TV Signal Levels
What do we mean when we talk about signal levels? There are several 'signals' that can be measured when dealing with CATV and cable based services such as Broadband Internet and IP Telephony. Measuring these levels and knowing their relevance is an essential function of a CATV Service Technician when diagnosing and repairing problems... so knowing a little about them can't hurt if we hope to successfully install or modify our Household CATV Network.
Following is a table listing some of the tests and measurements used in CATV(don't worry, you don't need to memorize this all right now! and much of this is used only when trying to diagnose and fix problems with one or more CATV services):
Testing  Test Equipment  Test Results/Units Measured 

Analog TV Signals Video Audio 
SLM(Signal Level Meter) aka Field Strength Meter  Verifies presence/strength of AM RF(radio frequency) signal* Measures video carrier strength in dBmV* Measures audio carrier strength in dBmV* 
Compressed Digital TV  SLM or DCT Diagnostic Menu 
Verifies presence/strength of Analog RF Carrier* Some meters have a QAM(Quadrature Amplitude Modulation) analyzer built in measures SNR & Correctable/Uncorrectable errors, otherwise use builtin DCT diagnostic tools 
Internet/Data  
SLM  Measures strength of Analog RF Carrier* Some meters measure signal strength required to communicate back to headend(requires test equipment in headend to receive signal from meter) Some meters have Modem/IP/UDP/Ping/Data Rate tests builtin 

Computer w/cable modem  Verify connectivity(http/dns/ip address) Test Latency(ping) aka packet delay time & Packet loss Test bandwidth upload & download speeds 

Telephony  
SLM  Same tests as for Internet  4 important RF readings regarding Internet
& Telephony: [1] Incoming RF signal strength dBmV [2] Outgoing RF signal strength dBmV [3] SNR Incoming [4] SNR Outgoing 

Telephone Buttset  Verifies dial tone/voltage/polarity  
Noise Balance Tester  Measures voltage in VDC, ambient noise in dB, impedance on phone loop in ohms  
Egress(Signal Leakage)  Signal Leakage Detector  Measures strength of leaked RF CATV signal in mV/m (millivolts per meter) 
Premises Wiring  RF Toner  Identifies untagged line/verifies continuity or fault 
*Not necessarily an indication of signal quality 
Whew! That's probably a lot to digest if you're new to the technical side of CATV. I wanted you to get a sense of the complexity of the CATV signal as compared to, say, measuring voltage on a pair of wires. Now let's boil it down to the basics....
Electrical or Radio?
Cable TV signals are a rather unique phenomenon. They are at the same time an electrical signal and a AM RF signal. When looking at a piece of coaxial cable, the electrical part of the signal is traveling down the center conductor(or more precisely on the outer surface of the center conductor). This electrical signal is very weak comparatively speaking... Where your automobile runs at 12 volts dc and your household electrical at 110 volts ac, we are dealing with 1 mV (millivolt), that's 1/1000 of a volt across a 75 ohm load with a power of 13 nW(nanowatt = 1 billionth of a watt). So what good is an electrical signal that weak? Answer: it induces, or drives, the RF signal along the dielectric(insulator) core that separates the center conductor from the outer conductor(shielding) of the coax cable.
Knowing that the RF CATV signal is both electrical and radio, it should now be apparent that a breakdown in the center conductor, outer shielding, or dielectric insulation will have a negative impact on the quality of service delivered. In other words, an impairment in either the electrical or RF portion of the signal can be a problem. More on this in the Troubleshooting TV Problems section.
Bels and Decibels
The one signal that we are most concerned with is the analog RF signal. We could measure the voltage(in fractions of volts) and power levels(in fractions of watts) but this would be cumbersome and impractical. These measurements require the multiplication and division of fractions to calculate gains and losses. Fractions, multiplication and division? yuk! nobody wants to do that stuff, right? So in the early days of CATV, some bigbrained cableguy looked at what test equipment and measurements the telephone business was using... bels!
A bel is simply a ratio of the input to output power. It was discovered that when telephone signals traveled over 10 miles of wire, the output power was 10%, or 1 tenth, of the input power. This attenuation(loss of power) was consistent regardless of the strength of the input signal. In honor of the inventor of telephone, this unit of measurement was name the 'bell'. The 10:1 ratio bel unit turns out to be a little too big to work with in everyday use, so the telephone company divides the bel into 10 equal parts, giving us the decibel, which represents the power loss over 1 mile of telephone wire. Since the decibel is a ratio of input to output power(aka logarithmic number and not linear), we can add or subtract to obtain power gains or losses. For this reason, the decibel was already being used in other electronic and science disciplines, so it seemed like a good fit for cable tv.
Attenuation(signal loss)
Although it takes far less than 10 miles of coax cable to reduce the RF signal by 1 bel, coaxial cable exhibits similar characteristics as telephone wire when dealing with attenuation. The actual length of coax cable required to create a 1 bel loss of RF signal varies depending on the type of coax being used(see below: Varying Attenuation by Coax Type and in CATV Cables and Components)
The output signal is always a given percentage/fraction/ratio of the input signal, regardless of the input signal power. In other words, the input/output ratio is consistent over a given length of coax, no matter how strong the incoming signal is.
Example #1: 

