RF Small Signal Design

Dec 12, 2005 7 Replies
R

For some large number of years now (comin' up hard on 40) I've used the y-parameter analysis of Hejhall in Motorola app note AN215 and it has served me well.



Recently I did a sort of oddball design and Hejhall's numbers weren't working. I decided to see of Hejhall and Norris' numbers (Motorola app note



166) correlated and much to my surprise they didn't.

So I attempted to convert y-parameters to s-parameters in HP's AN-95 routines and things got squirrelyer yet.



Hardy's book (HF Circuit Design) gave different answers. Krauss, Bostian, and Raab (Solid State Radio Engineering) ... different again.



One of two things is true. Either all these folks have a little difference in their analysis and reasoning OR Excel's method of handling imaginary numbers isn't the real world. Before I go into a long song and dance to check Excel's routines (nontrivial task), I'd appreciate comments from you all about the subject.



BTW, it's not rocket science ... a simple 2N5770 (2N918 in drag) common emitter at 75 MHz..


Jim


Vote
Twitter Facebook LinkedIn
W

You may be better off with intelligent lumped-circuit analysis, r_e from Ebers-Moll at Ic, Cbe from datasheet f_T plots at Ic, and Ccb and Cce estimated from some datasheet values or curves.

Thanks, - Win
Vote
Twitter Facebook LinkedIn
J

I've always hung my hat on Roy Hejhall and Pete Norris' works. Precisely what isn't matching up (no pun intended ?:-)

I don't know about Excel and complex numbers... haven't used it.

qrk/Mark ??

...Jim Thompson

| James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC\'s and Discrete Systems | manus | | Phoenix, Arizona Voice:(480)460-2350 | | | E-mail Address at Website Fax:(480)460-2142 | Brass Rat | | http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.
Vote
Twitter Facebook LinkedIn
M

I'm not an RF designer. Excel sucks in so many other ways that it wouldn't surprise me at all if its complex number handling is sub-par.

The first time I used Excel for an FFT was an eye-opener. This was around eight years ago, so it is may not be fair to compare with the modern version. But I tried a 4096 point FFT, and it was a go away and come back in half an hour kind of thing. I am pretty sure that whoever wrote it actually didn't implement the FFT algorithm at all.

It was only like a 200 MHz Pentium or something, but still. Matlab could do it fast enough that it felt instantaneous.

--Mac

Vote
Twitter Facebook LinkedIn
R

Ya wanna try that again in engineeringlish?

Jim

Vote
Twitter Facebook LinkedIn
J

Bwahahahaha!

...Jim Thompson

| James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC\'s and Discrete Systems | manus | | Phoenix, Arizona Voice:(480)460-2350 | | | E-mail Address at Website Fax:(480)460-2142 | Brass Rat | | http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.
Vote
Twitter Facebook LinkedIn
M

Model it as a voltage controlled current source, include the input RC time constant and the collector to base capacitance, as well.

Marc

Vote
Twitter Facebook LinkedIn
J

You might try being a civil engineer, I expect all electrical engineers worth a damn to be aware of Gummel-Poon as well as Ebers-Moll. Both are more complex than the hybrid-pi model, which would probably serve for your application.

JosephKK
Vote
Twitter Facebook LinkedIn

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required