LTspice parametric sweep: how to force the simulator to continue with the next step?

I'm running transient simulations as I am sweeping a couple of parameter values (resistances, capacitances) to give a first hint about what range of values make a certain circuit stable. In addition, I want to extract certain measurements from the wave form, in order to basically give me a figure of merit for the given parameter combination. These measurements limit how short time period I can use for the simulation (as in, it turns out I can't go below about 200µs if I want the measurements from that run).

Each run takes a fairly long time, so I would like to avoid, if possible, running the whole transient simulation if I can already see that these particular values causes oscillations.

Is there any way to tell LTSpice that "no, this parameter combination is dead in the water, skip this and continue with the next"?

• Remember, LTSpice is ultimately derived from a program written in the batch processing era of computing. After you submitted your punch-card deck to the computer operator, you wouldn't expect to interact with the program until you got your print-out back the next morning. Some of the features of LTSpice still reflect the assumptions of that obsolete interaction model. – The Photon Sep 28 '18 at 20:30

It would be nice if programs could read our minds, but they can't. They will only do what you tell them to, within their conception limits. What @ThePhoton says is also true. But you can try to adapt to their way of thinking. If you say you have oscillations, you could add a check for the offending node, and if it's not behaving, maybe add a comparator (Schmitt trigger) that grounds the nodes, or some critical nodes that force the outcome to go to ground, thus forcing the simulation to calculate only zero values and going fast(est) until the end, then resuming with the next stepped value. If you have an oscillator, replace it with the builtin MODULATE(2), controlled externally, which can be disabled, too. Logical gates should be easy to disable. The disabled nodes can be the inputs, which will simplify things, but not entirely as additional elements will need placing, that deal with the interrupts. Otherwise, and if I am not mistaking, I don't think there are SPICE programs that have the requirement you need -- though I'd be glad to find out I am wrong.