Simple answer: NM=Rx Power - (-91.0). You can simplify that to be NM=91- absolute value (RX power) for RX Power < 0. You'd have to be less than a mile from a 1 MW station for Rx Power to be calculated to be greater than 0.
Signal power and ATSC:
Doing some math (elsewhere), it can be calculated that the theoretical noise floor for a 6 MHz-wide channel is -106.2 dBm. ATSC specification requires a signal to noise ratio of 15.2 dB. Therefore, in a theoretical circuit, at the input to the decoder portion of the tuner, there must be a signal power > -91.0 dBm. The site's software establishes this as a 0 dB Noise Margin (NM) in that there's no margin for any additional reduction in signal power as that would cause decoding to fail.
See FAQs for expansion of below.
http://www.tvfool.com/index.php?opti...d=57&Itemid=78
Quote:
|
These values tell you if you are above or below the detection threshold for each station and by how much. Since these values represent the amount of signal "in the air" at your location, you need to have enough margin to account for building penetration, cable loss, splitters, tuner sensitivity, and other factors specific to your setup. If you take the initial NM value for a given channel, add your antenna gain, subtract all the other system losses, and still end up with a value above 0, then you should be able to detect that channel.
|
Emphasis added to the above quote. One factor not emphasized enough in the above is the tuner's sensitivity known as its 'noise figure'. For a non-amplified system the noise figure of the system would include EVERYTHING from the antenna's terminals to the decoder chip inside the tuner. Balun loss, cabling loss, connector loss, splitter loss, and the input to the tuner (noise figure) must all be summed and then subtracted (by the end user) from the calculated noise margin. As an example, assume a balun loss of 1 dB, 100' of RG6 (about 5.5 dB @ 700 MHz, it's less at lower frequencies), a two-way splitter of 3.5 dB insertion loss per leg, 0.5 dB of connector loss, and a tuner with a noise figure of 6 dB at the channel of interest. That means the system noise figure at the decoder is 16.5 dB and that you need a NM of 16.5 or higher in order for reception to take place. This is without any adjustments for antenna gain (antenna gain improves NF if gain is positive), amplifiers, or real-world signal impairments such as an elevated noise floor due to background noise, multi-path, penetration losses due to trees or building materials, and the like.
For an amplified system, the noise figure of the amplifier dominates the setting of the system noise figure. If you have a preamp with a 2 dB NF, a balun with 1 dB loss, and miscellaneous connector loss of .5 dB before the amp, the noise floor is raised by 3.5 dB at that point, all assuming that the gain of the amp is sufficient to cover the downstream losses by several dB. At the input to the decoder, the noise floor is a bit higher, you'd have to run the cascaded noise figure calculations to get a final estimate.
I've found that 1-edge and 2-edge calculations in the simulator can be wildly inaccurate. My field tests have show results that vary by as much as 20 dB for 2-edge distant signals. For signal locations that "right behind" a significant terrain object like the back side of a hill or mountain, they're next to useless. Therefore, take those types of calculations to be worth less than a grain of salt. LOS signal calculations are usually close enough for government work.
If you Have ANY impairments such as trees or buildings, the calculations cannot include them so any number calculated must be reduced by some "fudge factor" which one can develop after a while.