ways to discharge a mosfet faster

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi all

I was doing a test on MOSFET. the circuit that i used is a resistive switching test circuit. So when a voltage pulse of 15 V with 95% dutu cycle is applied to the gate. The drain to source would be 0 for 95% duty cycle. But the problem here is when the gate pulse goes to 0V after the 15V pulse, the mosfet is not discharging properly and the drain to source voltage is merely 0V. what are the possible ways to make the mosfet discharge faster.

Thank You
 
Dear jegandren,
For proper function of Mosfets, parasitic capacitance created in the gate has to be discharged quickly and for that appropriate gate driver with appropriate configuration is required. I have put forward 3 configurations for easy understanding. For normal and medium power, fig. 1 and fig. 2 and for higher power, fig. 3 can be adopted.
Hope, it will be helpful for understanding Mosfet Driving mechanism.


_____________________________________________________________________________
IDEAS ARE MONEY IF THEY CAN BE USED ON GROUND.
বাস্তবে ব্যবহৃত চিন্তাধারাই মানুষের প্রকৃত সম্পদ।
 

Attachments

  • mosfet _ igbt driver configurations.pdf
    26 KB · Views: 639
Discharging the MOSFET

I could understand that the MOSFET you are operating in 95% duty cycle is not discharging fast enough.

Normally i use a reverse diode from the Gate to Pulse generator.

One more option will be using a discharging transistor.

I have attached 2 sample schematics. for both the solutions.


This component need to be selected depending on the Frequency you operate.
 

Attachments

  • untitled1.jpg
    untitled1.jpg
    13 KB · Views: 814
Hi jtkraj,

Your second circuit is alright but first circuit requires a diode as shown in s.g. mahbub's first one - fig.1. If diode is not provided after 47 ohm resistor, then when the gate charge will be grounded, then the input gate driving pulse will be also grounded which will distort the gate driving pulse.
For all sort of circuits, high or low frequency, totem pole configuration is the best.
Thanks.
 
Another Option is to replace the gate resistor with a small inductor. The inductor will block RF oscillation while providing maximum current flow.

Couple of comments:

How much is the voltage drop across the BJT transistor your using. If your supply voltage is 12V and the VCEsat voltage is 2 volts your Gate voltage is reduced to 10 volts.

How fast are you switching and what is the response time of the opti-coupler your using. If the opti-coupler switching speed is below or near your Mosfet switching speed you have a problem.


FWIW: You should consider using Gate drivers instead of using BJT or Opti-Couplers to drive your Mosfets or IGBT. Modern Gate drivers uses a Mosfet totem pole instead of a BJT totem pole which avoids the VCEsat issue of BJT's. A Modern gate driver will almost certainly deliver better switching performance than the circuits your looking at. Why try to re-invent the wheel?
 
Hi,
Techguy, you made a mistake. TLP250 as shown in option 3 in s.g. mahbub's attachment, is not just an optocoupler. It is an optically isolated dedicated MOSFET / IGBT driver made by Toshiba and has current capability of upto 1.5A and frequency upto 25kHz.

Yes, modern drivers use MOSFET totem-pole stages instead of BJTs, but there is a problem. To drive the driver MOSFETs in the totem-pole stage, you need another gate driver, but for BJT, it is not required, so for novices, BJT is preferable for learning purpose, so that afterwards they can use dedicated MOSFET drivers.

To omit the VCEsat voltage problem, you can place the pnp on top and the npn below. It will act as an inverter totem-pole stage but will still drive the MOSFET at 12v (that is, if the supply voltage is 12v).
This stage practically has no frequency problem, as BJTs can operate at VERY HIGH frequencies. e.g. I use the BC327 and BC337 which can operate upto 10MHz.

But, for learning purpose for novices, a BJT totem-pole stage is the best choice.




