FURTHER TO PACKAGING, 3D PRINTING, AND PACKAGE/INTERCONNECTION DEGRADATION OF MICROWAVE SIGNAL PATHS
Greg Merrell
Retired Sr Design Engineer/Project Manager at Teledyne Technologies Incorporated
This is an addition to the following post:
Following are some more links with related information.? Default simulation environments (small signal,? ADS, AWR Microwave Office, etc.) assume perfect package conditions, or semi-infinite above-circuit environments, uninterrupted and unperturbed ground planes.? In many situations these effects are negligible. The challenge is to always be aware when they ARE negligible, and if they are not, how to handle it. Feedthrough pins in housings can be inductances that mismatch. MMICs can have port impedances that are inductive, and are only further degraded by the addition of a bond wire. MMIC mounting grounds can be connected to housing grounds through steps and dips in a ground plane before reaching the output at the feedthrough. Depending on their lengths, with respect to operating wavelength, the effect can be negligible or very damaging. Outside world return losses can be destroyed, and in-band transmission characteristic ripples can appear (large ones). In extreme cases package and MMIC signal interactions can result in positive feedback with amplifying port reflections (reflection loss of opposite polarity vs. a good match)? and/or oscillations.?
Companies and engineering departments have varying levels of robustness in their design of packages and microwave device integration. Some have (have had) design libraries with hardware guidelines and templates to implement and manufacture for good performance.
Bond wire inductance:
A good rule of thumb is that for one-mil wire bonds (25 micron diameter), length in millimeters is nearly equal to inductance in nano-henries. Note that for this equation, L and D are in mils (thousandths of an inch).? ? Microwaves 101
The guideline I was told, and used, was 0.067 nH per mil. For a 10 mil wire, that would be 0.67 nH.
The above Microwaves 101 guideline would lead to 10 mils = 0.254 mm = 0.254 nH.
I sometimes checked the geometry against a round wire above an infinite ground plane. Shorting one end of the wire, calculating its characteristic impedance, then calculating the impedance of the line (reactive, inductive) would then lead to an inductance and inductance per mil for that geometry, given the frequency. The measured result was usually reasonable.
领英推荐
https://ieeexplore.ieee.org/document/4139566 (requires IEEE membership for access).
Minicircuits AN75-011. A Practical Approach to the Design and Implementation of Scalable, High-Performance, Custom SMT Packages for mmWave Applications
Senior Telecom consultant
1 年Greg, your expertise in RF and Microwave field will be useful to many. I hope a new generation of engineers follow you and get benefited that is what industry needs today. RF, Microwave and mmWave industry is relying on ageing population and I hope this changes soon.