spiney
Guest
- Joined
- Apr 29, 2005
- Messages
- 1,514
- Reaction score
- 1
- Points
- 0
- My Satellite Setup
- Pace 2200 Sky digibox with ftv card, Comag SL65 FTA sat receiver, 40cm Sky minidish, Setpal terrestrial receiver (for free uk tv only!).
- My Location
- Midlands
CAVEAT EMPTOR (bit 'o Latin, "let the buyer beware"!). I've got an "installing digital tv yourself" guide, from a major DIY retailer. It's complete rubbish, from beginning to end, so be careful!
(also see: www.megalithia.com/elect/aerialsite/ . Very good, recommended!).
My "newer version" Freeview guide follows ......
FREEVIEW RECEPTION INFORMATION ON THE INTERNET.
There's two main "postcode predictors":
The "official" one, at www.freeview.co.uk .
An "unofficial" one, at www.wolfbane.com/cgi-bin/tvd.exe? .
The official one gives a simple "yes/no" answer, and is "conservative" (possibly for legal reasons!), ie, you may well get Freeview in many places it claims you can't, but unfortunately no further info is given, it doesn't tell you whether you're "only just" or "a long way" outside the reception zone!
The Wolfbane site gives far more detailed info - predicted field strength from and other stuff about the nearest transmitters - but the accuracy is unknown. It possibly can't take into account the terrain between transmitting and receiving aerials, and the directional information may not be accurate.
(It's unwise to say what's a "usable" signal (!), but most Freeview receivers should be "ok" on 30uV/m, and 50 is "good").
At best, using both these together gives a only a "very rough idea". By far the best guide - obviously - is other nearby dwellings. If some neighbours can get Freeview (ask!), then their aerials are a good guide to what you'll need. Is reception generally fine on existing older analogue aerials? (then, you're lucky!). Or, is a very large aerial needed, possibly with a masthead amplifier (and reception still dodgy!)? Or, is reception still completely impossible in your area (in many it is, and in some it will always remain so!)?
Note that, common sense applies! Despite some extraordinary claims originally made for digital terrestrial tv, generally speaking the digital aerial needs to be at least as good as the previous analogue one (and often significantly better!).
The following link has some info on digital tv technical principles, and possible reception problems, with (limited) practical advice:
www.dtg.org.uk/publications/books/r_book2.pdf .
More generally, the above dtg website - www.dtg.org.uk - is an excellent information source, having various technical stuff, and comprehensive web links.
For more detailed technical information on UK digital terrestrial tv, Google search for "MPEG2" (video and audio compression details), and "COFDM" (the transmission system used).
An "official" list of the actual DTT transmitters is at:
www.dtt-tx-info.org .
There's a more informative version here:
www.dtg.org.uk/retailer/transmitters.html .
(once again, it's actually the dtg website!).
GETTING FREEVIEW, WITH "THE RIGHT STUFF" (ie, equipment!).
1) Digital tv is a "different animal" from older analogue tv. Generally speaking, within a reception area you get it (perfectly!), and outside you don't (and no use trying!). That's oversimplifying, but basically true.
2) With an extremely strong signal, ie, in Greater London, or central Birmingham, or very close to (within sight of) most transmitters, DTT (Freeview/Topup) receivers should work on "just a bit of wire" stuck into the aerial socket.
Slightly further away, you may still be ok with a set top aerial and "signal booster" amplifier. However, there's no guarantee these will work, and better ones (well designed aerial with low noise figure amp) are more expensive, so some "financial risk" is involved!
(note that - on an indoor aerial - DTT reception might be good, but the picture can still occasionally "jump" when people move around nearby).
Then, at slightly further away still - and throughout most of the remaining "possible reception zone" - a roof aerial will be needed.
3) After which, it's a matter of getting the appropriate roof aerial:
a) You may be able to get the newer digital signals on an exisiting roof aerial, previously meant only for analogue ones. If so, then that's fine (and you're lucky!). Otherwise:
You need the correct frequency band, for the transmitter you're using, so check. Also, some transmitters have the newer Freeview signals partly "out of band" - ie outside the "frequency plan" previously used for just analogue tv - in which case you'll need a new wideband aerial; again, check!
The UHF bands are A, B, C, D, E . These days, aerials tend to be AB or CD or E. Unfortunately, for the same gain, wideband aerials are larger and heavier!
c) You need sufficient aerial gain, to get a high enough signal (digital tv being "all or nothing", the signal being "not quite good enough" means you get nothing!).
Generally, the biggest available aerials have a gain of 18dB (that's "100 elements up", for higher gain the number of metal rods starts to get impractical!).
For some idea of different aerials available, see:
www.tlc-direct.co.uk/Main_Index/Aerials_Index/AerialsTv/index.html .
Also: http://www.blake-uk.com/prod_products_aerials_tv_menu_selection.aspx .
The highest gain aerials available are ones like the DAT 75 (19dB max). See www.hdtvprimer.com/ANTENNAS/DAT75.html .
For higher still gain, it's most usual to use a masthead preamplifier. Otherwise, you can combine single aerials together, but this needs specialist knowledge (and you end up with "lots of ironmongery" on the roof!).
However, at distances where such high gain is needed, it's usually a matter of fringe reception conditions, an the signal may well be affected by diffraction across mountains, differing day/ night propogation, etc, so things can get unreliable (although, you might still be relatively lucky), for particular locations it's very hard to say!
(sometimes you see a different aerial type, several stacked parallel rods in front of "chicken wire". This is low gain - for local transmitter - but with high off-axis rejection, useful if there's strong interference).
(Also, there's the wideband log periodic types, characterised by a centre "double strut" and roughly triangular shape; sometimes used on caravans, but not recommended for roofs).
4) "Digital" aerials tend to have extra gimicks, eg, things like "ghostkiller" directors and larger reflectors. While these may be of use with some digital tv reception problems - eg co-channel interference in some regions, or impulsive interference - for most people the main "reception problem" is achieving sufficient gain!
5) Then you need the appropriate connecting cable! Old style "UHF" coaxial cable - previously used for analogue aerials, and widely sold in DIY shops - is "usable", but satellite grade cable is better! For 2 reasons: a slightly lower signal loss at higher UHF frequencies, also the continuous foil outer condutor (which better resists impulse interference, very much a digital tv "known problem"!).
(Note that, ideally this should be moisture proofed (foam dielectric) type, as the "satellite cable" sold in DIY stores is usually RG6u, usable, but with an air-spaced interior which will gradually absorb moisture and slowly degrade if used outdoors. Not as serious a problem at digital terrestrial tv frequencies as it would be at satellite ones, so can be used, but a waterproof type would be better!).
