[?] Avago MG-71543 amplifier IC, and understanding S-parameters
David Chapman
I've just come across the Avago MGA-71543 Low Noise Amplifier IC which
is available at low cost in the UK (Farnell Part No. 134-7205 @ �1.34
each).
I'd like to use this IC as a receiver pre-amplifier at 434MHz since it
appears to give good gain and low noise whilst drawing just 10mA from a
3 volt supply. Although the data sheet quotes an operating frequency
range of 100 - 6000MHz for this device, it only gives input/output
matching component values for 1900MHz and 800MHz.
Unfortunately I have no experience of using S parameters such as those
shown in the Avago data sheet and am therefore rather clueless about how
to start matching the input and output of an amplifier using them. I did
learn about using the Smith Chart in college nearly 50 years ago but, to
my shame, have had no reason to use it since then, more is the pity.
Has anyone used this device at or near 434MHz and can recommend
suitable input/output matching component values OR can anyone suggest an
easy method for a beginner to learn how to design matching networks
using S-parameters?
Any constructive suggestions offered will be much appreciated,
Many thanks,
--
David C.Chapman - (dcch...@minda.co.uk)
--------------------------------------------------------------------------
oopere
David,
The S11 parameter gives you the input reflection coefficient when the
output port is terminated into 50R.
[using data from page 7]
For 500 MHz, roin = 0.921*exp(-j*16.4*pi/180)=0.8835 - 0.2600i
This means that Zin = 50*(1+roin)/(1-roin) = 93 - 320i [ohm]
So, for maximum signal transfer, you need to match 93 - 320i to your
input source (50 R). You can do this with the matching network of your
choice (*).
OTOH, if you want to optimize for noise, the datasheet should give you
the Gamma_opt value for 433MHz. You may extrapolate this info from the
data at 900 and 700 MHz. For 700 MHz
gamma=0.61*exp(j*16.3*pi/180)=0.5855 + 0.1712i
Zopt=50*(1+gamma)/(1-gamma)=156 + 85i [ohm]
This means that for minimum noise, you have to build a matching network
that transforms your source impedance (50R, assuming) to 156 + 85j R.
Again, use the topology of your choice (*).
(*) Consider biasing issues, as described in the datasheet.
Pere
mi...@sushi.com
> --------------------------------------------------------------------------
David Chapman
<8e70c278-ab32-41ce...@z3g2000prd.googlegroups.com>,
"mi...@sushi.com" <mi...@sushi.com> writes
Many thanks to those who took the trouble to reply, and for the
useful pointers.
- Dave
mi...@sushi.com
> In article
> <8e70c278-ab32-41ce-aea2-950a899e5...@z3g2000prd.googlegroups.com>,
> "m...@sushi.com" <m...@sushi.com> writes
> --------------------------------------------------------------------------
Knowing what S-parameters are is only half the problem. Next up is the
matching network. Plenty of rants about not doing conjugate matches,
but little advice on what you should do.