_____________________________________________________________________________

The strongest man in the world is one, who stands alone for achieving something good.
পৃথিবীতে সেই সবচেয়ে শক্তিশালী ব্যক্তি, যে ভাল কিছু অর্জনের জন্য একাই নিরন্তর চেষ্টা চালিয়ে যায়।
 
Tahmid said:
Hi,
Techguy, you made a mistake. TLP250 as shown in option 3 in s.g. mahbub's attachment, is not just an optocoupler. It is an optically isolated dedicated MOSFET / IGBT driver made by Toshiba and has current capability of upto 1.5A and frequency upto 25kHz.

Yes, modern drivers use MOSFET totem-pole stages instead of BJTs, but there is a problem. To drive the driver MOSFETs in the totem-pole stage, you need another gate driver, but for BJT, it is not required, so for novices, BJT is preferable for learning purpose, so that afterwards they can use dedicated MOSFET drivers.

To omit the VCEsat voltage problem, you can place the pnp on top and the npn below. It will act as an inverter totem-pole stage but will still drive the MOSFET at 12v (that is, if the supply voltage is 12v).
This stage practically has no frequency problem, as BJTs can operate at VERY HIGH frequencies. e.g. I use the BC327 and BC337 which can operate upto 10MHz.

But, for learning purpose for novices, a BJT totem-pole stage is the best choice.
I realized that the TLP250 was a gate driver. A BJT totem pole will still have a voltage drop across the transistors. The purpose of using a totem pole is to elminate resistance and to permit faster turn ons but it doesn't eliminate Vce(sat) as well as reduce switching losses created when using a series resistor. Take the TLP250 or a set of NPN and PNP transistors and measure the gate voltage when its turned on. If your supply voltage is 12V it will probably be between 11.7 V and 11.3 V (lower if the NPN has a really lousy Vce(sat)) You will have a voltage drop caused by the Vce(sat) of the NPN transistor. When you reset the BJT totem pole, the gate voltage will fall to the Vce(sat) of the PNP transistor. Although this shouldn't be a big deal unless the Vce(sat) is so large that the MOSFET never turns completely off.

A modern Gate Driver such as the IR2110 use built-in Mosfet totem poles instead of builtin BJT totem poles. See Page 4 on http://www.irf.com/product-info/datasheets/data/ir2110.pdf

The TLP250 looks like an outdated design. 25Khz is a pretty low switching speed. Put the Tr and Tf are very slow (measured in microseconds). Tr and Tf for the IR2110 is measured in nanoseconds. The slower the Tr and Tf is, the more switching losses are created. When a Mosfet gate voltage changes the it switches out of full conduction or from non-conduction into a semi-conduction state causing resistance. This resistance causes energy to disapate into heat. Ideally you want fast gate voltage rises and falls when switching.
 
Hi Tech Guy,
Thank you for thread bare analysis of TLP 250 and in depth analysis regarding modalities of modern gate drivers. You are absolutely correct in your perspective, but here the scenario is different. Here, a new entrant in the world of power electronics wants to know the all pros and cons of gate drive techniques, which requires lot of theoretical as well as costly, lethal and time consume practical knowledge. So, question to you,

What do you prefer for a new entrant, try to learn through basic configurations within the frequency of < 25KHZ and when perfection is achieved, then going to greater KHZ freq. using the ICs like IR 2110/2181/2185 etc. or straight way landing in higher frequency range with the drivers you mentioned, when even basic modalities are still unknown to him.
With thanks.

____________________________________________
Ideas are money if they can be used on ground.
বাস্তবে ব্যবহারযোগ্য চিন্তাধারাই মানুষের প্রকৃত সম্পদ।
 
Hi,
Jegandren, you should move slowly but steadily in the vast field of Power Electronics. Don't try to move fast. If you do, you will find your lot of energy, effort, money, time and material is wasted. Make your foundation stronger, and then you can move fast. In dynamics, there is a formula, time of ascent = time of descent. It is applicable in the Power Electronics field. Keep it in mind. Do not rush and land up in S.M.P.S. minefield.
Thanks.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.