how to measure input impedance

Hi all,

I have to design an antenna to fit onto a microchip. The problem is I do not know the input impedance of the microchip. How can I measure the chip input impedance so I can design my antenna to match the impedance of the microchip.

Any help will be appreciated.

Thanks in advance.

Cheers,
Element7k

Hi Element7k

Theoretically you can solder your chip to a PCB, connect a 50 ohm trace to your antenna port and measure the impedance with a network analyzer. You have to pay attention that your chip is driven correctly (all powers connected, not in a sleep mode etc.)
But mostly the chip is already matched to 50 ohm or has a very high impedance so you can use a shunt resistor. What kind of chip do you want to use?

Bye

Hi ghz,

thanks 4 the reply. I do have any data about chip. :( It is unmarked. I was only asked to design the antenna. The chip is already on a pcb with other components. However, when I mounted a close to "50ohm" antenna to the input of the chip, I got very poor performance. I suspect I have to match the antenna to the chip impedance. Are you suggesting I should detach the chip and connect it to 50ohm trace? I will have the problem of having to figure out what each leg of the chip is connected to then?! Supposing that this can be done, what should be the setup to measure the chip impedance then?

Many thanks.

Cheers,
Element7k

If you have lines for matching it's a transistor.

Post here your scanned or snaped pcb layout.

Hi element7k

Try another way: I guess it's only one device you have to optimize? Connect a transmitter over a big attenuation directly to your receiver. Measure your (bad) demodulated output signal. Now you can try to optimize the receiver with stubs on your RF line. These stubs will change the input impedance of your chip. Theoretically you will find a point with best S/N performance. This is optimal match to your 50ohm input device.
One point you have to take care about is that your RF line has to be lambda/4 in length or more.

Bye

Thanks mr_ghz, sounds like a brilliant ideal but can you suggest an easy way where I can estimate the approximate dimension of the tuning stub?

Cheers,
Element7k

Hi Element7k

Try a lambda/4 stub (width=50ohm) in a distance of lambda/8 of the chip. First cut the stub for best performance then move the stub. When you found optimal position try longer/short stubs and so on.
Question to other users: Who has experience in 'experimental' matching a circuit?

Bye

Looks like the best way to do the measurement you want is to use a variable tuner connected to a directional coupler and the isolated port of the directional coupler connected to a power meter or spectrum analyzer.
Also a RF Gen should be connected to the other end of the tuner with the right level to drive the device.
So, change the tuner to get the minimum signal on the power meter. Disconnect the device from the tuner and measure the tuner impedance using a VNA. The conjugated of this impedance is exactly the input impedance of the device.
You can buy variable tuners on e-bay. I've seeing some last week.

NandoPG