[TowerTalk] Snow and rain attenuation

VR2BrettGraham vr2bg at harts.org.hk
Tue Nov 1 17:49:33 EST 2005

W6RMK added:

>>Any dish not pointed by somebody who knows what they are
>>doing & with a spec-an could be as much as 1 dB down from
>>where it could be.
>>Dish size on the ground is always too small - despite filings
>>to ITU for coordination of the bird in the first place.
>And there you have the basic conflict in high gain antennas.  If you make 
>it big enough to have the link margin you want, then you've got to point 
>it too accurately, which raises the installation cost.

Yes, installation monkeys usually aren't given spec-ans, but those that I
have watched who do have them will do silly things like try to do so with
silly setting like 20 dB/div on amplitude scale, or even earth station
staff who are too lazy to carry spec-an to roof, so try to point reasonable
contribution feed antenna by having somebody on an HT inside try to
describe what is being seen to those twisting the dish.

>There ARE solutions, of course.. you can make an electronically steered 
>phased array for fine pointing, for instance. (This is what I've been 
>doing for the past two years at work, except for transmitting video back 
>to earth from Jupiter or Mars)

That sounds like fun stuff!

>So far, though, nothing is cheap enough for the exceedingly cost sensitive 
>world of consumer electronics.  I personally find it amazing that they can 
>mass produce LNBs with 20K noise temperatures at 12 GHz and sell them for $50.

Errrr... that's a bit dear, though generally only talked about container
load quantities or about setting up factory in VU-land to avoid extreme
import duties.

>That whole industry (as Brett can tell you) calculates everything down to 
>the penny.  Since they subsidize the receiver and dish/LNB cost, that 
>factors in.  "Do we make the dish 2" bigger, and spend $0.50 per unit for 
>aluminum on millions of dishes, or do we add another square meter of solar 
>panel to the bird."  I used to work with a Set Top Box manufacturer, and 
>they agonized about pennies: "Do we really need that bypass capacitor there?"

Oh yes - though not lucky enough like Charlie Ergen to start launching
our own fleet of BSS birds, removal of components is important
cost-down technique in consumer electronics.  When I ran factories
in BY7, had some really silly suggestion from our assembly partners.

I dabble in consulting work now & most recent job was to convince the
leading chipset manufacturer it had serious problem in DSP core &
also teaching a rather significant set-top maker that all the other
problems in their product were not because of use of low-end
chipset, but their skimping on power supply rail decoupling.  Now
hoping that VU cable industry moving to 256-QAM will mean more
work - I think due to ham radio the front end side of the boxes is fun
stuff & having shown another box maker who does know their stuff
a number of serious problems with some of the early tuner-in-a-chip
silicon they used, will expect all sorts of problems that should be
good at filling my rice bowl for a while longer yet.

The extreme cost-down trend in consumer electronics is not good
for EMC, as hams know all too well.  And all sorts of problems
nobody anticipated or seem to comprehend (like consequences
of ripping out selectivity in LNBs to get down to US$20 range).

Anyway, a lot of junk out there.  Big DTH platforms used to employ
folks to deal with this.  I reckon the typical WISP to be totally
clueless, so beware - a good indication I have learned is whether
they do their own link budgets or use something they got from
somewhere & plug in numbers... the latter will also accept at face
value claimed specs of the hardware, or not even realize that
crucial parameters aren't even spec'd.  Guaranteed grief will follow.

73, VR2BrettGraham

More information about the TowerTalk mailing list