2

I am working on a parasitic steerable phased array antenna, designed at 15 GHz. I need to terminate one of the array elements with a 0.01 pF chip capacitor to achieve desired steering direction. I have explored a lot, but sadly I could not find a 0.01 pF chip capacitor available in the hardware with any vendor. The least I could find was 0.05 pF manufactured by AVX RF. Does anyone know of any vendor who manufactures 0.01 pF chip capacitor having self resonant frequency (SRF) more than 15 GHz? Also, I need another chip capacitor of 1 pF with SRF more than 15 GHz. 1 pF capacitors are easily available but the SRF is far below 15 GHz and hence cannot be used in my design. Any suggestion or solution to deal with this issue would be highly welcomed.

HSU
  • 21
  • 3
  • 4
    sounds like you need to design with geometry rather than components. A 0.01pF capacitor is a short length of open transmission line, the pads of an 0201 component would contribute more stray C than the component itself. If you're stuck with wanting a lumped 1pF at 15GHz, then you need to redesign with transmission lines. – Neil_UK Mar 30 '18 at 10:38
  • More importantly, 0.01 pF is such a small value, that if you haven't very carefully designed your board to avoid parasitic capacitances, there could easily be a higher value present already. You should probably be re-designing your circuit so that adding or removing of a few 0.01's of pF don't affect the circuit much. – The Photon Mar 30 '18 at 15:25
  • Have you used the parallel-plate formula C = E0*Er*Area/Distance for this purpose? How about the wire-over-plate Capacitance (which I don't have memorized)? You want parasitic-type thinking here, unless you are developing custom silicon for this. – analogsystemsrf Mar 30 '18 at 17:17

2 Answers2

3

You are talking about so small capacitances that you don't need any discretes to achieve them. According to [1] the capacitance of an 1206-size components pad is about 0.72 pF. Therefore there is no point in creating discrete capacitors smaller than about 10 pF because it is easier to embed them into the PCB and you get smaller parasitics too [2].

The second reference I used talks about some fancy dedicated PCB materials for embedded capacitors, but you can also use regular FR4 PCBs. Howevere, you are most probably going to need impedance control. Calculators and equations for PCB embedded capacitors are readily available by googling i.e. http://chemandy.com/calculators/rectangular-capacitor-calculator.htm.

[1] http://www.sigcon.com/Pubs/edn/parasiticpads.htm

[2] https://multimedia.3m.com/mws/media/251029O/fabrication-of-embedded-capacitance-printed-circuit-boards.pdf

Trafi
  • 435
  • 3
  • 8
0

At 15GHz nobody uses lumped element components, 10ff sounds to me like a very tiny stub or triangle feature etched above the ground plane, are you sure your fab is even capable of that level of precision etching (And is your substrate sufficiently well qualified for that to make sense)?

1pF you might manage something in 010005 or such, but it still sounds like a printed feature to me, the loop area due to coming up away from the ground plane will introduce significant inductance... I would print the thing on the board (Which had better be a THIN RF substrate).

Dan Mills
  • 17,266
  • 1
  • 20
  • 38