diode impedance
Can someone give me a clue about the best way to measure the impedance of a Schottky diode? It's an Avago HSMS-285Y in an 0402 SMD package. There's a circuit for testing shown in "Approximate Antenna Modelling for CAD" by Visser, but I've tried building it and it doesn't work.
I've got access to a VNA, a spectrum analyser and as many other parts as I need...
Any ideas, or does someone already know the impedance, or a good way to estimate it?
Thanks,
________
Agronski.
You should get this impedance by simulation, if you have a model. It's easy and give you something close to measurement.
I checked, you have the spice model in the datasheet.
Use it in a simulator and go on...
Diodes are very simple things, but hard to measure. That is because they change their characteristics dramatically depending on the RF input power and load impedance on them.
If you just try to measure a schottky diode mounted in-shunt, with a standard network analyzer, you would probably see that it looks like a pure 0.1 pF capacitor. If you instead used a high power network analyzer, so that the diode was rectifying the RF incident on it into DC current, and there was a DC load connected through it via a bias network, you would see something that looked more like a larger capacitor in series with 30-200 ohm resistance. But even that would be a suspect impedance measurement, since if you then designed and added a matching network in front of the diode, its operating point will change again, and give you a different result.
Thanks for those - a bit of a follow-up though:
Small-signal modelling isn't my strongest subject, so I found an old hardware test-rig for varactor diodes - my guess is it'll work well enough to test Schottky's too. It's a 50Ω microstrip line: One end to the VNA, one end shorted to ground. Then you replace the short with the diode, and you can subtract the the combined diode+microstrip S parameters from just the microstrip and bingo!
....one problem, though - I'm not 100% sure how to subtract S parameters from one another. My VNA only gives me a Smith chart, dB return loss and SWR anyway. Is the subtracting one set of S parameters from the other the so-called "de-embedding" I've heard about?
________
Agronski.
Don't know if it's right, we normally use a very good computer-control power supply to measure the impedance. Just sweep the voltage from, say, -0.5 to +1v and measure the current to plot the DC-IV curve of the diode. Then calculate the gradient of the curve to work out the impedance, and you could find the responsivity from that too.