Wiring/Ribbon bonding between two different Er substrate

Hi,

I am curious,

if i wire/ribbon bonding, two 50 ohm lines of two substrates (with different Er) together:

1)Ideally, is a 50 ohm matches with 50 ohm, but both transmission line have different width (different Er), will it affect and degrade the performance?

Also the wire/ribbon will tend to introduce some inductances between the two substrate, and that may cause some mismatches, any idea, how to reduce the mismatch effect cause by the wire/ribbon? Anyone had the experience on such issue? I heard the use of pad to reduce the mismatch effect cause by the wire/ribbon, anyone know how to go about? And how to model the wire/ribbon bonding in circuit simulation eg, ADS?

Hi John,

The bondwire will introduce some inductance as you said. And the inductance will increase with frequency climbing(or a thinner and longer wire).

To reduce the affection to your circuit performance, you can use a compensating circuit to form a LPF structure(LCL or CLC 3rd order low pass filter). As you mentioned , pads will be introduced in this compensating circuit. Those Pads at each 50ohm TL which your wire attached will act as capacitance and the wire itself will act as inductance(C-L-C structure). To compensate the draw back of your wire, just set the cut-off frequency of this LPF above your operating frequency which can balance your wire affection to the circuit return loss.

Best regards,
K

You must imagine what happens to the electric fields in the vicinity of the transition. In the lower er substrate there is more E field in the air, and less in the substrate than in the higher er material. So, even though both transmisson lines are "50 ohm", there is a mismatch of electric fields in the two media.

Now microstrip is NOT a TEM structure. That means that the apparent solution to the fields are actually a summation of many different modes. This change in Efield at the transistion can force some of the power into non-propagating modes. These non propagatin (evanescent) modes act as a lumped reactance to your circuit.

So, you might have to do some sort of minor matching structure between the two substrates, even though by themselves each looks like a 50 ohm line.

Thanks for enlightening me.... seem to have a slight overview on what is going on... For both substrates of diff Er, thought have the same 50 ohm.... may still have some mismatches due to the field effect...

A more obvious example would be going from a 50 ohm coaxial connector to a 50 ohm microstrip line. There is a relatively big discontinuity there, which is typically modeled as a finging capacitance and a series inductance.

Hi
You should use HFSS to simulate or optimize such structure