6) The tv receiver can also make a small (but again important, it's "all or nothing"!) difference. Most DTT receivers claim a (working) input sensitivity level of - 75 dBm (see decibels explanation). Some cheaper sets are not as good, so a "known name" brand increases the chance of better reception (for example, maybe avoiding the need to upgrade and older existing aerial!).
(The specially designed Novapal tuner, as used in Setpal - and some other - receivers, has a sensitivity of -83 dBm, the best there is! However, Setpal are now defunct (although still available as I write this, eg, www.cpc.co.uk type "digital tv receiver" in searchbox, also in some Tesco stores)).
DIGITAL RECEPTION, A FEW PRACTICAL TIPS!
1) Once you're sure about which "bands" aerial to use, go for the highest gain one you can find. Many digital tv reception problems come from an inadequate aerial, and you're extremely unlikely ever to get "too much" signal (and - in this very unlikely event - UHF attenuators are very cheap!).
2) If considering re-using your exisiting analogue aerial for digital reception, then if possible first briefly borrow a neighbour's Freeview receiver, and try that on the aeiral (use the existing stored channels, don't re-scan!). This will give you some idea. If the reception isn't good enough, is it a long way off, or "only just insufficient" (ie, intermittent partial reception with pixilation - "blocking" - maybe ob just some channels. If the former, then it's a new aerial needed! But, if the latter, then you just might get away with a "signal booster". A "small financial risk" (or zero, if you can borrow one to try!).
3) When adjusting the aerial, what you're after is the minimum bit error rate (from Viterbi decoding)! This is shown by the "quality" reading on the signal strength display page (available on nearly all Freeview sets!). Usually, some muxes (groups of channels) - particularly ITV/ Topup ones - will be worse than others, and you should try for the best quality reading on the worst received channels (then, the "stonger" channels should be ok too!).
(note that, high signal quality means low received error rate (!), so if your receiver shows a number as well as a quality bargraph, then as quality goes up this number will go down!).
4) A particular "known problem" with DTT is impulse interference (due to how COFDM works, Google "cofdm impulse interference" for further info). If you have this, the picture can keep freezing, or just disappear. Mainly due to nearby electrical devices, especially motors and/or flourescent lights elsewhere in house, possibly cars passing by just outside, etc.
Here's where "digital aerials" score! The higher directivity of "crossed" directors - and a bigger reflector - are some help. Also, a balun is a good idea, possibly one at both ends of cable (digital aerials often have one built in, check the specs!).
A further "tactic" is to try and partially screen the aeiral from interference sources. This is already partly true of any roof aeiral (ie, screened from any ground directly beneath!). However, it's also possible to "put brick" - eg, the side of a house - between aerial and a known interference source. Another ploy is to slant aeiral slightly upwards, further decreasing pickup from the ground below!
ANALOGUE AND DIGITAL TV, RECEPTION DIFFERENCES!
As the transmitted tv signal spreads outwards from the transmitter, it gets weaker with increasing distance away.
(At any instant, think of total signal energy as contained in a thin spherical "shell", slowly expanding outwards. As this gets bigger, signal energy gets "more spread out" - across the shell's surface - hence it's weaker. From simple goemetry, surface area of a sphere is 4xPIx(radius squared), so signal power actually falls off according to the inverse square of distance from transmitter. Note that - if using measuring equipment - electric field stength only falls off as inverse distance (not squared!), but it's the power that matters!
(Across entirely urban areas, scattering from buildings changes this to an inverse 4th power law! However, except for maybe in London, this wouldn't apply to digital terrestrial tv in most of the UK!).
As the tv signal weakens, it's increasingly affected by "electrical noise", at first picked up via aerial (from various sources), and -ultimately when the signal gets too weak to use - from devices (transistors) inside the tv receiver itself!
There's also radio wave propogation, the UHF terrestrial tv frequencies - pretty much! - travel in a straight line, so reception starts getting difficult "below the transmitter horizon" (when the receiving aerial can no longer "see" the transmitter, due to earth's curvature). However, distance weakening effects are usually more important, "kicking in" before the horizon effect does!
As signal weakens, the receiver agc (automatic gain control)compensates, but the relative noise level rises. As we say, the signal to noise ratio (SNR) worsens (gets lower!). This affects analogue and digital tv reception quite differently:
Analogue tv - the video signal voltage represents picture brightness (ignoring colour here!), and receiver timebase "synchronising pulses" are added. This (vestigal sideband) amplitude modulates a carrier signal - so the transmitted "wave shape" closely follows picture brightness - then at the receiver the original video is (fairly easily) extracted.
As noise levels increase, the picture gets ever more "snowy" - and faint - but you can still make it out, even when quite bad! And the receiver timebases will continue locking into the sync pulses - even if they're badly distorted. So, even when reception gets fairly bad, usually you can still watch/listen.
(In the jargon of Information Theory, we say analogue tv is "highly redundant", ie, the signal can be hugely distorted but remain "watchable", as long as the receiver timebases can still "lock in", with the human eye/brain then "integrating" across many picture frames to compensate for picture noise ("snowiness")).
Digital tv - is very different! The picture is coded as numbers, which are transmitted as on/off pulses (ones and zeros). To decode the picture, nearly ALL these pulses must be correctly received! However, as the relative background noise level rises, there's an increasing chance of a "1" becomming a "0" (or vice versa!). The solution is the one always used in digital telecoms, have an error correction system! For digital tv, several methods are combined - bit interleaving, convolution, galois fields - with the correction information then being added to data as extra bits ("Forward Error Correction", ie, it works one way only, there's no "backwards path" as in computer communications, which - for example - can allow modems to slow down under "very noisy" conditions!).
This means that - as the signal weakens and the (relative) noise level rises, the error correction system "works increasingly harder" - correcting more wrong bits - but decoding still works properly, and the picture remains "perfect"! Up until a certain point, where the noise becomes so bad that - quite sudenly - decoding is no longer possible, and the picture "completely freezes" (again, I'm oversimplifying, but that's basically correct!).
So, we get the famous "digital cliff" (or "sudden fallover") effect! In fact, it even has a specific number, a ratio of 17 dB ("Rayleigh" modal value to variance ratio), or approx pre MPEG decode (post Viterbi) error rate of around 1 in 1000.
(correction - this magic figure is actually CNR, carrier to noise ratio, but it effectively becomes the above in very bad multipath conditions).