let's say we have a 100' length of RG6 coax. Let's randomly choose 579 MHz, which is just a point in the CATV RF band(generally between 5 MHz and 1000 MHz) to measure at the input and output of this cable. We will get a different number at the input and at the output, the difference being referred to as the attenuation, or more commonly the loss, at the 579 MHz frequency over this span of cable. Let's say that difference is 5 dB, now if we could crank up the input power, we would still measure the same 5 dB loss at 579 MHz. Even though the levels are higher, we still have the same loss over that span of cable. 
Don't worry if you aren't familiar with terms like RG6, these will make sense later.
Example #2: 

Here's another analogy of the power loss/attenuation: we definitely lose more in scenario 1 than in scenario 2, but the input/output ratios are identical. This is how RF signal attenuation works over coax and related devices, both passive and active(more on these in CATV Cables and Components) 
Now we aren't done talking about RF Signal Attenuation just yet... just when you think you understand RF signal loss, we're going to throw a wrench into the whole equation just for fun...
Varying Attenuation by Frequency
In Example #1, just to recap, we lost 5 dB at 579 MHz over 100 feet of RG6 coaxial cable.
Example #3: 

let's say we have 100' of RG11 coax(bigger gauge than RG6 therefore
lower attenuation per foot). Let's stick with 579 MHz, but let's also
measure 108 MHz and 862 MHz as well. The losses(rounded to nearest
half dB) that we should expect to have on 100' of RG 11 are: at 108 MHz ... 1.5 dB at 579 MHz ... 3.0 dB at 860 MHz ... 4.0 dB 
Remember we said that the attenuation on a section of coax would be the same regardless of the input power? This is true as long as we stick with the same frequency at any input power, but when we change frequencies we experience different degrees of attenuation.
In fact, as you can see in Example #3, the higher the frequency, the
greater the attenuation. This is caused by 2 factors:
[1] the science of radio waves; the higher frequency wave(faster oscillating)
uses its energy quicker than the lower frequency wave(slower oscillating).
[2] the higher frequency(faster oscillating) RF signal encounters a greater
degree of impedance from the coax cable.
Varying Attenuation by Coax Type
Larger gauge cables allow for a larger surface area on the center conductor = less resistance to the electrical component of the CATV signal, and a larger dielectric = more space for the RF component of the CATV signal. Both of these factors make for less attenuation per foot. Also affecting the attenuation is the type and quality of material used in the coax cable. For example, two different coax cables that are supposedly both the same gauge(ie. RG59, RG6, RG11 etc...) can exhibit wildly different electrical characteristics  thus providing very different amounts of attenuation. Remember, the three important components of the coax cable affecting signal transmission: center conductor, dielectric, outer conductor(aka shielding). What all this means is: use the largest gaggle coax that is practical, and use the best quality coax available.
Varying Attenuation by Coax Condition
The physical condition of the coax cable and the connectors on either end of the cable also affect the signal attenuation. Two similar cables can have very different electrical characteristics due to poorly prepared ends/connectors, corrosion in/on the cable, mechanical damage such as overstretching, bending, crushing or shorting between conductors. More on this in Troubleshooting TV Problems.
dBmV
Up to this point we have discussed measuring signal attenuation using
decibels... but how do we place a value on a particular signal stength?
We don't use dB, as a decibel is used to measure a ratio(to explain signal
gain or loss) , not to measure a finite amount of signal.
To measure a particular CATV RF signal at a given frequency we use a new unit: the dBmV  the decibelmillivolt. Early on in cabledom, it was observed that a television set required about 0.001 volts  or 1 millivolt  to display a good picture. So 1 mV became the reference point for CATV RF Signal measurement. Since a tv is designed to offer 75 ohms impedance, 1 mV across 75 ohms resistance became the CATV measurement of 0 dBmV. Also important to note is the fact that decibels are logarithmic and not linear; so negative dBmV levels still have meaning.
Remember that gains or losses are measured in dB's and actual power levels
are measured in dBmV's. When dB's are added to dBmV's, their respective
power levels are multiplied. When dB's are subtracted from dBmV's, their
power levels are divided.... what does this mean? Well this takes almost
a whole book to explain, so here's the basics:
 dBmV represents the power ratio of a signal, and only makes sense as
long as the impedance remains constant(this is why all modern televisions
and related equipment have an impedance of 75 ohms);
 an increase of 3 dB represents a doubling of the power;
 a decrease of 3 dB represents a halving of the power;
 an increase of 6 dB represents four times more power;
 a decrease of 6 dB represents four times less power.
 dBmV's do not progress in a linear fashion as they represent a log of
the power ratios
Working with dBmV's
Example #4: 

Let's take our SLM(signal level meter) and have some fun... throw your 28' ladder up against the pole and climb up to the LE(lineextender amplifier) and DT(distribution tap). Plug your SLM into the test port on the LE and 'check the levels'. That means take dBmV readings at several predetermined test points. Let's say that the signal strength is 40 dBmV's at a particular frequency. Then climb down the ladder, walk down to the next LE, throw the ladder up and climb up. Take a reading at this LE. Let's assume a reading of 30. This 10 dB loss represents about a 90% power loss. This 10 dB loss is like losing 9 billion volts in Example #2. Now let's go to the side of the house and open up the CSE. Take a reading at the ground block(ie. the end of the input line); let's assume you get a reading of 10 dBmV. Now go into the living room and take a reading at the cable outlet; let's say you get 0 dBmV. This 10 dB is a 90% loss of power. This 10 dB loss is like losing 9 volts in Example #2. Note that in both cases we lost 10 dB's but the actual amount of power lost is quite different. This is what we mean by dBmV's not being linear. A 3 dB increase represents a doubling of power, but not a doubling of the dBmV reading. 
Real life dBmV readings
Now when you look at a 2way splitter, you will see a 3.5 dB stamped on both output legs and understand why... if you have a certain signal strength coming into the splitter, you get a little less than half of the signal coming out of each leg. Why not exactly half of the signal out of each leg? because no passive(nonpowered) splitter can be 100% efficient; there is always a small amount of attenuation to the signal as it passes through the electronics of the splitter. This is known as insertion loss or through loss. Whether we have a million dBmV's or 1 dBmV coming in, we still only lose 3.5 dB's passing the signal through this splitter. We lose half of the power at a million dBmV's or half of the power at one dBmV, either way we lose 3.5 dB which is half of the signal power, not a finite amount of signal in dBmV's ... this again reinforces what you have learned, that dBmV's are not linear. This is really easier than you might think.... and only requires adding or subtracting small numbers (rather than working with percentages or fractions) to calculate gain/loss of power.
Imagine if dBmV's were linear: if you had 7.6 dBmV's coming into a splitter, you would have to divide that 7.6 by 2(or 3 for a 3way splitter, or 4 for a 4way splitter and so on...) to determine what the output power would be. Now you can see that working with a logbased counting system can actually make things easier in the everyday world of cable! Those crazy mathematicians might be onto something after all!
Summary
There is a lot more to all of this power/voltage/dBmV ratio stuff, but frankly you don't need to know all of the ins and outs unless you are a CATV Technician(watch for us to offer a course on this subject in the future)
After all of this, we have only discussed measuring analog RF CATV signals. We have not touched on Compressed Digital Signals, QAM analysis, measuring HUM/Distortion/Noise/Egress, analyzing bitstreams and bitrates, signaltonoise ratios and more.... but don't worry, much of this falls into place when good RF levels exist coupled with proper craftsmanship and good quality cabling and equipment, besides, you need specific CATV TEST EQUIPMENT in order to quantify many of these components of the CATV Signal.
Now that we have beaten this Signal Level horse to death, let's take a look at CATV Cables & Components.