Maker Pro
Maker Pro

More woes driving LEDs via PWM

FuzzyWombatSoup

Nov 18, 2014
35
Joined
Nov 18, 2014
Messages
35
I have 16 feet of RGB LED strips, cut into 4 foot segments that are physically located in parallel, wired in parallel. I've been having issues with some channels lighting up when powering the circuit resulting in the MOSFETs becoming very hot within seconds. I've found a workaround(when connected to an arduino) was to perform and analogWrite(0) to the channels. After that, during normal operation even at full duty cycle they would barely be warm to the touch. First question, what could be causing this?

I'm trying to move to a RaspberryPi 3 instead of the Arduino to drive these LEDs, but when I connect them to the circuit via respective GPIO ports, I get insane flicker of all channels(about 10hz). I'm not sure how to resolve this, either. I tried other LEDs and did not get this flickering, so I know its something with my design.

This is the circuit design I used .. https://cdn-learn.adafruit.com/asse...dium800/led_strips_ledstripfet.gif?1448059609 except the MOSFETs I have are NTD3055L104G.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Add a resistor between gate and SOURCE of each MOSFET. A value between 1k and 10k should be fine.

edit: I said drain (aaaagh)
 
Last edited:

cjdelphi

Oct 26, 2011
1,166
Joined
Oct 26, 2011
Messages
1,166
Yup what steve said, fets need a pull down or they float...

Arduino's better for this than a pi
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Well, they are necessary if what is connected to them is floating. It sounds like the OP did not have them set as outputs at some point.
 

FuzzyWombatSoup

Nov 18, 2014
35
Joined
Nov 18, 2014
Messages
35
Can you confirm if it's gate to source or gate to drain? Gate to drain made no difference with my first issue. I tried gate to source which helped, but the MOSFETs started to heat up again when turned on, although not as rapidly. I tried 1k ohms, 10k, and 1M. Are these particular devices actually the best fit for this application?
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Definitely gate to source (what was I thinking!).

Do the mosfets heat up when the gate resistors are fitted, but no connection is made to the Arduino? (they shouldn't)

Are you certain you set the pins to be outputs and pull them low as part of the arduino initialization code?

Do you know the PWM frequency being used? The Arduino can't supply a lot of current and the combination of high frequency and slow switching will cause heating of the mosfets.
 

FuzzyWombatSoup

Nov 18, 2014
35
Joined
Nov 18, 2014
Messages
35
Definitely gate to source (what was I thinking!).

Do the mosfets heat up when the gate resistors are fitted, but no connection is made to the Arduino? (they shouldn't)

Are you certain you set the pins to be outputs and pull them low as part of the arduino initialization code?

Do you know the PWM frequency being used? The Arduino can't supply a lot of current and the combination of high frequency and slow switching will cause heating of the mosfets.
They are cool to the touch(and the lights off as desired) with the resistor and nothing connected. I'm using the RPI 3. Even when I set it to full duty cycle(after I connect), the LEDs are only a fraction of their brightness compared to when I had the connected to the arduino... I'm not sure why. As such, they are not getting warm. I have not mucked with the frequency, I think RPI defaults to 100hz?

Edit: even with the very dim light, they are getting veryhot.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
What voltage do you get with the rpi outputs? is it 3.3V? What current are you switching?

Have you tied the rpi ground to the source of the mosfets?

Have you looked at what is happening at the rpi output with an oscilloscope (or even measured it with a multimeter)?
 

FuzzyWombatSoup

Nov 18, 2014
35
Joined
Nov 18, 2014
Messages
35
What voltage do you get with the rpi outputs? is it 3.3V? What current are you switching?

Have you tied the rpi ground to the source of the mosfets?

Have you looked at what is happening at the rpi output with an oscilloscope (or even measured it with a multimeter)?
That cued me in to the problem.. I was not grounding to the RPI as well as the power supply for the LEDs. The MOSFET for red seems to get hotter than others(which are no longer heating up), but I guess I just need to heatsink that.
 

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
If the PI is outputting 3.3V, that might not be enough to turn the MOSFET on completely, resulting in heat.

What is the part number of the MOSFET? You need ones with very low threshold voltage, which will turn on completely (or nearly so) at 3.3V.

Bob
 

FuzzyWombatSoup

Nov 18, 2014
35
Joined
Nov 18, 2014
Messages
35
If the PI is outputting 3.3V, that might not be enough to turn the MOSFET on completely, resulting in heat.

What is the part number of the MOSFET? You need ones with very low threshold voltage, which will turn on completely (or nearly so) at 3.3V.

Bob
It is 3.3v output, I verified that. It's the MOSFET is a NTD3055L104G
 

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
Those MOSFETS are not fully turned on at 3.3V. You need to find MOSFETS that have their on state characterized at 3.3V with a low enough Rdson to avoid heating up.

Bob
 

FuzzyWombatSoup

Nov 18, 2014
35
Joined
Nov 18, 2014
Messages
35
Those MOSFETS are not fully turned on at 3.3V. You need to find MOSFETS that have their on state characterized at 3.3V with a low enough Rdson to avoid heating up.

Bob
If I'm reading the datasheet right, the VGS(th) is minimum 1V and max 2V. Why is that not sufficient, it's under the 3.3V supplied.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Because that's the threshold voltage. The current through the device at the threshold is defined as some small (and typically useless) value.

Check the graphs of Ids vs Vds for varying values of Vgs. You probably want a couple of amps, and to do this without significant voltage drop across the MOSFET, you will want to turn the MOSFET on hard enough to carry significantly more current. Also note that these values are typically typical ( not min or max).
 

FuzzyWombatSoup

Nov 18, 2014
35
Joined
Nov 18, 2014
Messages
35
Because that's the threshold voltage. The current through the device at the threshold is defined as some small (and typically useless) value.

Check the graphs of Ids vs Vds for varying values of Vgs. You probably want a couple of amps, and to do this without significant voltage drop across the MOSFET, you will want to turn the MOSFET on hard enough to carry significantly more current. Also note that these values are typically typical ( not min or max).
I think I understand. So would a PSMN1R1-30PL,127 meet this requirement?
 

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
Before we can select an appropriate MOSFET we need to know how much current you are going to switch.

Bob
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
That MOSFET might work for you. The gate charge is pretty high, so you may need to consider this if you need to switch rapidly or at high frequency.
 
Top