3
\$\begingroup\$

So, I'm trying to control a LED strip with a NodeMCU using N-MOSFETs FQP30N06L, one for each channel. However, the strip is not nearly as bright as with the original IR controller. (I saw the similar question, however my transistors are definitely logic-level)

The main problem SEEMS to be the mosfets, as if I connect the LED strip chanels directly to ground i get a much much brighter light. Also, there seems so be a quiet high pitch sound when using the transistors.

The circuit is very simple, strip (5m, 30 leds/m) runs on 12V and should take 1.5A, so 0.5A for each channel, and that's what the power supplies. Looking at the datasheet for the MOSFETs, it looks like at 3V G_gs is should let through 10A, way more than the 0.5 I require. (However it is my first time looking at a datasheet so that's probably where my mistake is, though I have no clue what).

I have a multimeter, just wouldn't know where to even start measuring things.

Also I'm not connecting the MCU to 12V as I know that's too much for the voltage regulator, I just couldn't find a good enough cirtcuit drawing tool, so any suggestion for that is welcome as well.

\$\endgroup\$
5
  • 1
    \$\begingroup\$ Are you driving this with pwm? If so you probably need a mosfet driver instead of the IO pin. mosfet gates are essentially a capacitor and a small microcontroller cannot charge and discharge it very fast for something like pwm switching. one way to test this would be to lower the pwm frequency greatly and see if your noise goes away and also your lights get brighter... I'll make this an answer if it works out for you. \$\endgroup\$
    – MadHatter
    Apr 25 '20 at 14:57
  • \$\begingroup\$ You could try measuring the voltage drop across your transistor when it's on. That could give us some information. \$\endgroup\$
    – MadHatter
    Apr 25 '20 at 15:03
  • \$\begingroup\$ @MadHatter seems like a plausible problem. Do you mean the voltage between drain and source? \$\endgroup\$
    – Lorenzo
    Apr 25 '20 at 17:56
  • 1
    \$\begingroup\$ Create a simple test setup without NodeMCU: A single MOSFET and the LED strip (one channel only). Then feed 12V to the drain, 3V or 3.3V to the gate and connect the source and all grounds. Now measure all the voltages and the current into the drain and report back. \$\endgroup\$
    – Codo
    Apr 25 '20 at 18:35
  • 1
    \$\begingroup\$ Turns out my MCU has a range for pwm of 0-1023 unlike the 0-255 from Arduino. So I was just running it about 25% duty cycle. Apologies for wasting everyone's time. I'm new to the electronics stackexchange so I'm not sure if I should accept an answer? \$\endgroup\$
    – Lorenzo
    Apr 25 '20 at 18:40
3
\$\begingroup\$

Fortunately, the graph you want is actually next to the one you used and I have marked red lines on to explain where you problem is: -

enter image description here

So, on the graph on the right (the one you used) I have marked a vertical red line corresponding with a gate source voltage of 3 volts. Now, at the point where the red line hits the traces if you looked across at figure 1 you can see what that means. Note - the current is about 13 amps.

Can you see that the 2nd line I drew (the horizontal red line) crosses a bunch of traces in figure 1? It doesn't cross the lower trace (not quite) and that is the trace when \$V_{GS}\$ is 3 volts - this means that driving it at 3 volts isn't going to deliver the goods (13 amps) typically.

Take the next trace up where \$V_{GS}\$ is 3.5 volts - you can see that my horizontal red line crosses it at a drain source voltage of about 0.7 volts (vertical blue line). This typically means that if you wanted to deliver 13 amps to your load, the MOSFET would drop about 0.7 volts from drain to source.

If you drove \$V_{GS}\$ at (say) 10 volts, can you see that to deliver 13 amps, the MOSFET would drop about 0.35 volts (vertical green line).

Given that you only need 0.5 amps I would say this MOSFET would drop significantly less than 0.1 volts across drain and source even when fed with a \$V_{GS}\$ voltage of only 3 volts.

So, my conclusion is that it can't be the MOSFET that's giving you the problem. Maybe it's the interconnecting wires or your power supply isn't able to deliver the current you need. Maybe your MOSFET is "fake"?

However it is my first time looking at a datasheet so that's probably where my mistake is, though I have no clue what

Hopefully this clears up where the problem doesn't appear to be - use fig 1 in future - it delivers what you need to know in this respect.

\$\endgroup\$
2
  • \$\begingroup\$ Thanks for the prompt answer. I would still say the main problem seem to be with the transistors (have a look at the second paragraph I just edited). Also, what do you mean by a problem with interconnecting wires? As in not plugged in properly? \$\endgroup\$
    – Lorenzo
    Apr 25 '20 at 12:08
  • \$\begingroup\$ @Lorenzo yes I read that 2nd paragraph - where did you acquire them? Interconnecting wires - bad connections (ohmic) or too much resistance due to their length. \$\endgroup\$
    – Andy aka
    Apr 25 '20 at 12:17
2
\$\begingroup\$

Based on you mentioning noise, I'm assuming you are using a PWM control scheme of some sort? Are you sure you are driving the devices on 100% duty cycle.

Also if you are switching with PWM, you may need a dedicated mosfet driver, the IO pin on your micro controller is not ideal for driving mosfets with a PWM like frequency. The gate acts like a capacitor that is charged and discharged. Doing this quickly requires more then the 20mA drive of most microcontrollers.

\$\endgroup\$
0
\$\begingroup\$

The datasheet says that the gate-source threshold voltage could be as high as 2.5 Volts. Maybe you have a batch that came out at the high end, and your 3-volt gate-source voltage is not quite enough. Do you have a way to temporarily apply a higher voltage to the gate as a test?

\$\endgroup\$
2
  • \$\begingroup\$ I'll try with the 12V from the strip power, though the MCU gives 3.3V which is almost 33% over the maximum gate threshold \$\endgroup\$
    – Lorenzo
    Apr 25 '20 at 13:40
  • 1
    \$\begingroup\$ @Lorenzo - "which is almost 33% over the maximum gate threshold" Please go to the data sheet. On the spec page where you get the gate threshold, Vgs(th), pay attention to the test conditions. At threshold, the FET will conduct 250 uA, or 1/4 mA. Adding another another volt to that will not give you a very low resistance, as Andy aka has pointed out. \$\endgroup\$ Apr 25 '20 at 14:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.