Issues with TPS2811 MOSFET Driver for IRF530 in Buck SMPS Design

[pc9460]

I'm trying to build a smps but want to simulate it on multisim first before I buy all the components. The buck has a 6.3V dc input and will have a 5V dc output at 2 amps. I've designed a buck converter and it seems to work when I supply 12V dc with 400kHZ at 83% duty cycle into the gate of the mosfet. I'm trying to get my design with a gate driver but it doesn't seem to work. I put VCC at 12 and put 5V dc at the input (I'm planning on using a microcontroller). For some reason it's not working! I don't have to use a microcontroller but would need some feedback circuitry.

 

[D.A.(Tony)Stewart]

High side switch should be P type unless boost V is available... such as Cap diode boost from PWM on lower side of synchronous totem pole output to bias upper switch.

Where are your results? Note Vgs

 

[pc9460]

These are the results I'm getting. Is there a gate driver in multisim that will work for my buck converter ckt? I was planning on using the 5v dc output from a microcontroller to do the pwm for the circuit. It works at 12V and I was looking into using a gate driver to help accomplish all this. Thanks for the help.

 

[BradtheRad]

Does the mosfet turn on to the same degree (conduct the same amount of Amperes) whether you apply 5V or 12V at the gate?

It may help if you try a transistor for the switching device instead of the mosfet.

 

[pc9460]

Does the mosfet turn on to the same degree (conduct the same amount of Amperes) whether you apply 5V or 12V at the gate?

It may help if you try a transistor for the switching device instead of the mosfet.

If I apply 5v to the gate it doesn't conduct the same amount of current versus 12v which would give me my desired 2amps.

Is there a specific type of transistor that I should use to do my switching? Something preferably on multisim. Would it work if I used a microcontroller as the pwm? I.e is the 5v output from the microcontroller too much for the transistor? Is there any downside to using a transistor versus a mosfet? This is a power supply that I need to build and use for a project.

 

[KlausST]

Hi,

To see your signal on the scope you might connect scope input A- to Gnd.

Connect B+ to the output of the FET and B- to Gnd.

Actually you should look for a driver that is able to drive a high side N ch FET. It needs the bootstrap circuit.

Klaus

 

[FvM]

A look at TPS2811 datasheet reveals a reason why your circuit can't work. The device doesn't perform the intended level translation, it can't be driven by 5V logic when supplied with 12V. Choose a different gate driver circuit.

Secondly, driving a floating NMOS transistor by a ground referred gate driver isn't a good idea. It can work under restricted conditions, e.g. the input voltage must not be higher than 6.5V, the transistor doesn't have a threshold voltage in the upper range of type variation. Better use a logic level MOSFET (e.g. Vth = 2.5V) or a bootstrap gate driver.

 

[Easy peasy]

As above, there is no proper return path for the gate drive in the N type mosfet, g - s. A P type with drive suitably inverted would work better...

 

[pc9460]

If I swap out my mosfet with this one https://www.sparkfun.com/products/10213 would and drove it with a microcontroller would it work without having to use a gate driver?

 

[FvM]

A level shifting gate driver is required by topology of the buck converter. Also the previously suggested PMOSFET needs some kind of level conversion.

 

[pc9460]

Is there a way to replace the mosfet with a transistor instead and that would still be called a smps, correct? I would like to possibly still control it with a microntroller unless something else is easiest. Is there a downside to using a transistor versus a mosfet? Will it still output the consistent 5v dc that I need?

 

[KlausST]

Hi,

A bjt could be faster or slower, select the right one.
A MOSFET usually has lower voltage drop and therefore lower power loss.
A MOSFET need low DC gate current, but for fast switching it needs a lot of peak current.

You will find that even with a bjt you need a base drive circuit. So it is almost the same effort than just choosing a gate driver IC for a MOSFET. I'd prefer the MOSFET.

The most easiest and reliable solution is to go to a manufacturer homepage (ti, linear, st, national...) and select a 8 pin step down switch mode IC with built in FET.

Klaus

 

[pc9460]

Is there any ic's available that will do feedback and adjust the pwm duty cycle so the output voltage stays at 5v dc? I will need to do some sort of feedback system.

 

[KlausST]

Hi,

Is there any ic's available that will do feedback and adjust the pwm duty cycle so the output voltage stays at 5v dc?

--> Thousands i think. At least hundreds.

I have given you manufacturer names to look at. Did you do this?

Here an example.
Got to linear.com (Other manufacturers are compareable)
Select the device:


then fill your data into the page:


and find 66 parts that meets your specification.

Klaus

 

[FvM]

Provided as is, an example of a buck converter based on classical 3524 switch-mode controller.

 

[KlausST]

Hi,

if you need low part count, then a LT1374-5 needs 8 parts and delivers up to 4.5A.

For mass production it may be too expensive. For a single project it may be ok.

There are a lot of switchers. National´s simple switchers for example.

Klaus

 

[spork891]

A look at TPS2811 datasheet reveals a reason why your circuit can't work. The device doesn't perform the intended level translation, it can't be driven by 5V logic when supplied with 12V. Choose a different gate driver circuit.

Secondly, driving a floating NMOS transistor by a ground referred gate driver isn't a good idea. It can work under restricted conditions, e.g. the input voltage must not be higher than 6.5V, the transistor doesn't have a threshold voltage in the upper range of type variation. Better use a logic level MOSFET (e.g. Vth = 2.5V) or a bootstrap gate driver.

This is something that has befuddled me for a while. By what term is a simple high side NMOS FET driver known? Whenever I go searching for them, I find devices that are heavily integrated (IE they do UVLO, OVLO, Current limiting) and are thus more expensive and complex than what I want. What do you look for when you are looking for a simple NMOS gate driver for a situation like above, assuming I wanted to go the the immense trouble of designing the feedback loop with an MCU.

 

[FvM]

The drivers are called high-side gate drivers. Single channel drivers without low-side driver are avialable e.g. from IRF or Fairchildsemi. You can also use standard half-bridge drivers like IR2110 with unconnected low side channel.

Due to the bootstrap operation principle, there are some operation restrictions regarding duty cycle and start-up to observe.

 

[pc9460]

Would a pwm controller ic and a gate driver ic be what I would need to do the pwm for my smps? Would the pwm controller adjust so that my output is a constant 5v dc?

 

[KlausST]

Hi,

We give informations on partnumbers. Look at at least one of the datasheets found via the linear selection guide.

Please read it - or at least look at the pictures - then you will find out, that there is no need for external drivers nor external FETs. All inside.

As i wrote: 8 parts in total.
How can I say it more clearly? No need to generate a PWM. PWM controller with regulation loop inside. Constant 5V output voltage.

Klaus