(In actual fact, this "digital cliff edge" can be rather wobbly! For example, day/night radio wave propogation can slightly differ, so you might get some digital channels by night but not by day (or vice versa!). If "only just on the edge", then there will be pixilation - the picture breaking up into individual squares - since only some "I" frames and partial P/B frames will get decoded).
(In Information Theory jargon, digital tv has "low redundancy", ie there's not much "room for error", and a just relatively small proportion of bit errors will make pictures/sound become completely undecodable!).
A WORD ABOUT DECIBELS (d.
An amplifier multiplies, ie, the output is so many times bigger than the input (ideally, that's ignoring various distortions!). So, for signal output level, you just multiply input by amplifier "gain"
A tv "receiving chain" - aerial, cable, various amplifiers (including tv receiver) - is basically many amplifiers cascaded (something that loses signal, eg cable, is just a "negative gain amplifier" with gain figure less than "one"). So, for total system gain you multiply together the individual component gains.
Doing lots of multiplications is awkward (even with a calculator!). So, usually, gains are first re-expressed logarithmically. Then, to multiply, you just add them (if anyone can remember "school maths"), which is much easier!
gain in dB = 10 log (base 10) actual power gain figure.
Roughly, x2 = +3dB, x10 = +10dB
(So, "x10x10x10x10x2x2x2" becomes "+10+10+10+10+3+3+3" instead, giving 49dB instead of 80,000; the dB version is easier to handle, and mistakes are less likely!).
(Note that, for voltages, the above formula value must be doubled (!), but tv signal strength is normally expressed in power ratios, often relative to dB(m), a fixed power level of 1mW. For example, Freeview receiver working sensitivity - when given at all! - is often -75 dBm).
SOME "EASY" ANTENNA THEORY (!).
From above, for digital reception it's clearly important to get a good signal to noise ratio (SNR). That's the job of the antenna.
The theoretically simplest -and physically unrealisable! - antenna type is the "isotrope", ie, "receives equally strongly from all directions". This has an "effective aperture" of: (wavelength squared)/(4)x(pi) . As you can see, this formula gives an area (!), in fact the antenna's "capture area"; it "collects" radio waves from this "cross-section" of space. So, the bigger this aperture is, the stronger the received signal will be.
But, wavelength is inversely proportional to frequency! So, as transmission frequency rises, the aperture falls, making it increasingly difficult to get a strong (good SNR) signal.
So, at AM frequencies, a single long wire will do, as used in portable AM radios and by radio amateurs (I'm cheating, the aerials inside AM sets are magnetic, but same principle applies!). Going up to FM/DAB, a single rod aerial on the set can suffice, but one on the roof is better, possibly having 3 or 4 elements. With higher terrestrial tv frequencies, a multi-pronged (Yagi) aerial on the roof is nearly always needed. Finally, for sat frequencies, we need a dish to collect ALL the signal from a certain area, dish size rising with falling signal strength (as readers know).
The simplest "realisable" (and usually used!) antenna is the Dipole, a metal rod one half wavelength long, with feeder cable connecting to a small gap in centre. This is "resonant" (not all antennas are), giving it several advantages:
a) directional, a null along the axis but 1.6 dB gain (over isotrope) and
bandlimited, it only responds to a a limited bandwidth around the centre resonant frequency.
Both these features help increase gain of wanted signals, and limit response to unwanted ones, exactly what you need for digital reception (for good SNR)!
Note that, it's the centre gap and feeder cable that makes the metal rod into a dipole. Otherwise, it's just a metal rod, and doesn't do anything! (you need a gap and cable to "extract" any received radio energy). However, if several parallel rods are placed sufficiently close together, then - by near field induction - some radio energy can "bleed across" between them. Then, they all "partially" behave as (limited) dipoles - each passing a bit of energy on to the next - until the final (actual) dipole, which transfers the total collected energy into its feeder cable.
Each rod "linked" this way to the main dipole transfers a bit more energy, and the cumulative effect is to gradually increase the aerial's total "aperture", ie to create a much larger "area" over which it can "collect" radio energy, hence improving the SNR.
This is the basis of the well known Yagi (inductively linked array) tv aerial. The supplementary dipoles are called "directors". Then, behind the main dipole is a "reflector" (by adjusting length/position, this feeds back to main dipole in antiphase, partially cancelling signals received from that direction. Modern aerials often use a mesh reflector, rather than just a rod, for greater effect).
But - roughly speaking - each further Yagi 3dB power gain requires (approx) doubling the number of "elements" (rods). However, we've said that transmitter power falls with inverse square of distance! So, it's a "law of diminishing returns", and with increasing distance the number of elements needed rapidly becomes impractical (from increasing weight!).
(Pictures: see www.sub-tv.co.uk/antennatheory.asp ).
(more theory, see: www.ee.surrey.ac.uk/Personal/D.Jefferies/yagiuda.html ).
IMPEDANCE/MATCHING.
With antenna systems, component sizes (aerials, cables, etc) are of comparable size to wavlength, hence transmission line theory applies, and antennas/cables/other devices must "match" in impedance value (otherwise, energy reflection occurs at mismatched interfaces, and you get analogue ghosting and digital increased error rate).
It just happens that a dipole has an "equivalent impedance" of 75 Ohms (another reason for using it, as energy losses along the connecting cable then become insignificant, non-resonant antennas have lower impedance and therefore higher cable losses!).
So, tv coax is made with 75 Ohm impedance (Ethernet/transmitter coax is often 50 Ohm, for different reasons).
(Yagi antennas use folded dipoles of 300 Ohm impedance, and the other added elements bring this down again to 75 Ohms).
(more, see: http://www.moonraker.com.au/techni/impedance.htm . This describes signal transmission, but applies equally to signal reception!).
COMBINING SEVERAL ANTENNAS.
(Before attempting this crazy stunt (!), it would be a good idea to check you're actually getting an identifiable digital mux signal, although at too low SNR for decoding. Remember, "doubling the metalwork" only increases received signal by 3 dB, and using 4 aerials only by 6 dB, that's not much! Although, given the "digital cliff", it could make the vital difference. Chance of success also increses with the best available tuner, currently still the Novapal one inside Setpal receivers, or the Echostar T101 if you can find it).
So, we've said a single Yagi has a practical size limit - due to increasing weight - the biggest being the above DAT 75 (really 3 Yagis stacked vertically, to reduce bending moment on chimney!)
If that's not enough, then - subject to there actually being a signal, albeit too weak for digital reception! - can we combine several Yagis for even higher gain? Yes, but it's a bit involved!
