choosing a chip inductor for Microwave/RF applications

[per_lube]

Hi all,

I have to use a SMT inductor (chip) for a biasing circuit of an amp operating at 10 GHz (to provide high impedance path to RF signals).

When choosing an inductor for this application:
1. what is the importance of Q?
2. what is the importance of self resonant frequency?

for my application, could someone please explain, what are the suitable components, Low or High Q inductors?
what should be the self resonant frequency of the inductor to use in a cct. operate at 10 GHz?

cheers,
per_lube

 

[D.A.(Tony)Stewart]

Q is determined by lossy series resistance or parallel leakage resistance whichever is worse. For decoupling LPF it is less important as you are designing a low Q bandstop filter (LPF)
SRF is very important as self capacitance SRF must be at least 5x break point of LPF. and ESR/ ESL of Capacitor must be very low. Often decoupling Caps of these frequencies are in low pF range and sourced by RF makers such as Murata.

 

[biff44]

I am not aware of any "chip style" inductors that work to 10 GHz. Most have resonant frequencies in the 1 to 3.5 Ghz range. Above resonance...you can no longer call it an "inductor".

A resistor may work. A distributed type bias tee may work. Some sort of printed spiral inductor may work, but needs to be analyzed properly.

 

[D.A.(Tony)Stewart]

Here are four 0402 inductors with SRF of 20GHz from Murata

https://psearch.murata.com/inductor/result/?size_l_w=W02_007&u_srf=ghz&v_srf=20&cat=i_spec

 

[biff44]

Well, you learn something new every day! 20 GHz chip inductors!
and 0402 isn't that small.

 

[jiripolivka]

For MMIC applications above 8...10 GHz, some companies sell conical chokes that do work up to >60 GHz. In my designs around 5...7 GHz I have used simple wire coils, with 0.15 mm dia wire wound on a 1.5 mm drill, then cut at 10-15 turns and pulled to extend the coil a bit. Yo can try the same. Using a swept-frequency test you can see if there is a resonance; by pulling and pushing the coil you can see if the resonance peak is caused by it, and modify the size and number of turns for best results. From my experience, capacitors cause the worst problems. Instead of using 1-nF blocking capacitors I found better in paralleling two-three 300-pF SMD capacitors. Now I saw some company offering "high-frequency" SMD caps, with large metal ends and a narrow ceramic gap left. Be ready to experiment.

 

[D.A.(Tony)Stewart]

Murata recommends the use of 100pF with high SRF for microwave and then for low frequency noise, of course bigger values can be used.
For every value on the V+ bus, SRF must be examined or impedance or return loss or insertion loss. Guessing is not allowed.

You may find these caps are metal film type not ceramic and also find metal film have the best ESR,ESL at lower freq but become very expensive so electrolytics are used when large values are needed > 1uF

 

[Ow@i$]

Why not to use ferrite core inductor or ferrite bead they can sustain high frequency i guess??
and why can't we use simply a resistor for this purpose!!
Enlighten me please!!!

 

[D.A.(Tony)Stewart]

ferrite core is not designed for microwave as the impedance is uncontrolled..

The issue is how to reduce coupling of 10GHz load with V+ by insertion loss and yet have low impedance for DC.
A resistor raised impedance but you want low impedance for RF load DC power source to improve performance.

If you understand that even a plated thru hole is an inductor at this frequency (10GHz) , then you can understand how physical coil and wire can be used for controlled impedance.

The S parameters are the preferred for each physical path and component, in order to provide the best power source to meet the supply ripple & regulation rejection from load to supply and visa versa.

 

[per_lube]

Thanks a lot for your great explanations....

is this self-resonance valid for the dc blocking capacitors used in between cct blocks in a microwave circuits (ex. in between two amplifier stages...)?

at this frequency do we need to worry about self-resonance frequency when choosing a cap. (say... to cascade two 10 GHz LNAs)?

cheers,
per_lube

 

[pstuckey]

A capacitor will always block DC - even above it's resonant frequency. However above it's resonant frequency it's impedance won't behave like a capacitor. Instead of decreasing with higher frequency it will increase (it will behave like a DC blocking inductor). You'll want the impedance to be low, usually less than a couple ohms, so you'll have to consider the rising impedance if you use the cap above SRF.