Driving LED with an N-channel MOSFET

Question

I am driving a red LED with a forward voltage of approx. 2 V as shown in the picture. The driving signal comes from a microcontroller with a logic HIGH of 1.8 V. The LED is being sourced from an independent supply of 3.3 V. The circuit works, but there is something in my measurements that I can't explain.

I measured 3 mA through the LED branch. I need more current through the LED (about 10 mA would be ideal). Then I reduced R1 in steps until I had about 10 Ω, yet the current in the LED did not change much. Sometimes I measured about 8 mA at the beginning of the measurement which after some seconds ended up at 4 mA.

• Why is this happening? I also reduced R2 but it seems to have no effect either.

• How could I increase the LED using a control signal of 1.8 V and powering the LED with 3.3 V?

^{simulate this circuit – Schematic created using CircuitLab}

Answer 1

The problem is most likely that the mosfet does not have enough voltage to fully turn on. Vgs is probably not high enough. I couldn't find the datasheet for the CMPD7002 (or any part named that). First make sure the Vgs will give you enough current, find how much Vgs you need for the Rds on. You'll need to do one or more of these things:

1) If the GPIO's on the microcontroller are 3.3V tolerant (or if 3.3V is fine with a pull down) Then use open drain on the GPIO and a pull up to 3.3V

2) Use a circuit or buffer to translate 1.8 to 3.3V

3) Select a different mosfet that turns on at 1.8V and has a low rdson at 1.8V.

4) Calculate the current vs LED drop and size R1 correctly.

Answer 2

Use a good ol' BJT, or find a logic-level FET rated for 1.8V. I've called out a 2N3904 in the suggested schematic below, but (A) you can use just about any small-signal NPN there, and (B) there are prebiased transistors that'll work and save you a resistor.

Design for an 0.7V drop from base to emitter, and a base current of about 1/10 of your desired collector current, to keep the transistor hard into saturation.

Answer 3

You need a MOSFET that turns on fully with 1.8V. Not many will. Do you have a datasheet for CMPD7002AG? I'm not getting any hits on it.

This one looks like it will turn fully with just 0.9V on the gate

https://www.onsemi.com/pub/Collateral/MCH3484-D.PDF

Digikey has them.

Answer 4

Although you have not linked any part numbers, I suspect you can drive the 2V LED Cathode directly from Vol with Anode to 3.3. The Ron= Vol/Iol should be in the 50 Ohm ballpark. Then when High , the LED will see 3.3-1.8 = 1.5V which will be very dim.

All you need is 10mA so when ON, the driver will rise to 3.3-2V= 1.1V with a current of 1.1V/50R or 22mA so adding 50 R in series will reduce the current to ~ 10mA and Vol to 0.6V .

For protection (opt.) in case of 3.3 rising before 1.8, just add a Schottky diode from output port to Vdd.

Then you need to invert the LED state in S/W.

Proof with Sim and specs

From Cypress datasheet p76
Iol = 4 mA at Vdd=1.8V Vol = 0.6 max, thus driver Ron= Vol/Iol=0.6V/4mA = 150 Ohms

I used a custom model of Vf=2.1V @ 20mA ( Isat=92pA, ESR 12 Ohms, Emission coefficent =3.73, breakdown voltage= 2V)

The above simulates minimum current from an ARM Core data port using 1.8V supply.

Adding 150 OHms will reduce brightness in half to 5mA min.