Although the hardware frequency counter can be wrong under certain circumstances, when I compared mine using the output of a calibrated function generator that I borrowed, it was spot on to the last decimal place.
Hi bitseeker, what I was referring to in that post was the high frequency of 440 MHz, which cannot be measured by the hardware freq. counter.
I just did an experiment, which I hadn't tried ever before, to see how high the hardware and "software" freq. counters could be usable.
I found that the hardware counter basically stopped being usable at 103 MHz, so it's in some way still tied to the "official" or hacked BW. After 103 MHz, the readouts it gave were non-sensical. I was feeding the scope about 440 mV RMS from my generator. Maybe someone else can confirm this.
But the software freq. calculation was working (roughly) up to 440 MHz. It ain't a very accurate readout (+/- 5 MHz) at that high a frequency, but it was hanging in there.