1) A Yagi works by "nearfield induction" between the elements, but we don't want this also happening between different Yagis! (if so, then they "take" some proportion of signal from each other, instead of from the wanted transmission). From theory, induction field falls to equal radiation field at distance: (wavelength/2xpi), which at UHF tv wavelength of around 0.4m = (v. approx) 7cm. Using the engineer approximation of "multiply by 10", then, multiple separate combined Yagis should - ideally, not always possible! - be 70cm or more apart.
2) As said above, maintaining the correct impedance is important, so you can't just "join the wire ends together"! Instead, use an aerial combiner. Unfortunately, commerically available ones are usually for combining differently banded aerials - hence have built in filters - or are for tv/fm etc. In which case, use a widband cable splitter "in reverse", 5-2400 Mhz, available 2-6 output versions:
(eg, see www.showmecables.com/showProducts.asp?category_id=414 ).
(that's a USA website, but similar splitters are widely available! Note that UHF tv only devices 5-1000 Mhz are fine, but sat only devices - 900 - 2400 Mhz won't work at UHF tv frequencies!).
(Also note, the "power pass" feature - on some/all/none splitter/combiner ports - is usually irrelevant, since dc is not fed along UHF cables, unless you've got a masthead preamplifier for which see later!)
3) This is the really tricky bit! OK, we've got "coherent" signals from a single source, so just adding them together will increase signal strength and - vitally for digital - the signal to noise ratio. But .....
We must add the signals together in phase (preferably exactly, or at least as nearly that as possible!). Otherwise, as they go out of phase, some cancellation will start to occur, effectively lowering the strength you're trying to increase. By a certain point, you're back where you started. After which, it gets worse, and at a half wavelength difference you get total signal cancellation, ie nothing!
Now, wavelength in UHF tv band is (approx) 0.6m. So, a "path difference" of half that - 0.3m - will give complete signal cancellation! Actually, to avoid significant losses you want much less than that, so let's say the difference in any cable lengths from aerials to combiner must be under 10cm, that's not very much!
But it gets even worse! Radio wave speed in coax cable is less than in "free space" - possibly sometimes just half - then the cable length difference goes down to just 5cm, before "significant" combiner losses start occurring!
So, yes it can be done, but must be very carefully!
(It's possible to "null out" any phaseshifts with r.f. engineering techniques, but this requires taking measurments and placing phase shift "stubs" in the right places, not to be considered unless you're already an r.f. expert!).
AMPLIFIERS AND "SIGNAL BOOSTERS".
Any aerial amplifier at all has electrical noise at its output, at a level far higher than any "atmospheric" noise picked up on a tv aerial. This will then create a new "noise floor", thus decreasing the sig to noise ratio of any digital tv signal passing through it (from above, exactly what we don't want!).
(Any transistor has thermal input noise - across its input resistance - and amplifies this by its own gain, then adds further "transistory" type noise, like "shot", "flicker", etc. If there's several transistors in "cascade" - or if an integrated circuit amplifier is used - then the noise is further amplified several times!).
(see: http://www.cisco.com/en/US/products/hw/cable/ps2209/products_white_paper0900aecd800fc94c.shtml .
Exactly the same principles apply to tv aerial amplifiers, where a tv aerial is a "noise source" just like a resistor!).
(If you are a keen masochist, there's also this: www.ieee.li/pdf/viewgraphs_mohr_noise.pdf ).
Usually, an in-line aerial amplifier will have a specified "noise figure"; then the SNR (or CNR, whichever you prefer) worsens by this amount. So, for a 26dB sig-to-noise signal going through a 5dB noise fig amp, SNR then degrades to 21 dB (aren't dBs nice and easy!).
(Note that the concept of "noise figure" is just a "calculating convenience", suggesting a ratio, whereas amplifier output noise is actually fixed (with some temperature dependance). However, it works well enough at the "expected" input level).
Well, ok then, aerial amplifiers always make the sig to noise ratio worse, which is bad for digital tv, so why would we ever use them? Well, there's one or two special situations ....
1) Received field strength not high enough to operate DTT receiver. And, there's 2 sub-cases of this:
a) Using a set-top tv aerial in a strong signal area, close to transmitter. This type of aeiral is very poor, and despite a strong signal may easily still not give enough micovolts at receiver input. In which case, an amplifier might just do the trick. It's going to make the SNR much worse, whatever else, but the local field is high enough for this still to be usable.
The amp "built into" most set top aeirals - or an added "signal booster" - is likely to be dire, with a noise figure of 6dB or more, so we're worsening ("throwing away"!) the received signal by at least that much, but if close enough to the transmitter you might get away with this!
In a fringe reception area, already using a high gain aerial (much trickier!).
In this case, the amplifier design is critical, since yes we can amplify, but anything that also worsens the SNR is bad news (this time, we can't afford to "throw away" some proportion of the SNR arriving at the receiver input!).
Looking at the above "noise" links, both give (slightly different versions of) the standard formula for noise in amplifiers in cascade. It can be seen that the first amplifier in any such "chain" is always the most critical, making the biggest single increase in the overall noise figure. In this case, that would be the aeiral amplifier!
So, what we have to do - for any chance at all of improving digital reception - is use the best possible type of masthead amplifier, for example see:
www.jwhardy.co.uk/shop/pages/pl_amp_mast.html .
A "good" noise figure would typically be about 1.9 dB, for UHF tv band only (you don't want wideband!). This is about the limit of what's reasonably possible, using just a few discrete components - not integrated circuits! - in simple well designed circuit configurations, for the lowest possible added noise (don't bother buying a slightly cheaper amp with worse specs, you'll be wasting your time and money!). Note:
1b1) A masthead amp should go as close as possible to the aerial. Because, you want the best achievable sig to noise ratio. And, as you move away from the aeiral, the received signal goes down, but amp noise output remains fixed, so SNR worsens! Yes, ok, you can use a masthead amp indoors - that's still better than a cheap "signal booster" - but will give a worse result than on aerial.
1b2) Get the highest possible gain - up to 30 dB - because you want the SNR as good as possible. A higher gain amp may have some spurious intermodulation products - especially if poorly designed - but this is not dissimilar in effect to multipath reflection, and COFDM can cope with that! But, it can't cope with a too low SNR (too high "noise floor").
1b3) Of course, a masthead amp also needs the appropriate power supply, usually around 12-15v. Just 2 or 3 transistors in the amp take a "trickle charge" of only a few milliamps, unnoticable!
1b4) If combining several aerials - as above - then the single masthead amp goes AFTER the combiner! In which case, whether or not the combiner has any "power pass" sockets is still irrelevant!
