Maker Pro
Maker Pro

calculating conjugate match, etc. (more on rf amplifier)

H

Hank

Jan 1, 1970
0
I have obtained several small rf transistors, including a NTE311 (similar to
a 2n3866) and a NTE278 (similar to a 2n5109). I have spent a few hours
reviewing several books I have (primarily RF Circuit Design by Chris Bowick)
and have at least indentified the fact that I am much more in the dark than
I first thought :)

I see now that I can dc bias a transistor for class A operation, and then
design a conjugate match for the source impedance and the load impedance.
Unfortunately, most of the examples assume you have Y or S parameters on
your data sheet, which I do not...not for any frequency, much less 110mhz.
So I either need to guess, measure or locate these parameters for the NTE278
and the NTE311. Suggestions?

Thanks,

Hank
 
G

gwhite

Jan 1, 1970
0
Hank said:
I have obtained several small rf transistors, including a NTE311 (similar to
a 2n3866) and a NTE278 (similar to a 2n5109). I have spent a few hours
reviewing several books I have (primarily RF Circuit Design by Chris Bowick)
and have at least indentified the fact that I am much more in the dark than
I first thought :)

I see now that I can dc bias a transistor for class A operation, and then
design a conjugate match for the source impedance and the load impedance.
Unfortunately, most of the examples assume you have Y or S parameters on
your data sheet, which I do not...not for any frequency, much less 110mhz.
So I either need to guess, measure or locate these parameters for the NTE278
and the NTE311. Suggestions?

Most folks find it easiest to hook up the network analyzer and measure it
(s-param) if they are not available from the manufacturer for the given bias
conditions. It isn't a bad idea even if you do have the manufacturer's data
sheet.
 
J

John Woodgate

Jan 1, 1970
0
I read in sci.electronics.design that gwhite <gwhite@hocuspocus_ti.com>
Most folks find it easiest to hook up the network analyzer and measure it
(s-param) if they are not available from the manufacturer for the given bias
conditions. It isn't a bad idea even if you do have the manufacturer's data
sheet.

'Most folks' (i.e. a majority of about 6 billion) don't have a network
analyser.

It's possible to calculate approximate y-parameters from data that IS on
the data sheet. To do this, you need to understand y-parameters in some
detail, and that's a bit too much maths to put on a newsgroup. Look for
textbooks rather than web sites for this sort of stuff.
 
J

John Woodgate

Jan 1, 1970
0
I have obtained several small rf transistors, including a NTE311 (similar to
a 2n3866) and a NTE278 (similar to a 2n5109). I have spent a few hours
reviewing several books I have (primarily RF Circuit Design by Chris Bowick)
and have at least indentified the fact that I am much more in the dark than
I first thought :)

I see now that I can dc bias a transistor for class A operation, and then
design a conjugate match for the source impedance and the load impedance.
Unfortunately, most of the examples assume you have Y or S parameters on
your data sheet, which I do not...not for any frequency, much less 110mhz.
So I either need to guess, measure or locate these parameters for the NTE278
and the NTE311. Suggestions?
The simplest thing for you to do is to choose other transistors for
which y or s parameters ARE on the data sheets. If the manufacturer
doesn't quote them the device isn't, prima facie, intended for
applications where they are needed for the design.

The ft of the 2N3866 is quoted in my book as 500 MHz. It isn't really
going to be too good at 110 MHz. The 2N5109 is more suitable (ft = 1200
MHz).
 
F

Fred Bloggs

Jan 1, 1970
0
John said:
The simplest thing for you to do is to choose other transistors for
which y or s parameters ARE on the data sheets. If the manufacturer
doesn't quote them the device isn't, prima facie, intended for
applications where they are needed for the design.

The ft of the 2N3866 is quoted in my book as 500 MHz. It isn't really
going to be too good at 110 MHz. The 2N5109 is more suitable (ft = 1200
MHz).

Well let's see that schematic for the 110MHz 30dBm 2N5109 amp?!!!!
 
H

Hank

Jan 1, 1970
0
Fred Bloggs said:
Well let's see that schematic for the 110MHz 30dBm 2N5109 amp?!!!!

Thanks, but I now want to come up with the design myself. I have never
devoted enough time to really understanding analog circuit design,
especially high frequency and I think it is about time I learn it instead of
expecting someone to post a schematic that is tweakable to my application.
Besides, this is fun stuff! :)

Thanks,

H
 
H

Hank

Jan 1, 1970
0
John Woodgate said:
I read in sci.electronics.design that gwhite <gwhite@hocuspocus_ti.com>


'Most folks' (i.e. a majority of about 6 billion) don't have a network
analyser.

True! I don't even have a spectrum analyzer, though I do have a 7104
mainframe :-(

It's possible to calculate approximate y-parameters from data that IS on
the data sheet. To do this, you need to understand y-parameters in some
detail, and that's a bit too much maths to put on a newsgroup. Look for
textbooks rather than web sites for this sort of stuff.

I looked in AoE, MicroElectronic Circuits, RF circuit design, and the 2001
ARRL handbook, but I was just skimming so I probably missed it. I need to
do some real studying of these books - the information seems to be in them
for the taking.

Thanks,

H
 
J

John Woodgate

Jan 1, 1970
0
I looked in AoE, MicroElectronic Circuits, RF circuit design, and the 2001
ARRL handbook, but I was just skimming so I probably missed it. I need to
do some real studying of these books - the information seems to be in them
for the taking.

There isn't much in AOE about matrix parameters of active devices. The
new edition, available on 2 million DVD-ROMs only, will have EVERYTHING
in electronics in it. (;-)

OTOH, you might find the shunt-series pair described on page 872 of the
Second Edition, interesting. Section 13.11 may also be interesting.
 
G

gwhite

Jan 1, 1970
0
John said:
I read in sci.electronics.design that gwhite <gwhite@hocuspocus_ti.com>


'Most folks' (i.e. a majority of about 6 billion) don't have a network
analyser.

"Most folks" meaning those people who give a crap about s-param's. Somehow I
don't think that approachs 6 bil.
 
Top