Minimum transmission line length

[GSarris]

Hello
Can i use a transmission line that has a length of 10 mils? Are there manufacturers that can achieve it? I want to insert a small inductance at the source of my transistor and a small transmission line is a solution.
Thank you

 

[albbg]

Did you simulate the beahviour of such a short line ? I think parasitic effects and inaccuracies could dominate. However I don't think manufacturers have problems to go down to 6 mils.
For curiosity what about the working frequency and the inductance you need ?

 

[GSarris]

albbg i agree with you about the parasitics..that's what troubles me too. I want an inductance of about 0.18 nH.

 

[jiripolivka]

Everything depends upon the signal wavelength you use. If you need 0.18 nH inductance, you can get one in SMD, or create yours by adjusting line length using the Smith diagram.

 

[tony_lth]

Some vias have inductance about 0.2nH, I saw someone use two vias to get 0.1nH.

 

[FvM]

You are talking about a piece of wire, either bond wire or PCB trace. You rarely need to employ a "manufacturer" to implement it.

Seeing the problem from the other side: You have plenty of unwanted inductance in most circuits. It's a more common problem to ignore the already existing inductances in a simulation by using a raw chip model without package parasitics or forget about circuit wiring inductances. But if you actually need to increase the "ubiquitary" circuit inductance, it's more than easy.

 

[GSarris]

Thank you for your answers. I thought of vias too, but my substrate is 1.524mm thick and the model i found on a Paper by Goldfarb and Purcel was for a sunstrate of 600um thickness max. Have you got any other references to suggest? Furthermore, i would like to ask you about the ADS components Ribbon and Wire. What are they exactly? Since they don't need a sunstrate are they being printed on the PCB? Or are they simple wires that are connected on the PCB after fabrication?

 

[biff44]

a transmission line is a uniform structure that can be modeled as a distributed series inductance per unit length and shunt capacitance per unit length.

If you have a very short length of "transmission line", where "short Length" means much less than 1/8th wavelength, it can be modeled as a lumped element.

For instance, if you have a very short length of 200 ohm transmission line on microstrip, it acts very much like a series inductor. Similarly a short length of 20 ohm microstrip will look like a shunt capacitor.

 

[GSarris]

I agree in general with what you say, but if you want to implement on microstrip a 200 Ohm line, this will be very thin (very small W). So thin, that firstly your response will suffer from fringing field effects (imagine a capacitor that the distance between its plates is much bigger that the area of the plates) and secondly the models that are generally used for microstrips are for negligible or zero conductor thickness (thickness<W/2 according to Gonzalez). If you design a 200 Ohm line, you will end up getting a width of some microns. Given that usually the conductor thickness is 1-2 oz, i don't know if the above assumption stands. So a correct simulation depends on the accuracy of the models that ADS uses...Generally i avoid very thin lines, to be on the safe side of the models as sometimes you cannot be sure about how they treat such extreme cases ...