Author Topic: MOSFET as switch  (Read 4138 times)

0 Members and 1 Guest are viewing this topic.

Offline jfgilbertTopic starter

  • Newbie
  • Posts: 7
  • Country: us
MOSFET as switch
« on: December 01, 2015, 05:48:10 am »
Hello, this is probably a dumb newbie question, but I can't figure it out.
I tried to use an N-channel MOSFET as a relatively fast switch, and I did not understand what was going on so I made a simple test circuit, as shown in the diagram. I am feeding the control signal to the gate with a function generator. The MOSFET is an IRF510.
At fairly low frequencies, the signal is as expected, as seen on the first scope picture. The signal is a square wave at 100Hz with 10V amplitude, as the data sheet for the transistor says 10V Vgs puts it in the sweet spot.
When I increase the frequency to 500KHz, things get really ugly as in the second scope picture.
The IRF510 data sheet gives rise and fall times in the few nanoseconds, so 500KHz should not be asking too much.
The 5V load power is provided by an older but solid HP power supply that is well below its limits.
The signal generator has a 25MHz sine bandwidth and 10MHz square wave capability, so, again it should be well within its limits. I understand that the transistor has some parasitic capacitance that may introduce some oscillation, but they are in the low pF range. I tried different values for the gate pull down resistor with little effect.
What am missing and why am I not getting a clean square wave?
Thank you for suggestions.
 

Offline Richard Head

  • Frequent Contributor
  • **
  • Posts: 685
  • Country: 00
Re: MOSFET as switch
« Reply #1 on: December 01, 2015, 06:14:08 am »
The MOSFET has an input capacitance that you have to take into account. The series 1k5 resistor is limiting the rate at which the gate capacitance can charge and discharge. Also, there is the Miller capacitance which increases the capacitance hugely when Vg gets to the threshold voltage (+/- 3V).
The solution is to decrease the 1k5 resistor value drastically. The faster you need to charge/discharge the gate the higher the drive current required. That's why there are dedicated MOSFET driver chips.
 

Online Monkeh

  • Super Contributor
  • ***
  • Posts: 7990
  • Country: gb
Re: MOSFET as switch
« Reply #2 on: December 01, 2015, 06:18:28 am »
The MOSFET's gate is a capacitor. You are slowing down the charging of said capacitor with your 1k5 gate resistor.

If you want to hit the rise and fall times in the datasheet, you will need to match the conditions in the datasheet. These include a gate resistor of 24 ohms, and involve driving a fairly substantial current into, and sinking from, the gate. Your signal generator could very easily be unable to source and sink such a current even if you reduce the gate resistance.
 

Offline rx8pilot

  • Super Contributor
  • ***
  • Posts: 3634
  • Country: us
  • If you want more money, be more valuable.
Re: MOSFET as switch
« Reply #3 on: December 01, 2015, 06:59:46 am »
have a look at 'gate drivers'. They charge and discharge the gate of a MOSFET, some of them faster than others. The data sheet will tell you the fastest it can be driven.

As the others said, your gate resistor only serves to slow things down.
Factory400 - the worlds smallest factory. https://www.youtube.com/c/Factory400
 

Offline rs20

  • Super Contributor
  • ***
  • Posts: 2317
  • Country: au
Re: MOSFET as switch
« Reply #4 on: December 01, 2015, 07:27:12 am »
It's sort of obvious in this particular case, but in future you should describe which curve is which on the oscilloscope traces (i.e. is Yellow the input, or the output?). Nothing is more frustrating for us than vague problem descriptions!
 

Offline Gyro

  • Super Contributor
  • ***
  • Posts: 9410
  • Country: gb
Re: MOSFET as switch
« Reply #5 on: December 01, 2015, 09:43:08 am »
Quote
I understand that the transistor has some parasitic capacitance that may introduce some oscillation, but they are in the low pF range.

^^ This is your fundamental error. As others have said, capacitance is the reason - time to go and read the datasheet gate capacitance spec and actually do some RC calculations for the required rise/fall time.
« Last Edit: December 01, 2015, 09:44:43 am by Gyro »
Best Regards, Chris
 

Offline BloodyCactus

  • Frequent Contributor
  • **
  • Posts: 482
  • Country: us
    • Kråketær
Re: MOSFET as switch
« Reply #6 on: December 01, 2015, 01:57:45 pm »
also I think your problem with driving it from your signal generator is, you chose IRF chip, the 'F' needs a driver chip, the IRL510 is logic level, the IRL might be a better choice if your just driving it with low voltage from the signal generator.
-- Aussie living in the USA --
 

Offline jfgilbertTopic starter

  • Newbie
  • Posts: 7
  • Country: us
Re: MOSFET as switch
« Reply #7 on: December 01, 2015, 02:04:09 pm »
Got it. Thanks for all the responses and advice.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf