How to measure the input and output impedance of an RF transistor

Hi how do, I measure the input and output impedance
of an RF-transistor without an S-parameter measuring device.

Kindest regards,
StoppTidigare

P.S. Does there exist simulation tools for valves circuits ?

I think it is easy to transfer the measured s parameters to impedance on the smith-chat. The precondition is that you have acuired the s parameters correctly.
Luck.

Look in old handbooks, like the ones by Terman for how they measured such things in the past. They used bridge type circuits with the valve (US = vacuum tube) in one leg.

Hi StoppTidigare,

You can measure the reflection coefficient (S11, S22) with a directional bridge, a sweep generator and a power meter (or spectrum analyzer). Is this what you want to measure?

Bye

mr_ghz, with only a power meter or spectrum analyzer, he cannot measure the phase of S11 or S22, only the amplitude, wich is a big deal to convert from S11 to Zin or from S22 to Zout.
Zin and Zout and very very dependent on the phase of S11 and S22, so without that info (phase info) he will get a trajectory of possible impedances (a circle centered in smithchart center) that includes resistive impedances, capactive impedances, and inductive impedances.
So, in short, he really needs to phase measurment (and amplitude of course, as described above by mr_ghz and flatulent) to get exact value of real resistive part as well as of reactive imaginery part.
No way to do that without a device that measures amplitude AND phase, which is basically a good old network analyzer. (or something similar that he can construct like using Analog Device's new phase/gain measurment chip up to ~2GHz only, but he needs then to worry about calculating mismatch losses...etc. Pretty much he will need then to build a full network analyzer :) )
Very difficult task indeed.
If you do not have Network analyzer (very expensive) I really advice you to at least rent one for a few days, to save yourself the hassle of building your own network analyzer hehe :) ...

Hi Gemini1706,

Yes you're right, with the simple method I mentioned you can only measure the amplitude. If this would be enough (it could be ... :D ) You can measure the SWR of your input/output.

Bye

Thanks for your responses.
I'll guess I'll borrow a networkanalyzer, or should it be a vector analyzer.
I don't know the difference..

Suppose that I just have an oscilloscope and the datasheet of the transistor, what is the procedure if I just want to tune for optimum power at the input and output.

How did they do it in the old days ?
Should I take any considerations when I cascade more stages ?

Regards,
StoppTidigare

Vector analyzer, that is what you need. Scalar network analyzer will give amplitude without phase info.

what is your frequency? oscilloscope freq range within your transmit frequency of the transistor? or just a low freq oscilloscope?

I am not sure what you can do with the osciloscope, but if you borrowed a vector network analyzer, and the datasheet of the device does have the values for optimum load for maximum power, then you can design matching networks for the imput and output that makes the device see the optimum impedances listed in the datasheet...

How much power you are talking about? 1W? 100W? what frequency? Looks to me like a PA design task.. not very easy, and may need more than just a network analyzer...

Hi
I think u can check one of the old Microwave books like collin ... he give a direct relation to transfer from Z matrix to S matrix
it's so easy
Thanks
Amr A.Hamid

[quote="Gemini1706"]

<snip>

< No way to do that without a device that measures amplitude AND phase,
< which is basically a good old network analyzer. (or something similar
< that he can construct like using Analog Device's new phase/gain
< measurment chip up to ~2GHz only, but he needs then to worry about
< calculating mismatch losses...etc.

You thinking AD8302 ?


< Pretty much he will need then to build a full network analyzer :) )
< Very difficult task indeed.

Yes, but guy from Cypress microsystem try to build
simple (incomplete in my view) VNA based on AD8302
for around 200S of component cost.

see "http://www.cypressmicro.com/support/appnotes/an2090.pdf"

But, this project (ie AD8302) lacks full 4-quadrant phase
measurement and a more or less unusable to extract
complex S-parameters for feed to smithchart. Just now
can only se data like horizon (real line) folded smith chart
and can not easy decise between inductive or capacitive part.