2) The other case where a masthead amp is used is when there's local electical interference. But - see above - using sat grade cable and positioning aerial "favourably" should screen out most of this. A UHF tv aerial is band limiting, as is the if stage(s) in the receiver, both together greatly reduce the received noise. It's also possible to add a cheap UHF bandpass filter between the aerial and receiver input, for even greater noise rejection.
(also see: www.megalithia.com/elect/aerialsite/ . Very good, recommended!).
My "newer version" Freeview guide follows ......
FREEVIEW RECEPTION INFORMATION ON THE INTERNET.
There's two main "postcode predictors":
The "official" one, at www.freeview.co.uk .
An "unofficial" one, at www.wolfbane.com/cgi-bin/tvd.exe? .
The official one gives a simple "yes/no" answer, and is "conservative" (possibly for legal reasons!), ie, you may well get Freeview in many places it claims you can't, but unfortunately no further info is given, it doesn't tell you whether you're "only just" or "a long way" outside the reception zone!
The Wolfbane site gives far more detailed info - predicted field strength from and other stuff about the nearest transmitters - but the accuracy is unknown. It possibly can't take into account the terrain between transmitting and receiving aerials, and the directional information may not be accurate.
(It's unwise to say what's a "usable" signal (!), but most Freeview receivers should be "ok" on 30uV/m, and 50 is "good").
At best, using both these together gives a only a "very rough idea". By far the best guide - obviously - is other nearby dwellings. If some neighbours can get Freeview (ask!), then their aerials are a good guide to what you'll need. Is reception generally fine on existing older analogue aerials? (then, you're lucky!). Or, is a very large aerial needed, possibly with a masthead amplifier (and reception still dodgy!)? Or, is reception still completely impossible in your area (in many it is, and in some it will always remain so!)?
Note that, common sense applies! Despite some extraordinary claims originally made for digital terrestrial tv, generally speaking the digital aerial needs to be at least as good as the previous analogue one (and often significantly better!).
The following link has some info on digital tv technical principles, and possible reception problems, with (limited) practical advice:
www.dtg.org.uk/publications/books/r_book2.pdf .
More generally, the above dtg website - www.dtg.org.uk - is an excellent information source, having various technical stuff, and comprehensive web links.
For more detailed technical information on UK digital terrestrial tv, Google search for "MPEG2" (video and audio compression details), and "COFDM" (the transmission system used).
An "official" list of the actual DTT transmitters is at:
www.dtt-tx-info.org .
There's a more informative version here:
www.dtg.org.uk/retailer/transmitters.html .
(once again, it's actually the dtg website!).
GETTING FREEVIEW, WITH "THE RIGHT STUFF" (ie, equipment!).
1) Digital tv is a "different animal" from older analogue tv. Generally speaking, within a reception area you get it (perfectly!), and outside you don't (and no use trying!). That's oversimplifying, but basically true.
2) With an extremely strong signal, ie, in Greater London, or central Birmingham, or very close to (within sight of) most transmitters, DTT (Freeview/Topup) receivers should work on "just a bit of wire" stuck into the aerial socket.
Slightly further away, you may still be ok with a set top aerial and "signal booster" amplifier. However, there's no guarantee these will work, and better ones (well designed aerial with low noise figure amp) are more expensive, so some "financial risk" is involved!
(note that - on an indoor aerial - DTT reception might be good, but the picture can still occasionally "jump" when people move around nearby).
Then, at slightly further away still - and throughout most of the remaining "possible reception zone" - a roof aerial will be needed.
3) After which, it's a matter of getting the appropriate roof aerial:
a) You may be able to get the newer digital signals on an exisiting roof aerial, previously meant only for analogue ones. If so, then that's fine (and you're lucky!). Otherwise:
You need the correct frequency band, for the transmitter you're using, so check. Also, some transmitters have the newer Freeview signals partly "out of band" - ie outside the "frequency plan" previously used for just analogue tv - in which case you'll need a new wideband aerial; again, check!The UHF bands are A, B, C, D, E . These days, aerials tend to be AB or CD or E. Unfortunately, for the same gain, wideband aerials are larger and heavier!
c) You need sufficient aerial gain, to get a high enough signal (digital tv being "all or nothing", the signal being "not quite good enough" means you get nothing!).
Generally, the biggest available aerials have a gain of 18dB (that's "100 elements up", for higher gain the number of metal rods starts to get impractical!).
For some idea of different aerials available, see:
www.tlc-direct.co.uk/Main_Index/Aerials_Index/AerialsTv/index.html .
Also: http://www.blake-uk.com/prod_products_aerials_tv_menu_selection.aspx .
The highest gain aerials available are ones like the DAT 75 (19dB max). See www.hdtvprimer.com/ANTENNAS/DAT75.html .
For higher still gain, it's most usual to use a masthead preamplifier. Otherwise, you can combine single aerials together, but this needs specialist knowledge (and you end up with "lots of ironmongery" on the roof!).
However, at distances where such high gain is needed, it's usually a matter of fringe reception conditions, an the signal may well be affected by diffraction across mountains, differing day/ night propogation, etc, so things can get unreliable (although, you might still be relatively lucky), for particular locations it's very hard to say!
(sometimes you see a different aerial type, several stacked parallel rods in front of "chicken wire". This is low gain - for local transmitter - but with high off-axis rejection, useful if there's strong interference).
(Also, there's the wideband log periodic types, characterised by a centre "double strut" and roughly triangular shape; sometimes used on caravans, but not recommended for roofs).
4) "Digital" aerials tend to have extra gimicks, eg, things like "ghostkiller" directors and larger reflectors. While these may be of use with some digital tv reception problems - eg co-channel interference in some regions, or impulsive interference - for most people the main "reception problem" is achieving sufficient gain!
5) Then you need the appropriate connecting cable! Old style "UHF" coaxial cable - previously used for analogue aerials, and widely sold in DIY shops - is "usable", but satellite grade cable is better! For 2 reasons: a slightly lower signal loss at higher UHF frequencies, also the continuous foil outer condutor (which better resists impulse interference, very much a digital tv "known problem"!).
(Note that, ideally this should be moisture proofed (foam dielectric) type, as the "satellite cable" sold in DIY stores is usually RG6u, usable, but with an air-spaced interior which will gradually absorb moisture and slowly degrade if used outdoors. Not as serious a problem at digital terrestrial tv frequencies as it would be at satellite ones, so can be used, but a waterproof type would be better!).
6) The tv receiver can also make a small (but again important, it's "all or nothing"!) difference. Most DTT receivers claim a (working) input sensitivity level of - 75 dBm (see decibels explanation). Some cheaper sets are not as good, so a "known name" brand increases the chance of better reception (for example, maybe avoiding the need to upgrade and older existing aerial!).
