Tool/software:
Hi
Customer want to use UCC23513 to drive IKQ100N120CH7(Ciss=13nF, Qg=700nC),20Khz switching frequency.
Could UCC23513 drive this IGBT? Should we add external push-pull driving circuit?
Besides, I want to know how to determine if a gate driver have a enough capability to drive a MOSFET? How to calculate? Could you please share the detailed steps?
Thank you,
Yishan Chen
Hi Yishan Chen,
It all comes down to switching losses. Gate driver selection can help minimize a switch's loss, but the switch sets a lower bound.
I have compared the switching losses of P2P gate drivers when driving IKQ75N120CS7, a comparable IGBT, with 5 ohms of external Rg. Here are my results up to 50A:
Eon Comparison:
Eoff Comparison:
You will have to estimate your average current and see if the heatsink can handle the corresponding thermal power. Extrapolating my 50A data by 2x to estimate losses for 100A worst case, I estimate that you will get around 14mJ of switching losses for Eon and 6mJ for Eoff. at 20kHz, that is 400W of power dissipation. Can your heatsink keep all the IGBTs below 175C with that thermal dissipation? To stay under 100W, you would need to limit the switching frequency to about 100W/ 20mJ = 5kHz. The IGBT is rated for the power, but you will have to take care of the temperature. This does not even include conduction losses yet. Maybe you do not switch 100% of the time too, such as in a motor inverter, where a switch might only have a 50% activity factor.
You will get 5x less losses with a SiC FET. The C3M0015065D will have an Eon of about 2mJ and Eoff of about 0.3mJ. That is 20k*2.3mJ= 46W for the same switching current. However, the voltage was only 400V, not 800V as in the previous test.
Eon:
Eoff:
Best regards,
Sean
Hi Yishan
1. I obtain Eon and Eoff by multiplying Id and Vds during a double pulse test (sample oscilloscope measurement below)
2. You have to measure it to get the resolution to compare gate drivers. Simulations are not very accurate. But you can get a good initial estimate from the IGBT datasheet, if the test condition is close to your use case.
3. Yes, it directly impacts efficiency. Your conduction losses will have to be added to the switching losses, but they will be relatively smaller. The IGBT datasheet has estimates for this as well.
4. 5 ohms was the lowest I tested for this IGBT. I damaged the devices when I tried to use 1 ohm in my circuit. I do data for higher Rg, and you can extrapolate:
5. With UCC23525 and 40A & 400V, your Eon will be ~5mJ, and Eoff is ~2mJ. For 20kHz continuous, that is 140W. In a three phase system, the activity factor is 1/2 (It does not PWM 100% of the time), so that would only actually only burn 70W per IGBT. The conduction losses will add an additional 1.5Vce * 40A * 1/2 =30W.
Best regards,
Sean
