Re: analog catv design
biff44 said:
Just search around on "multi tone intermodulation". Here is one:
h**p://www.maxim-ic.com/appnotes.cfm/appnote_number/728
The problem in CATV is that you have 200 RF channels, right next to one another. So, a 2 tone intermod test is a little too easy for the components. You put in 3 or 5 tones, and see what the highest intermod products are. Multitones can constructively add or subtract to really have a big peaking factor--really stressing things like amplifiers.
If you have high intermods, one or more CATV channels interfer with a desired one, and the system stops working.
----------------------------------------------------------------------------
But 2-tone test is very importent test to caracterize equipment.
from 2-tone can calculate IP3 and from IP3 value you can
recalculate multiton situation and different wanted
carrier vs inteference dynamic.
Two-tone give steady spuriose/IM and easy to measure
but > 2 tone test give spuriose with sum of cycling envelopes,
give lower RMS-value but still high peak value etc.
and more carrier give lower RMS-value on each IM-product, but
envelope is more and more puls shaped and still
high peak value.
you can calculate ruffly 10 * log10(number of carriers / 2) = give how
many dB lesser signal strengt each carrier compare to two-tone test
and approx. same IM and spuriose RMS-level floor.
for example you measure IM-product -36 dBm on 15 dBm output
in two tone from amplifier and can calulate IP3
IP3 = (dBc/2) + dB above 0 dBm
here:
dBc = (36 + 15 = 51 (differense between carrier and IM-product, C/I)
and
IP3 = (51 dBc / 2) + 15 dBm = 40.5 dBm IP3
from this value can calculate in four carrier situation
ie 10 * log10( 4 / 2 ) = 3dB lower each carrier compare
to two tone test for approx. same IM-product RMS-value
(-36 dBm)
(ie 3 dB lower/carrier each doubling number of carrier)
- total output power is a constant and spread out of number
of carriers
if peak-value on IM is importent (depend of type of modulation if adjecent
channel affected on this or not), you need calculate 20 * log10( number
of carrier / 2)
(ie 6 dB lower/carrier each doubling number of carriers)
---
this i a ruffly calculate, but works well in many cases..
if you not have equal level of carrier, you count strongest
carrier (represent minimum 3/4 of total power) and
skip carrier 6-10 dB lower on strongest carrier
in calculate-process
---
if you know systems IP3 value, and know allow max level
of IM-products in dBm and want 2*carrier level and C/I,
you can calculate:
IP3 = +67.5 dBm (yes is a strong amplifier)
IM = -36 dBm
2xcarriers = ?
and try formula (and hopfylly written right now)
2xcarrier = ((2 * IP3) + IM) / 3 = ((2 * 67.5) + (-36)) / 3 =
= 33 dBm * 2 carrier
and C/I_two_tone = 33 + 36 = 69 dB
and using formula early to recalculate to multicarrier system.
Ie. four carrier = 30 dBm each, eight carrier = 27 dBm each carrier etc.
If IM peak is more importent:
four carrier = 27 dBm each, eight carrier = 21 dBm each carrier etc.
- but still ruffly calculate.
---
common used formula for calculate C/I in eqvidistant multitone
situation and spuriose free dynamic is importent factor, use:
C/I_total_dBm = C/I_2_tone_dB - 6 dB + 10 log10(x)
(C/I = carrier over interferences)
there 'x' are
n carrier x 6dB + 10 log10 (x)
2 0 dB
3 1 6.0 dB
4 2.3 9.6 dB
5 4.5 12.5 dB
6 7.5 14.8 dB
7 11.5 16.6 dB
8 15.5 17.9 dB
9 20 19.0 dB
10 26 20.1 dB
11 33 21.1 dB
12 40 22.0 dB
13 48 22.8 dB
14 57 23.6 dB
15 67 24.3 dB
16 77 24.9 dB
n > 16 ~ (3/8) n^2 6 dB + 10 log10((3/8) n^2 ) dB
( why I not can use old fasion way 'whitespace' to format list
and fractal in formla - list above collapse now, 'submit' take away
all exept one withespace betwen words in text - and i already spent
for much time to writing text... - if possible fix it, educate me how!!!)
for equidistance carrier in frequency and same level
on all carrier
---
On a 67.5dBm IP3 amplifier you can give out 8.1 dBm * 16 carrier
and IM below -36 dBm in all time (44.1 dB C/I Spur-Free Dynamic Range)
and 200 carrier give -8.7 dBm/carrier and 27.3 db C/I SFDR
---
ok, answer is litte messy - going from simple and fast answer
to little more accurate answer in writing process... :*)