Often when I run a .step param simulation, I get a "QNANA" error in one or more of the steps, even though they may be close together in value. How do I avoid this?
I also find that the LTspice simulator sometimes does this with a fixed parameter, NOT stepped. Then if I change it by just a tad (~0.1, say), it will run fine. I don't understand this rather peculiar behavior. Anyone?
I've never seen this in LTspice - but my understanding is that NAN stands for Not A Number. The last A may stand for Amps, i.e. there's a current generating the error. There's a thread about a QNANV error over at the users' group.
Last edited:
"....my understanding is that NAN stands for Not A Number."
I understand that as well---so it cannot compute that particular circuit. But, as I said, if I change a value by even a small amount, and then run it again, sometimes (not always) it works fine---which makes no sense---there is no 'divide by zero' or other causes and nothing obvious to me that triggers this error.
Do you have a link to that users' group?
I understand that as well---so it cannot compute that particular circuit. But, as I said, if I change a value by even a small amount, and then run it again, sometimes (not always) it works fine---which makes no sense---there is no 'divide by zero' or other causes and nothing obvious to me that triggers this error.
Do you have a link to that users' group?
I tried that group, and even signed up for it. But it doesn't seem to DO anything---clicking on any of the links on that page just loops you back to the same page over and over again. Is there a trick?
There's the messages link top left below the home icon, try clicking on that, you should then see the messages to the group, then at the top right will be the search box or try this direct link to the messages page. I'm not a member of that site so I don't know what happens when you sign-in.
The link for the actual discussion of this error is
LTspice@groups.io | Issue with .#QNANV
Here's another one that mentions QNAN errors:
LTspice@groups.io | How to fine run simulation in LTspice with SiC SCT3022KL model ?
If you click on Messages on the home page and do a search for QNAN, a whole load of threads are listed that relate to this error.
dotneck335 said:
I really can't imagine why you are having problems. Have you made it work now?
Might try asking this question in the LTSpice thread here:
Installing and using LTspice IV (now including LTXVII). From beginner to advanced.
Installing and using LTspice IV (now including LTXVII). From beginner to advanced.
Thanks, Kevin---I did just that. We'll see if I get a usable response. Have you ever seen this result in any of your simulations?
In fact I do see that error, but rarely - it's more common when I simulate tube circuits. I usually look for errors I have made, and sometimes I just need to slightly adjust a value to fix it. I have no actual idea what causes it, but I am assuming it is a math related error. Sometimes changing the solver or selecting a less rigorous set of solver parameters solves the problem.
I haven't seen the error lately though.
It's oddest when you have an actual circuit on the bench that works well and LTSpice tells you it doesn't work at all.
I haven't seen the error lately though.
It's oddest when you have an actual circuit on the bench that works well and LTSpice tells you it doesn't work at all.
That's the CRAZY thing about the #Q NANA error---sometimes, if I change a resistor by as little as 1 ohm, the sim runs perfectly. In a range of values, some work and others don't---there doesn't seem to be any pattern to it that I can discern. It is some sort of math error; I just wish I could figure a way to minimize the NANA occurrence.
I suggest uploading diagrams with such a disaster, so that we can collectively find ways to combat it.