I have been using intensively LTSpice in the last week always in the same PC. I find that when I change the circuit configuration adding / deleting a connection or components and ask for a new run, I have high chances that it will freeze in the first steps.

After many failures I found that clicking on the "save" icon, prior clicking on "Run", the freezing is almost not happening again. Is it any possible setting that I could have wrong? I discarded everyghing related to time step because the circuits involve more or less similar components and basically change little by little, and, after saving, it starts and simulates normally.

Comments appreciated.

 

How big is the circuit (number of nodes, components)?
How big is the .raw file?
Are you saving every node voltage and branch current, even in .subcircuits?

 

How big is the circuit (number of nodes, components)?
How big is the .raw file?
Are you saving every node voltage and branch current, even in .subcircuits?

Right now, I do not know how to find out how big.

The .raw one used to be long thus heavy. Simulated time, was progressively shortened for 600ms of real time but error persisted. Even in simple projects.

I kept settings as LTSpice was installed.

Why all this should produce a freezing?

 

Likely not freezing. You are just not waiting long enough for the sim to finish and finish writing the .raw file. What does the status message in the lower left corner of the schematic window say as the sim progresses?

As the circuit gets more complicated and the sim run times get longer, the .raw file (which captures the time history of node voltages and branch currents) for subsequent plotting gets bigger and bigger.

There are settings to avoid saving every node voltage and branch current. It is time for you to get more familiar with these settings...

 

On several occasions I've changed a component value while a sim is running, then foolishly forgotten to stop the sim before clicking 'Run' again. That's guaranteed to screw things up and likely get a 'not responding' message .
On other occasions I've waited for a sim to stop, made only a slight alteration to a value, then found that the sim fails to converge or a 'time step too small' message pops up. Strange. Sometimes LTS can be persuaded to converge by adding a ".options gmin=1e-7" command to the schem (a hint I was given in another thread).
We live and learn .

 

Sometimes a UIC (Skip initial operating point) or "Start external DC supply voltages at 0V" in transient analysis can help.
Changing the Digital Integration Method in the Tools/Control Panel/Spice window from "modified trap" to "trapezoidal" or "Gear" can also help.

 

Could be RAM issue too.If RAM gets full its going to start storing data on virtual memory.

 

Sometimes a UIC (Skip initial operating point)...can help.

I take issue with this suggestion. Mike E. (the developer of LTSpice) says exactly the opposite. Read the help file entry for .UIC

 

Could be RAM issue too.If RAM gets full its going to start storing data on virtual memory.

Yes, if you only have 1G of ram or even less, then kill all other applications (like your email client and browser) before starting the LTSpice sim.

 

I think the freezing of LTspice is real. For example, this morning it hung at about 25% done. When I closed LTspice and re-opoened it with exactly the same circuit it ran to completion normally. This is not an uncommon occurrence for me. I have not seen that doing a save helps.

 

Likely not freezing. You are just not waiting long enough for the sim to finish and finish writing the .raw file. What does the status message in the lower left corner of the schematic window say as the sim progresses?

As the circuit gets more complicated and the sim run times get longer, the .raw file (which captures the time history of node voltages and branch currents) for subsequent plotting gets bigger and bigger.

There are settings to avoid saving every node voltage and branch current. It is time for you to get more familiar with these settings...

The totally frozen status line keeps informing the las step and the percent worked out up to then, usually around 1% or 1. something.

My changes are the type of adding / taking out a DC blocking cap.

Maybe a pattern: if the last thing I did was to zoom in the raw file, I can expect this happening more frequently. But not really sure.

 

I take issue with this suggestion. Mike E. (the developer of LTSpice) says exactly the opposite. Read the help file entry for .UIC

I know .UIC is stated to possibly cause problems in initial convergence but I've used it frequently and never have had a problem.
For example it's useful if you are starting a transient sinewave simulation at the 90° phase point but don't what the circuit initial bias calculations to set the circuit to the 90° voltage from the sinewave.

 

I have been using intensively LTSpice in the last week always in the same PC. I find that when I change the circuit configuration adding / deleting a connection or components and ask for a new run, I have high chances that it will freeze in the first steps.

After many failures I found that clicking on the "save" icon, prior clicking on "Run", the freezing is almost not happening again. Is it any possible setting that I could have wrong? I discarded everyghing related to time step because the circuits involve more or less similar components and basically change little by little, and, after saving, it starts and simulates normally.

Comments appreciated.

Hi
I've used LTSpice on XP, Win7 and Win8.
Can you provide info on your PC? OS, CPU, RAM, Disk and Virtual MEM size?

eT