(The specially designed Novapal tuner, as used in Setpal - and some other - receivers, has a sensitivity of -83 dBm, the best there is! However, Setpal are now defunct (although still available as I write this, eg, www.cpc.co.uk type "digital tv receiver" in searchbox, also in some Tesco stores)).
DIGITAL RECEPTION, A FEW PRACTICAL TIPS!
1) Once you're sure about which "bands" aerial to use, go for the highest gain one you can find. Many digital tv reception problems come from an inadequate aerial, and you're extremely unlikely ever to get "too much" signal (and - in this very unlikely event - UHF attenuators are very cheap!).
2) If considering re-using your exisiting analogue aerial for digital reception, then if possible first briefly borrow a neighbour's Freeview receiver, and try that on the aeiral (use the existing stored channels, don't re-scan!). This will give you some idea. If the reception isn't good enough, is it a long way off, or "only just insufficient" (ie, intermittent partial reception with pixilation - "blocking" - maybe ob just some channels. If the former, then it's a new aerial needed! But, if the latter, then you just might get away with a "signal booster". A "small financial risk" (or zero, if you can borrow one to try!).
3) When adjusting the aerial, what you're after is the minimum bit error rate (from Viterbi decoding)! This is shown by the "quality" reading on the signal strength display page (available on nearly all Freeview sets!). Usually, some muxes (groups of channels) - particularly ITV/ Topup ones - will be worse than others, and you should try for the best quality reading on the worst received channels (then, the "stonger" channels should be ok too!).
(note that, high signal quality means low received error rate (!), so if your receiver shows a number as well as a quality bargraph, then as quality goes up this number will go down!).
4) A particular "known problem" with DTT is impulse interference (due to how COFDM works, Google "cofdm impulse interference" for further info). If you have this, the picture can keep freezing, or just disappear. Mainly due to nearby electrical devices, especially motors and/or flourescent lights elsewhere in house, possibly cars passing by just outside, etc.
Here's where "digital aerials" score! The higher directivity of "crossed" directors - and a bigger reflector - are some help. Also, a balun is a good idea, possibly one at both ends of cable (digital aerials often have one built in, check the specs!).
A further "tactic" is to try and partially screen the aeiral from interference sources. This is already partly true of any roof aeiral (ie, screened from any ground directly beneath!). However, it's also possible to "put brick" - eg, the side of a house - between aerial and a known interference source. Another ploy is to slant aeiral slightly upwards, further decreasing pickup from the ground below!
ANALOGUE AND DIGITAL TV, RECEPTION DIFFERENCES!
As the transmitted tv signal spreads outwards from the transmitter, it gets weaker with increasing distance away.
(At any instant, think of total signal energy as contained in a thin spherical "shell", slowly expanding outwards. As this gets bigger, signal energy gets "more spread out" - across the shell's surface - hence it's weaker. From simple goemetry, surface area of a sphere is 4xPIx(radius squared), so signal power actually falls off according to the inverse square of distance from transmitter. Note that - if using measuring equipment - electric field stength only falls off as inverse distance (not squared!), but it's the power that matters!
(Across entirely urban areas, scattering from buildings changes this to an inverse 4th power law! However, except for maybe in London, this wouldn't apply to digital terrestrial tv in most of the UK!).
As the tv signal weakens, it's increasingly affected by "electrical noise", at first picked up via aerial (from various sources), and -ultimately when the signal gets too weak to use - from devices (transistors) inside the tv receiver itself!
There's also radio wave propogation, the UHF terrestrial tv frequencies - pretty much! - travel in a straight line, so reception starts getting difficult "below the transmitter horizon" (when the receiving aerial can no longer "see" the transmitter, due to earth's curvature). However, distance weakening effects are usually more important, "kicking in" before the horizon effect does!
As signal weakens, the receiver agc (automatic gain control)compensates, but the relative noise level rises. As we say, the signal to noise ratio (SNR) worsens (gets lower!). This affects analogue and digital tv reception quite differently:
Analogue tv - the video signal voltage represents picture brightness (ignoring colour here!), and receiver timebase "synchronising pulses" are added. This (vestigal sideband) amplitude modulates a carrier signal - so the transmitted "wave shape" closely follows picture brightness - then at the receiver the original video is (fairly easily) extracted.
As noise levels increase, the picture gets ever more "snowy" - and faint - but you can still make it out, even when quite bad! And the receiver timebases will continue locking into the sync pulses - even if they're badly distorted. So, even when reception gets fairly bad, usually you can still watch/listen.
(In the jargon of Information Theory, we say analogue tv is "highly redundant", ie, the signal can be hugely distorted but remain "watchable", as long as the receiver timebases can still "lock in", with the human eye/brain then "integrating" across many picture frames to compensate for picture noise ("snowiness")).
Digital tv - is very different! The picture is coded as numbers, which are transmitted as on/off pulses (ones and zeros). To decode the picture, nearly ALL these pulses must be correctly received! However, as the relative background noise level rises, there's an increasing chance of a "1" becomming a "0" (or vice versa!). The solution is the one always used in digital telecoms, have an error correction system! For digital tv, several methods are combined - bit interleaving, convolution, galois fields - with the correction information then being added to data as extra bits ("Forward Error Correction", ie, it works one way only, there's no "backwards path" as in computer communications, which - for example - can allow modems to slow down under "very noisy" conditions!).
This means that - as the signal weakens and the (relative) noise level rises, the error correction system "works increasingly harder" - correcting more wrong bits - but decoding still works properly, and the picture remains "perfect"! Up until a certain point, where the noise becomes so bad that - quite sudenly - decoding is no longer possible, and the picture "completely freezes" (again, I'm oversimplifying, but that's basically correct!).
So, we get the famous "digital cliff" (or "sudden fallover") effect! In fact, it even has a specific number, a ratio of 17 dB ("Rayleigh" modal value to variance ratio), or approx pre MPEG decode (post Viterbi) error rate of around 1 in 1000.
(correction - this magic figure is actually CNR, carrier to noise ratio, but it effectively becomes the above in very bad multipath conditions).
(In actual fact, this "digital cliff edge" can be rather wobbly! For example, day/night radio wave propogation can slightly differ, so you might get some digital channels by night but not by day (or vice versa!). If "only just on the edge", then there will be pixilation - the picture breaking up into individual squares - since only some "I" frames and partial P/B frames will get decoded).
