On Sun, 3 Aug 2014 20:28:38 +0100, Ian Jackson
wrote:

[I, for one, can't see any point in using large numbers when small
numbers will do just as well.]

I presume you haven't met the marketing people, who believe that
larger numbers are always better.

50/75 ohm matching pads (and vice versa) can sometimes be more of a
hindrance than a help (the 'horses-for-courses' principle needs to be
sensibly applied).

Finding where someone hid the pads is often the biggest problem. There
usually aren't enough around the various labs for every piece of
equipment.

Of course, I made things worse. I got tired of subtracting 5.72 dB
for the usual minimum loss L-pads, so I made a few 10.0 db Pi-pads.
For the curious, that's 77.1 ohms in, 207 ohms in series, and 87.1
ohms out. Roll your own:
http://chemandy.com/calculators/matching-pi-attenuator-calculator.htm
At the time, I had access to a laser trimmer, so I bought some
commodity parts and trimmed them to value. The 10.0 dB loss made it
much easier to do the math, as long as I didn't do my usual order of
magnitude screwup. However, these only added to the confusion, with
two types of matching pads in circulation. Moral: I can't win.

In practice, dBm (or the rare dBmW) generally implies 50 ohms to the RF
guy and 600 ohms to the audio/phone guy, whereas dBmV and dBuV generally
implies 75 ohms. However, as usual, it usually depends on the context,
and it's a case of "if in doubt, tell it as it is".

Except for FM tuners, which use dBf (dB above 1 femtowatt).

One oddity I found (many years ago) was a piece of cable TV equipment
had one RF parameter was specified in dBm, while all the others were in
the usual dBmV. This was not a mistake, and the reason given was that
the dBm signal was CW, while the rest were vision and sound IF and RF
signals. I hope I may be forgiven for not understanding or agreeing with
the reasoning for this decision, as not only was it unnecessary, but it
could also have been considered a typo, and caused confusion.

I don't understand their logic either. However, I can sorta see it
with a spectrum analyzer. Maybe the signal levels were in peak power
(dBm) where one needs to adjust the value depending the modulation
bandwidth. For example, a flat 0 dBm channel signal on a spectrum
analyzer, that's 6 MHz wide, using a 100 KHz analyzer bandwidth, needs
to have the 0 dBm adjusted by:
10 * log(6000/100) = 17.7 dBm average power.
No compensation is needed with a CW signal, where the peak, CW, RMS,
and average levels are identical, thanks to zero bandwidth.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558