Hello everybody! I am new here.

I have a little grasp on theory, but I know *nothing* about practical electronics, that's why I came here.

In my project, a reed(magnetic) switch controls the powering of 9x3.6V 300mA LEDs in parallel.

(Note: I have read that it's not a good idea to put leds in parallel, and without resistors, as I did. However, the resistance needed is probably less than this of whatever cable I am using. Even more importantly, I have this already set up, and at a high price, so please don't waste advice on this).

This is a total of 2.7A, certainly off range of the switch. Also, I hope to find and use some PWM circuit to be able to dim the LEDs. Whatever chip does this, obviously also cannot handle 3A.

Also, I understand many chips need 5V. I use a 18650 battery, nominal voltage 3.7V, true maximum 4.2: I also got a small dc step-up converter to 5V.

Can you please propose a solution including a specific transistor that can do this job? If this presents you too much freedom, consider that my project is tight about space.

Thanks beforehand!

 

If you're reluctant to use current-limiting resistors how will you prevent a 4.2V battery or 5V converter from frying your 3.6V (nominal) LEDs?

I have this already set up, and at a high price, so please don't waste advice on this

It's your prerogative not to follow advice already given; but your 'high price' is likely to get even higher if/when you have to replace the LDs

 

I am just anxious to prevent wasting your time.

4.2V(max)-3.6V led voltage drop= 0.6V

target current of 0.3A
Resistance of resistor needed: 0.6/0.3=2 Ohms

The wire I use has more than that.

 

I would recommend using a MOSFET instead of a transistor, since it is being used as a switch. You need a logic level MOSFET, meaning the gate threshold voltage is fairly low. I have used a NTD4815N in a similar application, driving parallel LEDs at 5V with a PWM signal. The gate threshold voltage on this part is 2.5V.

If I couldn't use current limiting resistors, it is possible to use the PWM as a current limiter by using a high frequency, and assuring the duty cycle never gets to 100%.

 

What control signal is available to control the MOSFET? Is the reed switch still part of the plan, or is it the PWM, or both?

I agree a MOSFET is the good choice here, although a power transistor might be better if the control voltage is low, and can supply enough current to the base. Just depends on the rest of the circuit.

 

The reed switch controls everything, and the PWM is the dimmer.
So, the switch gives power to the PWM, the PWM creates the control signal for the MOSFET, and the MOSFET amplifies the PWMed control signal to the needed 3A.

You mentioned NTD4815N? May I have a schematic too? Where to connect what? I am really bad at this...

 

...
In my project, a reed(magnetic) switch controls the powering of 9x3.6V 300mA LEDs in parallel.
...

1W white power LEDs like those have significant internal resistance and a soft Vf knee, and can be safely paralleled without resistors (assuming same brand and type of LEDs). Most manufacturers do it now.

Of course you need some resistance or current limiting in total.

 

Where to connect what? I am really bad at this...

What are you using to provide the PWM signal? And is the entire system powered by the same battery?

Basically the N-channel MOSFET is like a switch on the main power line for the LEDs. It controls the return path to ground, though, not the V+ line. So the PWM goes to the gate pin, the source pin is connected to battery ground, and the V- wire from the LEDs goes to the drain pin.

As you said, you can power your PWM thru the reed switch. Since the MOSFET and the LEDs will be turned on when the PWM signal is removed, you need to add a resistor from gate to ground. This will ensure the MOSFET and LEDs are turned off when the PWM signal is removed. The gate voltage of the MOSFET might become undefined otherwise.