my question is what H (substrate thickness) should i set in the MSUB? Spiral should not have any ground plane underneath. due to this, shall i set the H to be the total H of the PCB since im placing the spiral on the top layer?
Dr Volker, thanks for your comments. just one input. L value on 2nd value will be higher than the 1st layer due to T difference. 1st layer is 0.7mil whereas 2nd layer is 1.4mils. thank you very much.With no ground below, the difference between your two inductor placements (top and middle layer) is only the effect of FR4 on parasitic capacitance betweem the turns. Or in other words: self resonance frequency. But I suspect that the "abused" microstrip inductor model isn't very accurate anyway, so I wouldn't worry about that second order effect.
~~
It's good to have no ground below, but I wonder how accurate the model is in that case. I would set a very large height for the model's MSUB, to minimize the (missing) eddy current effect of the (missing) ground underneath.
Looking at your previous layouts, I think you will have a large effect (inductancce error) from metal on other layers that creates a closed loop around your inductor. That isn't included in your microstrip inductor model, so you need to check that using EM anyway.
just one input. L value on 2nd value will be higher than the 1st layer due to T difference. 1st layer is 0.7mil whereas 2nd layer is 1.4mils.
since we are removing all the layers underneath, I believe the H has to be the total H of the PCB.
I don't understand why you worried about the accuracy.
and also increase the inductance because 2nd layer T is double the T of top layer.
For best inductor performance we want to remove the ground. The microstrip inductor model is designed for microstrip inductor. That's the difficulty. I have doubts how accurate the model is when removing the ground, because it was not designed to for that case. My proposal is to set H to a large value (ground far away), but I never tried how good results are then.
LnH shows clean response without any resonance. please see the pictures below.
Ri=20mils ( 0.508mm)
S=3mils ( 76.2um )
W=3mils ( 76.2um)
These values are absurd.
In order to get a realistic values from ANY model of a component ( active or passive ) the dimensions ( or other parameters ) should be constrained by the model itself.
I mean,the model cannot carry on its validity using with ANY dimension as you wish.If you work at RF/MW frequencies, these oddness may augment.
its a PCB spiral inductor. due to this, I cannot go below 3mils on trace thickness and the gap.
i dont have ground underneath my inductor. i used 3mils to reduce the side of the inductor. i believe increasing thickness of the spiral and the gap will reduce the SRF. please correct me if im wrong.
target inductor value and implementation is given below.
2.2nH - spiral
0.2nH - stub
You may use also a Serpentine Inductor instead of spiral here..
2.2nH is physically small using PCB technology, I think there is no need to meander the line.
Straight line on FR4 (no ground on backside) with w=0.2 mm and l=2.2mm gives 2.2 nH.
View attachment 151403
Straight line on 1.5mm FR4 (with ground on backside) with w=0.2 mm and l=2.2mm gives 1.9 nH.
View attachment 151404
im trying to replicate this in ADS but im unable to. how to design stub inductor without any ground at the background? MLIN and CPWG have ground underneath. only CPW dont have ground underneath?
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?