Need help displaying linearity plot of circuit with LTSpice

Thread Starter

bordonbert

Joined Feb 21, 2012
40
Hi guys. I found you here on a Google search for advice on LTSpice. I am a new user to that software but have managed to work my way into getting very simple simulations drawn, up and running.

My current task is to try to use LTSpice to display the linearity of a basic 3 diode ladder network when a sinusoidal signal is passed over it.

I have a simple generator driving a resistor in series with my network and I am looking at the generator voltage and the resulting voltage across the ladder network. I can see the output voltage changing the sine's shape as it should, flattened peaks to a degree depending on the values of the series resistors, but I can't see a way of displaying a visual representation, effectively a Lissajous figure, of the changing linearity of the circuit. I would like to have the generator voltage displayed on one axis against the ladder voltage on the other. I should then see the transfer curve bending under the influence of the diodes. This will of course be a static display which I can then alter by trimming the circuit values.

Is this possible or am I on a wild goose chase once again?
 

Papabravo

Joined Feb 24, 2006
21,159
Linearity seems to be a popular topic this week. Can you give us your understanding of what a linear circuit is supposed to do?
Changing the shape of a sine wave is not one of those things a linear circuit is supposed to do, and diodes are inherently non-linear devices.
 

Thread Starter

bordonbert

Joined Feb 21, 2012
40
Papabravo. I understand what you are asking, I phrased it badly. All I meant was display the transfer function of the ladder network graphically. I realise the diodes will of course produce a non-linear transfer function, it is that non-linearity which I am trying to simulate and investigate. I am not trying to produce a clean amplifying circuit where ultimate linearity is a key goal, quite the opposite.

This is a popular way of shaping guitar signals to simulate valve amplifier overdrive characteristics. The simple method almost always used of using a single pair of back to back diodes to clip the peaks of both sides of the signal is too rough and raw to my ears. (I don't play Death Metal or Grunge, I'm way too old for that!) A ladder of perhaps 3 diodes, each with their own series resistor, (and for each polarity of course), could be used to distort/limit the signal more gradually but the values of the associated resistors would need to be carefully chosen to balance the various break points. Each polarity will be shaped differently to introduce sweeter even harmonics as well as the odd harmonics from the shared clipping characteristics.
 

Thread Starter

bordonbert

Joined Feb 21, 2012
40
Thanks Crutschow, that's exactly what I needed. Got it working and its displaying exactly what I was asking for. Many thanks for your help. I can already see a long way towards the solution of my problem.
 

MikeML

Joined Oct 2, 2009
5,444
If you sweep a voltage or current source and use it as input to your network, you might get what looks like a "linear" response. To see how it deviates from being truly linear, you can use a behavioral voltage or current source that is a function of the original swept voltage or current to create a function that really is linear. Plotting both the actual response minus the linear function makes the lack of linearity more visible...

For example, look at V(accel) {linear}, V(out) {actual, non-linear response} and the difference between the two: V(out)-V(accel) in this simulation
 
Top