Saturday, December 18, 2010
【 Weak current College 】 on IGBT Driver understanding 】
I just combined with practical application on a number of questions about my understanding, the wrong place, please advise.
IGBT devices how to open and close, speed is fast is slow, in fact, until now, there is no clear conclusions, ultimately practical application environment.
1. first opened. Everybody knows that, in theory, open faster and sooner into saturation, the smaller the switching losses. Receive the actual conditions, opening speed cannot be done too quickly.
The effects of stray inductance gate
The inductance is widespread, especially the use of magnetic coupling of driver circuit is particularly pronounced. It produces resonance and gate capacitance, increase Rg, you can mitigate this phenomenon. Neither the Lgs, or Rg, affects the opening speed, increasing the switching loss. Figure 1.
The actual effect of the circuit
Use reasonable circuit and reduce the impact, Lgs Rg selection is very small, whether it solved the problem? no, as in Figure 2, applied to the double forward topology.
As the transformer leakage Lts, magnetic Reset PIN inductance Lds, excessively high di/dt, will result in excessive Vce (and Cce oscillation). These inductors are difficult to predict, in order to improve performance, both from the wiring is decreasing. Energy reduction makes the Vce receive limit. If you cannot resolve this problem, the need for increased absorption. Absorption in Figure 3. Additional circuit need close to the root, otherwise harmful IGBT useless. Absorption of increased wear and tear, it can offset as the opening speed of faster, need actual validation.
(Single) double forward is a hard switch topology, opened the fastest one, corresponding to the other topologies, such as the bridge, the situation worse. For soft-switching, in particular, to increase the voltage soft switch opened speed is a good choice.
2. repeat off. Closing speed on IGBT close loss.
Some say that quickly turn off the high dv/dt, destruction of IGBT. Then see what conditions, only very high dv/dt, IGBT switch speed to keep up with time. Small just negative feedback effect, increase the flow of electric current on the driver. Due to the IGBT, bus, often with limited absorption speed. Difficult to produce high dv/dt. Really so high DV/dt, as long as the driver circuit output impedance low enough, the gate will not be destroyed.
Then there are more than ten years, I have never seen due excessively high dv/dt IGBT failure phenomenon caused, only theoretically established assumptions not to consider.
Real reason: because it is less child devices IGBT, he cannot close the gate to reach quickly close the IGBT purposes. Therefore quickly close on it useless.
3. say driven negative bias
As long as the control system good bus voltage IGBT transient overshoot, negative bias is not required. Selecting the right power topology, such as zero flow resonant converter, you can minimize the impact on the bus.
And MOSFET, negative bias can prevent bus bar too high dv/dt caused gate misleading. I said I am ten years never seen due excessively high dv/dt IGBT failure caused mainly refers to the opposite extreme.
Gate drive are often low impedance loads, especially the source drive, low-impedance, anti-interference ability is strong. Negative bias select-5V to, of course, select-15V fine. Higher negative bias increases driving loss.
4. on the over-current protection
1. high frequency converter (not directly access load)
In designing the circuits, the choice of the IGBT, almost always in the current left margin on the protection of the point are often lower than the rated value of IGBT. The so-called 3-protection only in more than 2 times the rated current function better. In 2 times the rated current protection is to close within the best. Not only reflect the speed, and can be individually pulsed current limiting. True if it's protected, not only delay a long time, but due to the close loss, continuous work to soon not bad is weird.
Available in almost all the IGBT drive preach the so-called 3-protection, but in practice, unless you are using a load switch, this feature basically useless. Even as load switch, if there are no restrictions on certain conditions, the protection function is useless.
Therefore, unless you have a flawed design, 3-protection of useless.
2. load switch
IGBT as load switches used, overcurrent protection is far more complex. Output current depends on the IGBT gate drive voltage. 3-protection only in more than 2 times the rated current function better. Gate drive voltage is too high, making short-circuit current is too large, because of latch, gate shut down has no effect. Therefore driving voltage too high, not excessive pursuit breakover voltage drop. Makes 3-overcurrent protection does not work. General latch phenomena in the output current is greater than 5 times the rated current is likely to occur, therefore if you want reliable 3-overcurrent protection, current protection point best smaller than this point.
But determining the gate voltage and short-circuit current relationship is almost impossible. As load switch, limit current up rate is a very good measure, it also limits the short time. The short-circuit protection can have a variety of options, in addition to 3-protection (essentially ease off), directly to the fast shutdown is also feasible.
In current of less than 2 times the rated current, direct and fast close method is more efficient. Suppose 400V, 100A load switch, select 300A IGBT as load switch, 1uH inductance as current up rate limit. Then:
Set point to protection, 100A current is less than 2 times the rated current, then:
Allows closing the maximum delay (2 * 300-100)/400 = 1.25us. This condition is not difficult to meet.
Namely: to meet the fast shutdown conditions, select IGBT capacity is large enough, so to better protectNurse-delay, smaller short-circuit suppression inductance. In other words, if the protection delay design is very small, can also reduce short restraining inductance, reduce IGBT capacity. In order to improve reliability. Or the previous example, if the protection delay reach 200ns then:
Actual current close point: 100 + 400 * 0.2 = 180A. If you do not take into account conduction losses, 100A of IGBT can meet use requirement.
Direct fast shutdown without substantial application of reason is mainly due to the fact that drive response quickly enough. Taking into account the reaction speed and driving ability to drive the market. JD10 series can be shut down within a 100ns 300A IGBT and close feet reflect only 50ns, even in isolated driven fore close, only 100ns delay, is seen in the market.
Annex: JD10 series MOS, the main performance IGBT drive:
Single-hose high power MOS or IGBT module drives.
Fast driving time delay: delay of typical open, close the delay time of ≤ 200ns = 100ns.
With up to 20A of transient driving ability
By falling edge up 50ns
Work dutyfactor 0-100%.
No input signal power.
Built-in buffer structure, not to drive on the rise.
Static power consumption low.
Output power and signal power independence.
With the broken legs, can rapidly switching 50ns.