Hello,
I haven’t been able to identify a method to determine the value that is used for the Diode On resistance in the “Dual Active Bridge Converter” Demo model (check the attachment). Please, could you provide information on how to obtain this value ?
Thanks,
Hello,
Have you looked at the model initialization code included within? There we mention the primary and secondary side MOSFETs used in the simulation, both from Wolfspeed, as well as their nominal (data sheet) values for diode on-resistance. Does that help? And just in case you can’t actually find the model initialization commands, I attached the Note from the model description with these instructions below.
Best regards,
Kris
Thanks Kris.
I can see that the ron_body_diode parameter has been defined as 5 miliohms in both cases (C3M0030090K and C3M0065090D MOSFETs). However, according to figure 9 of the corresponding datasheets and my calculations, the slope resistance @ VGS = 0 (dynamic resistance @ VGS = 0) of the body diode is not equal to 5 miliohms but is approximately 33.5 miliohms for the C3M0030090K and 68.8 miliohms C3M0065090D. I use the criteria that the ron_body_diode parameter is the dynamic resistance of this diode, i.e. the slope of the curves of Figure 9 @ VGS = 0. Please, could you help me to clarify this?
Thanks,
Hello,
For C3M0065090D, since it is rated as 36A @ 25C then we measure Ron (the slope) at 36A from that figure. You might choose a value closer to 10 ohms, but I don’t think it is as high as the 30’s. Anyway, you are welcome to change the value in the model. Let me know if this helps.
Regards,
Kris
Hello, thanks.
I am using Ohm’s law to calculate the slope (1/Ron) from figures 9s of both datasheets C3M0030090K and C3M0065090D. Now I have changed to 33.5 miliohms and 68.8 miliohms the Ron body diode and the efficiency is not degraded as expected. I realized that the current is flowing through the MOSFET channel and there is no current flowing through the body diode. I expect that the MOSFET and the body diode share the drain current. Is that ok?
Thanks.
Here is the relevant plot with markup directly from the manufacturer. Around the rated current of 36A, the on resistance is probably closer to 10 ohms than 5:
But anyway, during reverse biasing, current sharing between the MOSFET channel and body diode is determined by precisely the on-resistances you are providing. It could just be that in your topology the switch is not turned on when the current is flowing in reverse, but rather it only freewheels through the body diode (and not the MOSFET). You should check with ammeters through the individual paths to be sure.
Thanks, Kris.
I noticed that the forward voltage of the body diode is set too high in the initialization parameters (vf_body_diode = 4.5 V). Hence, the body diode never gets forward polarized during all the simulation.
On the other hand, I can see that the slope of the curve that you uploaded is not Ron but 1/Ron. Therefore, if the current is 36 A, then 1/Ron = I / V is approx. 18.95 for the C3M0030090K (i.e Ron = 52.8 miliohms) and 24 (i.e 41.67 miliohms) for the C3M0065090D.
When I run the simulation, nothing changes because vf_body_diode = 1.8 V is still too high. However, if I use a dummy vf_body_diode of 0.7 V, it appears conduction losses in the body diodes of the primary side because they become forward polarized at the Load change (at t=0.4s).
Let me know your comments please.
I have attached the model with these changes made.
DualActiveBridgeBORRAR.plecs (190 KB)
Hello again,
I agree with your comment about the curve slope, the on-resistance is higher than what the model specifies. However, the Vfd of the diode of 4.5V is what the manufacturer specifies in the data sheet: https://www.wolfspeed.com/downloads/dl/file/id/1185/product/253/c3m0030090k.pdf. But you are right that this will force the diode to stay off. Anyway, does that answer your questions?
Best regards,
Kris
Hello Kris.
Yes, and thanks, you’ve helped make it clearer for me.
See you!
Regards,
Carlos
