When to use a Mosfet vs BJT (high current (2a))

Goal

I’m wanting to “dim” LEDs using PWM where varying the on/off frequency alters the brightness.
I’ve done this before with single LEDs but now I’m wanting to control lots of leds including things like this and this

BJT

I’ve been experimenting with just a single green waterclear LED.
I’ve got this circuit working with the S9013.

My understanding is that S9013 is high current but maybe maximum 500ma.
Those cobs worry me, at 600ma EACH and I don’t think this transistor will scale.

question

Is there a transistor that is designed to handle high current fast switching scenarios?

MOSFETS

I know MOSFET are good in these situations, and I’ve done some reading on them.
I own a IRLB8721 that seems right for the task.

question

I’m writing because I know there are lots of rules for transistors.
e.g. 0.7v between the base and the emitter
e.g. the collector must be higher than the emitter

I imagine there are also lots of rules for mosfets too, but googling around there is nothing jumping out at me. I’ve tried searches like Rules for MOSFETS, MOSFETS conditions.
I think I could wire it like this… it seem too easy

question

Why would I chose a MOSFET solution over a high current BJT solution?

Thanks in advance :slight_smile:

2 Likes

Hi Jonny,

What timing, I just finished a quick dive into them at uni.

For power control applications FET’s are generally better, with this application a N-channel will be best.

On MOSFETs
From my time making mistakes and Bob fixing them I’ve made the following rules:

  • Gate resistor - the gate of a FET has a non-negligible capacitance, this can be solved by including an inline resistor (I go for around 100 ohms)
  • Make sure to discharge that - a pull-down resistor can be used to ensure the FET capacitor is discharged
  • Make sure to read the datasheet - Vgs vs Ron
  • Remember, MOSFETs are voltage devices

Its also worth getting rid of the potentiometer/rheostat there, you’ve got PWM now!

2 Likes

Hi Pix.
Firstly get rid of RV1.
Size R1 for maximum allowable current.
Insert a resistor of 1kΩ in series with the gate.
Fit 10kΩ between gate and ground.

High current BJT solutions are usually IGBT (Insulated Gate Bi-polar Transistor). There are pros and cons with both devices. You will have to look it up and decide.
Cheers Bob

3 Likes

Thanks Jim, edited for consistency
EDIT: Sorry Jim! Was just rushed home and was on my phone :grimacing:

Hey Pix, Great Scott has a good video on them from memory.

1 Like

Thanks for your thoughts bob

What is this resistor doing? :slight_smile:

1 Like

It protects the gate and the GPIO of the Arduino reducing the inrush current.
Mosfets have a significant gate capacitance.
The 10K to GND allows the capacitance to discharge and keeps the Mosfet off when the Arduino is not driving it.

@Robert93820 might have a better description, but that is my understanding and what works from past experience.

Cheers
Jim

PS In a previous thread we discussed the gate capacitance. When charged it is enough to keep the Mosfet on for days. I did a test. Mosfet switching a LED. Triggered it on by making the gate positive then removed the trigger. Nothing connected to the gate. After the LED had been on for about 3 days I discontinued the test.

4 Likes

Hi James.

That is it in a nutshell although most of the protection is for the GPIO.
I do think that this has been explained to Pix in the past. If I am wrong I apologise but the subject seems to come up every time someone uses a Mosfet.

This is not helped when some manufacturers omit these components (especially the series gate resistor) when designing some “general purpose” modules.
Cheers Bob

Ps i remember that switching experiment I suggested you do. The LED could stay on for months. The Mosfet will even stay ON when you remove power. When power is reapplied the Mosfet will still be ON.

4 Likes

Hi @James46717 and @Robert93820.

Very nice explanation. Thanks James. :slight_smile:

The only time I’ve worked with fets in the past is as a logic level inverter.
This is the first time I’m working with FETs and PWM. We have discussed similar situations with transistors, maybe that’s what you’re remembering :man_shrugging:.


This is surprising. Is the oxide acting as a seperator between the metal layer and… what? Doesn’t it go straight from the oxide into the doped silicon?

1 Like

This might help you understand it.
I am not the person to explain it in detail. The symbol used on the data sheet, to me, shows a type of capacitor. The data sheet lists the input and output capacitance.
image

The oxide layer insulates the metal gate from the source and drain and this results in capacitance.

Example from datasheet.
image
image

Regards
Jim

2 Likes

From the link that james provided.

A conducting layer of aluminum is laid over the entire channel

Maybe this is the other metal plate completing the capacitor? :man_shrugging:
Electronics are so interesting. :slight_smile:

1 Like

Hi Pix
Th save a lot of grey hairs just believe the Data sheet.
Cheers Bob
The capacitor in question here is the gate to source cap.

1 Like

For future readers, I found this lovely summary within the Wikipedia page for IGBTs

In general, high voltage, high current and lower frequencies favor the IGBT while low voltage, medium current and high switching frequencies are the domain of the MOSFET.

To answer my original questions about which is better, it seems like rewiring to the MOSFET was the right choice for my PWM application.

3 Likes

On a practical note, I’ve just used this MOSFET module to dim 5 3W LED’s powered by this step-down module.

1 Like