(In Information Theory jargon, digital tv has "low redundancy", ie there's not much "room for error", and a just relatively small proportion of bit errors will make pictures/sound become completely undecodable!).
A WORD ABOUT DECIBELS (d
.An amplifier multiplies, ie, the output is so many times bigger than the input (ideally, that's ignoring various distortions!). So, for signal output level, you just multiply input by amplifier "gain"
A tv "receiving chain" - aerial, cable, various amplifiers (including tv receiver) - is basically many amplifiers cascaded (something that loses signal, eg cable, is just a "negative gain amplifier" with gain figure less than "one"). So, for total system gain you multiply together the individual component gains.
Doing lots of multiplications is awkward (even with a calculator!). So, usually, gains are first re-expressed logarithmically. Then, to multiply, you just add them (if anyone can remember "school maths"), which is much easier!
gain in dB = 10 log (base 10) actual power gain figure.
Roughly, x2 = +3dB, x10 = +10dB
(So, "x10x10x10x10x2x2x2" becomes "+10+10+10+10+3+3+3" instead, giving 49dB instead of 80,000; the dB version is easier to handle, and mistakes are less likely!).
(Note that, for voltages, the above formula value must be doubled (!), but tv signal strength is normally expressed in power ratios, often relative to dB(m), a fixed power level of 1mW. For example, Freeview receiver working sensitivity - when given at all! - is often -75 dBm).
SOME "EASY" ANTENNA THEORY (!).
From above, for digital reception it's clearly important to get a good signal to noise ratio (SNR). That's the job of the antenna.
The theoretically simplest -and physically unrealisable! - antenna type is the "isotrope", ie, "receives equally strongly from all directions". This has an "effective aperture" of: (wavelength squared)/(4)x(pi) . As you can see, this formula gives an area (!), in fact the antenna's "capture area"; it "collects" radio waves from this "cross-section" of space. So, the bigger this aperture is, the stronger the received signal will be.
But, wavelength is inversely proportional to frequency! So, as transmission frequency rises, the aperture falls, making it increasingly difficult to get a strong (good SNR) signal.
So, at AM frequencies, a single long wire will do, as used in portable AM radios and by radio amateurs (I'm cheating, the aerials inside AM sets are magnetic, but same principle applies!). Going up to FM/DAB, a single rod aerial on the set can suffice, but one on the roof is better, possibly having 3 or 4 elements. With higher terrestrial tv frequencies, a multi-pronged (Yagi) aerial on the roof is nearly always needed. Finally, for sat frequencies, we need a dish to collect ALL the signal from a certain area, dish size rising with falling signal strength (as readers know).
The simplest "realisable" (and usually used!) antenna is the Dipole, a metal rod one half wavelength long, with feeder cable connecting to a small gap in centre. This is "resonant" (not all antennas are), giving it several advantages:
a) directional, a null along the axis but 1.6 dB gain (over isotrope) and
bandlimited, it only responds to a a limited bandwidth around the centre resonant frequency.Both these features help increase gain of wanted signals, and limit response to unwanted ones, exactly what you need for digital reception (for good SNR)!
Note that, it's the centre gap and feeder cable that makes the metal rod into a dipole. Otherwise, it's just a metal rod, and doesn't do anything! (you need a gap and cable to "extract" any received radio energy). However, if several parallel rods are placed sufficiently close together, then - by near field induction - some radio energy can "bleed across" between them. Then, they all "partially" behave as (limited) dipoles - each passing a bit of energy on to the next - until the final (actual) dipole, which transfers the total collected energy into its feeder cable.
Each rod "linked" this way to the main dipole transfers a bit more energy, and the cumulative effect is to gradually increase the aerial's total "aperture", ie to create a much larger "area" over which it can "collect" radio energy, hence improving the SNR.
This is the basis of the well known Yagi (inductively linked array) tv aerial. The supplementary dipoles are called "directors". Then, behind the main dipole is a "reflector" (by adjusting length/position, this feeds back to main dipole in antiphase, partially cancelling signals received from that direction. Modern aerials often use a mesh reflector, rather than just a rod, for greater effect).
But - roughly speaking - each further Yagi 3dB power gain requires (approx) doubling the number of "elements" (rods). However, we've said that transmitter power falls with inverse square of distance! So, it's a "law of diminishing returns", and with increasing distance the number of elements needed rapidly becomes impractical (from increasing weight!).
(Pictures: see www.sub-tv.co.uk/antennatheory.asp ).
(more theory, see: www.ee.surrey.ac.uk/Personal/D.Jefferies/yagiuda.html ).
IMPEDANCE/MATCHING.
With antenna systems, component sizes (aerials, cables, etc) are of comparable size to wavlength, hence transmission line theory applies, and antennas/cables/other devices must "match" in impedance value (otherwise, energy reflection occurs at mismatched interfaces, and you get analogue ghosting and digital increased error rate).
It just happens that a dipole has an "equivalent impedance" of 75 Ohms (another reason for using it, as energy losses along the connecting cable then become insignificant, non-resonant antennas have lower impedance and therefore higher cable losses!).
So, tv coax is made with 75 Ohm impedance (Ethernet/transmitter coax is often 50 Ohm, for different reasons).
(Yagi antennas use folded dipoles of 300 Ohm impedance, and the other added elements bring this down again to 75 Ohms).
(more, see: http://www.moonraker.com.au/techni/impedance.htm . This describes signal transmission, but applies equally to signal reception!).
COMBINING SEVERAL ANTENNAS.
(Before attempting this crazy stunt (!), it would be a good idea to check you're actually getting an identifiable digital mux signal, although at too low SNR for decoding. Remember, "doubling the metalwork" only increases received signal by 3 dB, and using 4 aerials only by 6 dB, that's not much! Although, given the "digital cliff", it could make the vital difference. Chance of success also increses with the best available tuner, currently still the Novapal one inside Setpal receivers, or the Echostar T101 if you can find it).
So, we've said a single Yagi has a practical size limit - due to increasing weight - the biggest being the above DAT 75 (really 3 Yagis stacked vertically, to reduce bending moment on chimney!)
If that's not enough, then - subject to there actually being a signal, albeit too weak for digital reception! - can we combine several Yagis for even higher gain? Yes, but it's a bit involved!
1) A Yagi works by "nearfield induction" between the elements, but we don't want this also happening between different Yagis! (if so, then they "take" some proportion of signal from each other, instead of from the wanted transmission). From theory, induction field falls to equal radiation field at distance: (wavelength/2xpi), which at UHF tv wavelength of around 0.4m = (v. approx) 7cm. Using the engineer approximation of "multiply by 10", then, multiple separate combined Yagis should - ideally, not always possible! - be 70cm or more apart.
