Is it possible to realize a 0.2nH bondwire indutor?

Hi all,

Currently I am designing an LNA around 2GHz, and I want to use a bondwire inductor about 0.2nH as the emitter degeneration part. Is it possible to realize it by multi bondwires in parallel? I need it to make a fast estimation, any suggestion is appreciated.

For the simulators, HFSS and ADS EMDS? Which one is better to simulate this case? Thank you very much.

Sure. You might get better results with a short ribbon bond.

0.2nH is very low for bonding wires even they are connected in parallel.A typical bonding wire has approx. 0.8nH ( I talk about HVQFN package) and if they are connected parallel with for instance 4 wires, they won't give you 0.2nH because of mutual coupling.
If I were you, I make this coil on the chip.It will occupy an amount of space but it will have higher Q factor and will be stable against stress and temperature.
Personally, I don't have so much confidence to boding wires.

in addition to wire bond , can you have microstrip with via in parallel. I may throughing stupid suggestion. I guess the parallel effect could be adjusted to 0.2nH.

BigBoss said:

0.2nH is very low for bonding wires even they are connected in parallel.A typical bonding wire has approx. 0.8nH ( I talk about HVQFN package) and if they are connected parallel with for instance 4 wires, they won't give you 0.2nH because of mutual coupling.
If I were you, I make this coil on the chip.It will occupy an amount of space but it will have higher Q factor and will be stable against stress and temperature.
Personally, I don't have so much confidence to boding wires.

Click to expand...

Yes, I am using HVSON lead-free package and the downbond is about 0.8nH actually.

Because this single-ended LNA is really large-area (Yes, differential arch doesn't suffer from the bondwire problem because of the virtual ground but I can't use it in this design.), I found it is really hard to realize an input impedance/noise match at the same time if the emitter inductor connected to the ground is larger than 0.01nH. This is impossible to realize a 0.01nH.

So in order to solve this problem, I added a shunt capacitor in parallel with the Cbe (Cpi) to decrease the cutoff frequency rudely and a transformer feedback to change the input impedance, then the emitter inductor could be about 0.2nH-0.3nH. Well, the power gain decreased inevitably. But I am still not sure if the 0.2nH-0.3nH is realizable.

For your suggestion, I don't know if I understand it. Because there is always a bondwire between the chip-ground and the package ground, even though you can use an on-chip inductor or microstrip inductor. How to solve it?

Thank you all.