LTSpice begginer

[PGC22]

Hello guys! So I'm not familiar at all with LTSpice and I didn't learn any in school.The thing is that I want to make a RLC model for a coaxial cable and calculate the impendence of the cable, but I really don't know where to start actually, in the attachement is something that I tried to do. If you guys could help me with some tips it'll be great. I know that maybe is a simple thing to do, but I really want to learn these things and the internet so far has not been very helpful ( english is not my first language). Thank you!

[Ratch]

Hello guys! So I'm not familiar at all with LTSpice and I didn't learn any in school.The thing is that I want to make a RLC model for a coaxial cable and calculate the impendence of the cable, but I really don't know where to start actually, in the attachement is something that I tried to do. If you guys could help me with some tips it'll be great. I know that maybe is a simple thing to do, but I really want to learn these things and the internet so far has not been very helpful ( english is not my first language). Thank you!

Try this site.  https://www.systemvision.com/

Ratch

[MagicSmoker]

Hello guys! So I'm not familiar at all with LTSpice and I didn't learn any in school.The thing is that I want to make a RLC model for a coaxial cable and calculate the impendence of the cable, but I really don't know where to start actually, in the attachement is something that I tried to do. If you guys could help me with some tips it'll be great. I know that maybe is a simple thing to do, but I really want to learn these things and the internet so far has not been very helpful ( english is not my first language). Thank you!

The first thing you need to do for pretty much any SPICE program is define ground. So add the ground symbol (upside down triangle) to your schematic and connect it to one of the nodes (usually the negative terminal of the voltage source).

Next is to set up your voltage source properly. Right now you have it configured with a 12V DC offset, a 1V amplitude, 1kHz frequency and a 0.1 second delay, and none of this appropriate (except, perhaps, the amplitude)! The DC offset should be 0V for a pure AC source, which is what you want here. The amplitude setting defines the peak change in voltage in either direction starting from the DC offset (which, again, should be 0V). So if you want a 10V peak to peak sine wave then set amplitude to 5V. If you want a 10Vrms sine wave then set amplitude to 14.14V... etc. The frequency also needs to be much higher for this particular task - since you like scientific notation then 10MHz, or 1E7, would be more appropriate than 1kHz. However, LTSpice also recognizes the common multiplier abbreviations like n = nano, p = pico, u = micro (which will be converted to mu), k = kilo and in order to distinguish between milli and mega, m = milli while Meg = mega.

Finally, the default type of simulation in LTSpice is "transient", in which you define a time period for the circuit to run, whereas with most SPICE programs you would probably select something like AC Analysis or AC Sweep. However, you can get meaningful results with transient analysis, so why make your life difficult? In this case, set a simulation time of, say, 10us (1E-5) then find the RMS value of the current through any of the components by holding down the Ctrl key then left clicking on the component.

With the voltage source set to 14.5V amplitude and 10MHz, I get an RMS current value of ~230mA at an RMS voltage value of 10V, so the impedance is ~43.5 ohms.

[PGC22]

You helped me allot. Thank you very much!

[Zero999]

The lumped model you've described for a co-axial cable isn't very accurate at higher frequencies.

Fortunately, LTSpice has a built-in model for a transmission line, called tline. If you place the symbol, right click on it, you can adjust the properties: Td = the time delay and Z0  = the characteristic impedance of the line.

[MagicSmoker]

The lumped model you've described for a co-axial cable isn't very accurate at higher frequencies.

Fortunately, LTSpice has a built-in model for a transmission line, called tline. If you place the symbol, right click on it, you can adjust the properties: Td = the time delay and Z0  = the characteristic impedance of the line.

Yeah, I was going to mention that, as well as using a pulse instead of a sine wave, and trying out both source and termination resistances to see the effect of reflections... but I wanted to make sure the OP got the basics down first.

[Ian.M]

LTspice can also handle fairly long multi-section lumped lines using its bussed wire and array component notation without having to manually repeat and connect the elements.  You just have to draft one element and connect it with busses.   Because its not using the built-in tline model you can probe anywhere within the line.

Here's a 256 element example:


N.B. make the component matrices and create bus labels *BEFORE* wiring in the busses, otherwise it may not connect them as you expect.   See here for the notation.  It can also get a bit confused and try to use alternative forms of the bus notation.  If that happens its probably best to edit the names to the canonical form,

Edit: updated attachments to get bus names to canonical form.

[darrellt]

Here's a blog that talks about how to use the SystemVision transmission line in a circuit application. 
https://www.systemvision.com/blog/it-wire-or-transmission-line-february-6-2017

Or you can go directly to the application circuit online:
https://www.systemvision.com/design/transmission-line-fed-led-driver-switching

[Zero999]

LTspice can also handle fairly long multi-section lumped lines using its bussed wire and array component notation without having to manually repeat and connect the elements.  You just have to draft one element and connect it with busses.   Because its not using the built-in tline model you can probe anywhere within the line.

That seems like a very useful feature but if you want to probe in the middle of a transmission line, then wouldn't be easier to use two transmission lines in series and put the probe between them?

[Ian.M]

if you want to probe in the middle of a transmission line, then wouldn't be easier to use two transmission lines in series and put the probe between them?

Far far easier, and almost certainly more accurate, which is good if your concern is solving the problem the transmission lines represent, but it wont teach you much about modelling a transmission line as lumped elements as all the workings are hidden 'under the hood'.