Strange Switchercad behavior

[capslock]

To all Switchercad experts (Sonnya?)

Tried to simulate a XO network to get started with Switchercad. I entered capacitors as eg. 33u, 10u or 6u8, resistors as plain numbers, and inductors as e.g. 1m5 or 0m82, as is usual spice notation and should be recognized according to the online help.

This seemed to work. E.g. Switchercad accepted these as 33µ. The simulation outcome, however, was strange to say the least. I then re-entered some values in scientific notation or play decimals, e.g. 33e-6 and 0.0015 and saw the result become more realistic. Changing all capacitor and inductor values gave a plausible result. I cannot say which values were foul, but it was at least two of them. Anybody know the problem?


Also, how do I add a voltage source from the menu? I had to resort to stealing one from an example file.

Puzzled,

Eric

[capslock]

Ok, I have narrowed down the problem. It seems the engineering multipier should not be used as a decimal point in this dialect. The programm does not issue a warning and will even display a value, i.e. I enter 6u8 and the display will contain 6µ8 (even the edit window contains this value when I click on it again).

However, the interpretation by the program is unpredicatable. 6µ8 seems to get interpreted as 6µ. I have no idea as to how it interprets 0m82 because this is where the results get really strange.

Remains the question of how to enter a voltage source...

[andy_c]

Hi,
The author of SwitcherCAD (LTSpice), Mike Englehardt, might want to hear about this one. He's a regular poster to the newsgroup sci.electronics.cad and will very likely respond if you post there. He's really quite helpful.

Another possibility is the Yahoo user's group for LTSpice, http://groups.yahoo.com/group/LTspice/
You'll need to sign up for that group (really a mailing list) to be able to read and post. If you sign up with the "no email" option, you can read it on the web only, like a newsgroup. It's a great group by the way. There's goodies available in the files section. For example, Helmut Sennewald has written a utility to convert the .raw binary output to formatted ASCII text for import into other programs. I've used this to export swept Vout vs. Vin curves for a BJT output stage. I brought these into Excel, where I did a best fit straight line with them, using the difference between the curve and the best fit line to analyze crossover distortion vs. bias current.