So this is part 3 (or 4 depending how you look at it) in a series of questions I've had around PWMing an RGB LED strip from an ESP8266 (check my question history if you desire).

I apparently keep misunderstanding/misreading/misinterpreting the MOSFET datasheets. I believe my previous mistake (with the IRFZ44N) was interpreting the "max" \$V_{GS(th)}\$ as the maximum I should give it vs the maximum it will require to fully turn on (please correct me if I'm wrong though).

This time around, I found the IRLZ44N that has a max \$R_{GS(th)}\$ of \$2V\$, great! Ordered a few of those and got a very similar result, damn!

This is my schematic:


TP1 (11.4 V coming from the power supply):

tp1 reading

TP2 (3.6 V PWM signal going to the gate of the MOSFET):

tp2 reading

TP3 (I was (again) expecting 11.4 V max, 0 V min here):

tp3 reading

My guess this time is that the \$R_{DS(on)}\$ being specified at \$4~\mathrm{V}\$ is the issue? So I assume that the MOSFET is fully on (due to the \$V_{GS(th)}\$ max value being exceeded), but has too high of a resistance, hence it being dim?

Is that right? Even though I've said that, I'm not sure it sounds correct and I could do with a bit of guidance please.

I'm also aware that I likely don't know the right questions to ask (hence me being back here for the 3rd time), so if you can see what I'm getting at, even though I may not have asked it directly, please can you provide the question and answer?


  • \$\begingroup\$ Your LEDs need current limiting and your MCU decoupling capacitor. \$\endgroup\$
    – winny
    Jan 17 at 19:29
  • \$\begingroup\$ Thanks. I don't know how it's wired up, but my strip has resistors on it, so I think I'm good from that angle. I assume the decoupling cap should be between TP1 and GND? Thanks for the advice. \$\endgroup\$
    – OdinX
    Jan 17 at 19:42
  • \$\begingroup\$ The resistors on the strip will determine the voltage on TP3, so it is a good idea to include them in your schematic, at least approximately. The way you've drawn it now, TP3 will switch between 9-10V and 12V as the transistor turns on/off. If your MCU is actually a complete board with regulator and decoupling caps, then you don't need to add your own. \$\endgroup\$ Jan 17 at 19:47
  • \$\begingroup\$ Yeah, I should have included the resistors in the schematic. Why do you say TP3 will switch between 9-10V and 10V? In the final image, it seems to switch between 4.5V and 5.5V. I don't know why though? \$\endgroup\$
    – OdinX
    Jan 17 at 19:54
  • \$\begingroup\$ TP4. Please name it Vcc. \$\endgroup\$
    – winny
    Jan 17 at 20:07

2 Answers 2


Threshold voltage (VGS(th)) is the gate to source voltage at which the conductive channel between the source and drain is just beginning to form. The datasheet specifies it vaguely as between 1.0 and 2.0 V since you aren't really going to design around it - the datasheet states that 250 microamps of IDS is the threshold.

VGS(max) is the absolute maximum voltage you can apply to the gate, referenced to the source, before damage occurs. For the IRLZ44, this is +/- 10 V.

In between, you have the transfer characteristics given in fig. 1 for 25 C and fig. 2 for 175 C. From the curve, it looks like you should be getting the behavior you expect but there might be some interaction with the internal circuitry of the LED strip or the circuit is connected differently than the schematic.

I would also add a gate resistor between the 10k pulldowns and the gate terminals.

  • \$\begingroup\$ As I said to winny on my question, there are internal resistors on the LED strip, but I don't think there's anything else aside from that. I'm also quite sure it's wired up as per the schematic. Is there something else that I could be missing or is worth trying? \$\endgroup\$
    – OdinX
    Jan 17 at 19:49

The threshold voltage is just that, a threshold. It is like entering a house , if you stop on the threshold you are still not in the house. \$V_{GSmax}\$ is the maximum value at which the threshold occurs. But it is still the threshold. \$R_{DS}\$ is still pretty high.

I use the following chart, from the data sheet, to choose a \$V_{GS}\$ value that I consider ON for PWM purpose.

The flat region where \$V_{GS}\$ barely changes is called the Miller region. In this region \$V_{DS}\$ is dropping requiring the gate-drain capacitance to charge from the gate perspective. This steals current from the gate slowing the voltage increase.

Once \$V_{DS}\$ is zero \$V_{GS}\$ can increase to lower \$R_{DSon}\$.

The gate pulses must rise quickly above this level for good performance. This is significantly higher than the threshold voltage.

Although the x-axis is in coulombs, this effect can be seen by monitoring the gate voltage (zoom in on the rising edge). It is particularly noticeable when switching inductive loads.

enter image description here

  • \$\begingroup\$ Just to rephrase what you've said to check my understanding: My Vgs voltage is still too low and going by that graph I should use something above the flat region, somewhere closer to 5V? \$\endgroup\$
    – OdinX
    Jan 17 at 19:47
  • \$\begingroup\$ Yes. 4volts minimum. \$\endgroup\$
    – RussellH
    Jan 17 at 20:12
  • \$\begingroup\$ And the only way to determine that value is from that graph? I've read quite a lot of material on MOSFETs and I don't think I've even heard it mentioned. Perhaps this is why I've been picking the wrong component, but it feels quite unintuitive that this would be an important metric and it be only depicted in a less prominent graph? \$\endgroup\$
    – OdinX
    Jan 17 at 20:20
  • \$\begingroup\$ Probably not the only way. But it is very convenient. I found it by circuit analysis, measuring, and comparing to data sheets. Once you understand how the miller capacitance interferes with switching it is quite clear. \$\endgroup\$
    – RussellH
    Jan 17 at 20:36
  • \$\begingroup\$ Ok, thank you for explaining. \$\endgroup\$
    – OdinX
    Jan 17 at 20:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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