. Compare this to the Yaesu 901-DM with the same narrow dynamic range but an SPI of less than 90.
A final comparison at the high end shows that the Flex 6600M has a narrow range in the high 90’s with an SPI of almost 140. The Apache Anan 200D has an almost identical narrow dynamic range but only an SPI of 110.
What do these comparisons have in common? At least two things come to mind.
One is that some were engineered during a previous decade in comparison to their bench-measured peer rig. This may entail the significant changes in engineering design, production scale capacity, and incorporation of newer technologies for receiver performance.
But a second is that those above the linear regression line are fuller-featured radios with a more complete package that appears to perform better on Rob’s bench test suite than the comparable one below the line. Think “comps” in a house purchase evaluation to determine sales market value except here we simply do not have all of the features identified or measured for this set of rigs. I can’t validate that interpretation form these data alone. But it is a systematic pattern in this scatterplot that is captured by the fuller SPI measurement over the critical but singular narrow dynamic range alone.
I hope that helps you better understand what Rob’s incredible set of bench tests reveal about these rigs, covering over 50 years of amateur radio.