Edit: See my post even further down which features a single bar chart graph showing each radar detector's sensitivity on each of the Ka band frequencies.
I have converted the most recent GOL Ka band on-axis August 2007 test results to actual radar detector sensitivities. To do so, I used the average sensitivity on each Ka band frequency of the three V1s which Michael B. of Whistler tested, and used the average detection distance for each radar detector as reported by GOL.
As per earlier Road & Track test reports from several years ago, I have assumed that there could be up to a 3dB variance in the actual radar detector sensitivities based upon GOL's average detection distances when using radar guns which are obstructed by radar absorbing foam. Anyway, my calculated radar detector dB sensitivities follow. Note that the yellow bar in each of the following images indicates a 3dB window since it is assumed, as Road & Track did years ago, that up to 3dB of measurement error could have occurred.
Here are my results:
Note that there are several radar detector models which, depending on the Ka radar frequency, have extremely good performance which is extremely close to the 3dB yellow shaded window.
Here is a FCC photo which shows the internals of the STi which was submitted for testing and FCC certification:
Note the numbers written in pencil on atop the radar horn. The numbers, from the front end of the horn towards the rear are:
X band -- 119
K band -- 116
35.5 Ka -- 108
34.7 Ka -- 113
33.8 Ka -- 112
The Ka numbers agree very well with my spreadsheet and graphs for the STi. I calculated:
35.5 Ka -- 107.1
34.7 Ka -- 112.5
33.8 Ka -- 110.4
My results are within 2dB of the results which Bel wrote on the STi which they sent in for testing in order to be certified by the FCC.
In conclusion, if GOL's distance measurements were done carefully, then my spreadsheet and bar graphs should be accurate to with 3dB.