Employee
So, you completed your solar converter design using M1H 1200 V CoolSiC™ MOSFETs from Infineon. Or, at least, that’s what you think at this moment. To celebrate the achievement, you make a trip to the coffee machine, where you meet some colleagues, sipping their drinks while engaging in a conversation.
You join them at their table and tell them about your accomplishment. They look impressed, but one of them asks you if you considered the stability of the threshold voltage under stress over the entire lifetime.
Suddenly, you have that eerie feeling that creeps over every engineer when they feel like they have missed something. Seeing the horror appearing in your eyes, your teammates explain: “With the higher switching frequencies of SiC MOSFETs, you might see a dynamic shift of the threshold voltage VGS(th) towards the end of your application’s life. However, this depends on the total number of switching cycles.”
One of them pulls a napkin out from under their cup and scribbles an equation on it, while explaining the parameters:
“You get the number of switching cycles over your product’s lifetime, Ncycles, by multiplying the number of seconds a year has, with the expected lifetime in years, tlifetime, the percentage your product is operating during that time, ractive, and your switching frequency fsw. Now, what is the life expectancy of your application? The switching frequency? And the average percentage your device is active over this period?”
Doing the math, using 20 years as the application’s life, 48 kHz as switching frequency and 12 hours or 50% per day as operating time for your solar inverter, you come up with approximately 1.53 x 1013 cycles over the lifetime.
“By the way,” another engineer jumps in, “this equation is not only valid for solar inverters, but also for all other applications operating over a longer period, like a DC charger for electrical vehicles, energy storage systems, or servo drives.”
You wonder how this will influence your design. “It slightly increases the channel resistance, Rch. This contributes to your drain-source on-state resistance, RDS(on), and therefore the on-state losses. There are other contributors as well, namely the resistances of the junction field-effect transistor, RJFET, of the drift layer’s epitaxial layer, Repi, and of the highly doped SiC substrate, RSub.”
A second formula is written on the napkin.
“I recently read an application note from Infineon,” another engineer chimes in. “It explains this phenomenon and the test they developed. And it has graphs, where you can extract the threshold voltage drift based on your application’s operating parameters. Look it up!”
Evaluating the impact
You take your cup and leave for your desk, but not without taking the napkin with you. Back at your workstation, you pull up the application note and study it. You discover a lot of useful information and references to a second application note and a video explaining the alternating current – high temperature gate-bias stress test. In the application note, you look up the drift percentage for your turn-on voltage of 18 V and your switching number of 1.53 x 1013 cycles and identify it to be around 6% at a Tvjop of 25°C and about 3% at 175°C.
As you are running your application at well-defined gate bias levels, you learned that you can expect an even lower change in RDS(on) for your project. And you are happy to find out that you made a good decision in choosing Infineon’s M1H 1200 V CoolSiC™ MOSFETs already at the start of your design cycle. Its great threshold voltage stability diminishes the impact of the drift phenomena much further.
So, you are on the safe side and can enjoy your remaining coffee, which has gotten cold in the meantime.
Read the full analysis in our whitepaper here.