2) As said above, maintaining the correct impedance is important, so you can't just "join the wire ends together"! Instead, use an aerial combiner. Unfortunately, commerically available ones are usually for combining differently banded aerials - hence have built in filters - or are for tv/fm etc. In which case, use a widband cable splitter "in reverse", 5-2400 Mhz, available 2-6 output versions:
(eg, see www.showmecables.com/showProducts.asp?category_id=414 ).
(that's a USA website, but similar splitters are widely available! Note that UHF tv only devices 5-1000 Mhz are fine, but sat only devices - 900 - 2400 Mhz won't work at UHF tv frequencies!).
(Also note, the "power pass" feature - on some/all/none splitter/combiner ports - is usually irrelevant, since dc is not fed along UHF cables, unless you've got a masthead preamplifier for which see later!)
3) This is the really tricky bit! OK, we've got "coherent" signals from a single source, so just adding them together will increase signal strength and - vitally for digital - the signal to noise ratio. But .....
We must add the signals together in phase (preferably exactly, or at least as nearly that as possible!). Otherwise, as they go out of phase, some cancellation will start to occur, effectively lowering the strength you're trying to increase. By a certain point, you're back where you started. After which, it gets worse, and at a half wavelength difference you get total signal cancellation, ie nothing!
Now, wavelength in UHF tv band is (approx) 0.6m. So, a "path difference" of half that - 0.3m - will give complete signal cancellation! Actually, to avoid significant losses you want much less than that, so let's say the difference in any cable lengths from aerials to combiner must be under 10cm, that's not very much!
But it gets even worse! Radio wave speed in coax cable is less than in "free space" - possibly sometimes just half - then the cable length difference goes down to just 5cm, before "significant" combiner losses start occurring!
So, yes it can be done, but must be very carefully!
(It's possible to "null out" any phaseshifts with r.f. engineering techniques, but this requires taking measurments and placing phase shift "stubs" in the right places, not to be considered unless you're already an r.f. expert!).
AMPLIFIERS AND "SIGNAL BOOSTERS".
Any aerial amplifier at all has electrical noise at its output, at a level far higher than any "atmospheric" noise picked up on a tv aerial. This will then create a new "noise floor", thus decreasing the sig to noise ratio of any digital tv signal passing through it (from above, exactly what we don't want!).
(Any transistor has thermal input noise - across its input resistance - and amplifies this by its own gain, then adds further "transistory" type noise, like "shot", "flicker", etc. If there's several transistors in "cascade" - or if an integrated circuit amplifier is used - then the noise is further amplified several times!).
(see: http://www.cisco.com/en/US/products/hw/cable/ps2209/products_white_paper0900aecd800fc94c.shtml .
Exactly the same principles apply to tv aerial amplifiers, where a tv aerial is a "noise source" just like a resistor!).
(If you are a keen masochist, there's also this: www.ieee.li/pdf/viewgraphs_mohr_noise.pdf ).
Usually, an in-line aerial amplifier will have a specified "noise figure"; then the SNR (or CNR, whichever you prefer) worsens by this amount. So, for a 26dB sig-to-noise signal going through a 5dB noise fig amp, SNR then degrades to 21 dB (aren't dBs nice and easy!).
(Note that the concept of "noise figure" is just a "calculating convenience", suggesting a ratio, whereas amplifier output noise is actually fixed (with some temperature dependance). However, it works well enough at the "expected" input level).
Well, ok then, aerial amplifiers always make the sig to noise ratio worse, which is bad for digital tv, so why would we ever use them? Well, there's one or two special situations ....
1) Received field strength not high enough to operate DTT receiver. And, there's 2 sub-cases of this:
a) Using a set-top tv aerial in a strong signal area, close to transmitter. This type of aeiral is very poor, and despite a strong signal may easily still not give enough micovolts at receiver input. In which case, an amplifier might just do the trick. It's going to make the SNR much worse, whatever else, but the local field is high enough for this still to be usable.
The amp "built into" most set top aeirals - or an added "signal booster" - is likely to be dire, with a noise figure of 6dB or more, so we're worsening ("throwing away"!) the received signal by at least that much, but if close enough to the transmitter you might get away with this!
In a fringe reception area, already using a high gain aerial (much trickier!).In this case, the amplifier design is critical, since yes we can amplify, but anything that also worsens the SNR is bad news (this time, we can't afford to "throw away" some proportion of the SNR arriving at the receiver input!).
Looking at the above "noise" links, both give (slightly different versions of) the standard formula for noise in amplifiers in cascade. It can be seen that the first amplifier in any such "chain" is always the most critical, making the biggest single increase in the overall noise figure. In this case, that would be the aeiral amplifier!
So, what we have to do - for any chance at all of improving digital reception - is use the best possible type of masthead amplifier, for example see:
www.jwhardy.co.uk/shop/pages/pl_amp_mast.html .
A "good" noise figure would typically be about 1.9 dB, for UHF tv band only (you don't want wideband!). This is about the limit of what's reasonably possible, using just a few discrete components - not integrated circuits! - in simple well designed circuit configurations, for the lowest possible added noise (don't bother buying a slightly cheaper amp with worse specs, you'll be wasting your time and money!). Note:
1b1) A masthead amp should go as close as possible to the aerial. Because, you want the best achievable sig to noise ratio. And, as you move away from the aeiral, the received signal goes down, but amp noise output remains fixed, so SNR worsens! Yes, ok, you can use a masthead amp indoors - that's still better than a cheap "signal booster" - but will give a worse result than on aerial.
1b2) Get the highest possible gain - up to 30 dB - because you want the SNR as good as possible. A higher gain amp may have some spurious intermodulation products - especially if poorly designed - but this is not dissimilar in effect to multipath reflection, and COFDM can cope with that! But, it can't cope with a too low SNR (too high "noise floor").
1b3) Of course, a masthead amp also needs the appropriate power supply, usually around 12-15v. Just 2 or 3 transistors in the amp take a "trickle charge" of only a few milliamps, unnoticable!
1b4) If combining several aerials - as above - then the single masthead amp goes AFTER the combiner! In which case, whether or not the combiner has any "power pass" sockets is still irrelevant!
2) The other case where a masthead amp is used is when there's local electical interference. But - see above - using sat grade cable and positioning aerial "favourably" should screen out most of this. A UHF tv aerial is band limiting, as is the if stage(s) in the receiver, both together greatly reduce the received noise. It's also possible to add a cheap UHF bandpass filter between the aerial and receiver input, for even greater noise rejection.