LTspice Integration Method -- Warning!

Thread Starter

Papabravo

Joined Feb 24, 2006
21,157
I have been experimenting with a half-bridge inverter using STP8NM60 MOSFETS. I had some results that were extremely puzzling and unsatisfactory. After eliminating all the factors I could think of with no positive outcomes, I decided to check the LTspice Control Panel, and noticed that I had at some point in the distant past set my default to "Gear". When I changed back to "modified trap" my "supposed" problem vanished entirely. In true Holmesian fashion when you have eliminated all the other possibilities, whtever is left must contain the explanation.

1637345758547.png
 

eetech00

Joined Jun 8, 2013
3,856
Expanding on PBs post, its a good idea to use a directive when an option needs to be changed from its default for special simulation cases. For example, to set an option so that the simulator will use the "Gear" method, the following directive can be placed on the schematic:

.options method=Gear

This will overide the default "Modified Trap" method for the current simulation and, instead, use Gear method.
 

Thread Starter

Papabravo

Joined Feb 24, 2006
21,157
Can you describe those results?
Yes. I had a Half Bridge Driver connected to a pair of MOSFETS. Instead of the normal signals at the gates there were a series of short sharp pulses. See my post #3 for the a link to the author of the subcircuits. I have ziped up the subcircuits and two test program(s):
  1. MyHBDriver -test1.asc works fine with all the normal defaults. Particularly the default method of"modified trap"
  2. MyHBDriver - test1_Gear.asc is the same file with the addition of the .options method=gear
I know you are still using exclusively LTspiceIV, so I'm kind of interested in your results.
 

Attachments

Alec_t

Joined Sep 17, 2013
14,280
Instead of the normal signals at the gates there were a series of short sharp pulses.
What I've found when using several 'A' devices in creating a Spice model is that glitches/hangups can occur if all devices have the default 0 propagation delay. I assume this is because of race conditions (although I'm no expert in the workings of Spice). Giving at least one of the devices a few nanoseconds delay usually solves the problem. I note that the 'A' devices in Papa's linked circuit have non-zero delays. This may or may not have any bearing on those weird pulses experienced, but I thought I'd add my 2 penny-worth.
 
Top