5v pwm to 24v dimming

Hello,

I'm trying to dim a led-strip that is running on 24v. When I hook it up to a power-supply it's nicely dimming from 24v to 18v. Turns off at +/- 18v.

Now I' trying to regulate the brightness with an esp32 (arduino like). I have created the following circuit:

But the problem is that when the PWM is 0, the led-strip doesn't turn off completely. It's 'stuck' at 19v.

I'm probably using the wrong FET. Could any of you guide me? I'm kinda hopeless at the moment.

Thanks

When I hook it up to a power-supply it's nicely dimming from 24v to 18v. Turns off at +/- 18v.

What LED strip do you have? If it's shutting off at 18V It probably has some driver circuitry and it's probably not designed to be dimmed.

But the problem is that when the PWM is 0, the led-strip doesn't turn off completely. It's 'stuck' at 19v.

How are you measuring that? A DMM usually is not reliable with PWM (but it will be read correctly at zero and 100% PWM where there is no actual PWM). Are you measuring across the LED strip, or are you measuring to ground.

Some experiments you can try:

Disconnect the MOSFET from the 74HCT245 and ground the Gate. (The LED strip should turn off). Connect the MOSFET Gate to 5V. (The LED strip should come-on.) If that works, re-connect the 74HTC245 and try writing high & low (no PWM).

Connect a regular LED & 1K series resistor to the output of the 74HCT245 and confirm it dims with PWM. Then connect the regular LED & resistor in place of the LED strip to see if it dims. (Or try it with the LED strip still connected.)

Hi,

Im using a led-strip of the following series: 404 Page not found :: Aquatlantis Aquarium - Aquariums, Terrariums & Accessories Manufacturing

There is a special device for dimming this, but it's very limited in its functionalities. So the led-strip must be able to dim, right?

I measured the pwm signal with a scope, the voltage of the drain with an regular multimeter.

I will try the experiments tomorrow.

I measured the pwm signal with a scope, the voltage of the drain with an regular multimeter.

Of course with the 'scope see 0-24V [u]PWM[/u] (or approximately 0V at 0% PWM and 24VDC at 100% PWM). If you're seeing 19VDC the PWM isn't working at all.

Usually the 'scope is referenced to earth ground so you may have to measure the MOSFET's Drain relative to Arduino-ground. In that case it will be "inverted" with 0V when the LEDs on and 24V with the LEDs off.

Im using a led-strip of the following series: 404 Page not found :: Aquatlantis Aquarium - Aquariums, Terrariums & Accessories Manufacturing

There is a special device for dimming this, but it's very limited in its functionalities. So the led-strip must be able to dim, right?

If we only knew how that dimming device works... Does it attach separately to the LED module, or inline with the power supply, or is it a separate power supply?

If it connects separately, and if you have one, maybe you could connect your 'scope and figure-out what it's doing???

There are no detailed specs but since the power supply is a regular (constant voltage) supply, not a special constant-current LED supply, I have to assume there are some active electronics in the enclosure with the LEDs. We don't know how that works, but it would typically be a constant current source and that would "fight" any attempts to dim it. If it were simply LEDs & current limiting resistors it would dim with PWM, but you loose about half of your efficiency.

The IRFB4610 is not a logic level MOSFET, so it won't be turning properly on at 5V.

As for the not turning off, try shorting gate to source - if the strip still doesn't turn off you must have zapped the
MOSFET (probably by not observing anti-static precautions, usually MOSFETs have no protection circuitry on the
gate and require very careful handling till they are in-circuit.)

The 1M pull down on the input to the '245 is a very weak pull-down, it may not be enough to protect against crosstalk from neighbouring signals. 100k is about the maximum logic pull-down or pull-up for local use.

Well, you were right. I 'zapped' it, I think.

The FET seems to be malfunctioning. As soon as I swapped it with a new one, it worked flawless.

Anyway, I want to thank you guys very much for your ideas and suggestions! Well appreciated!