Enhanced LTSpice Model Simplifies GaN Power Designs

For power-system design engineers who want to be fast, accurate and confident with their simulated designs prior to building hardware, GaN Systems’ latest set of LTSpice models are available for download that allow for a variety of inputs and simulations options. Users select the product of interest and then select the LTSpice button. Additionally, LTSpice application notes GN007 and GN008 are available on the GaN Systems website.

 

The LTSpice model user guide helps engineers model systems at three levels, ranging from an initial overview of circuit performance to detailed analysis and fine tuning of the design:

  • Level 1: Basic adjustment and analysis of switching speeds, optimized for quick simulation.
  • Level 2: In additional to Level 1 features, includes thermal inputs and Cauer thermal RC network transient models for simulating the device junction temperature and self-heating effect.
  • Level 3: In addition to Level 2 features, includes parasitic losses, provides the most accurate model with longest simulation times.  

To confirm the accuracy of the LTSpice model, laboratory measurements of GaN E-HEMT switching losses were recorded using a half-bridge, double-pulse test circuit. The switching losses measured in the test were then compared with the LTSpice model simulations. The comparison demonstrates a strong correlation between the simulated results and real-time circuit measurements.

 

With a 400V, 0 to 30A switching current setup using a 650V, 50-mΩ GS65008T device, the difference between actual measurement and the simulated model is less than 5%, a very good number for Eon/Eoff accuracy. The outcome is a simulation tool that provides a convenient and accurate way to understand GaN switching characteristics, evaluate GaN switching performance under different electrical conditions and build overall confidence in a new product design. Visit the company’s download page to acquire the LTSpice models. Also visit the product page and app note page.