2
\$\begingroup\$

Following circuit drives a 24V LED strip. Vin is a 3.3V 1kHz PWM signal coming from an ESP32 GPIO. The MCP1416 MOSFET driver converts the input signal to a 5V PWM with low output impedence. The MOSFET SQ2362ES drives the LED strip with a max drain current of 1A. A TVS diode SMAJ10CA+ provide the overvoltage protection for the gate, while the flyback diode is a SS115 schottky. The 5V and 24V come from 2 bench power supply. Circuit schematics

I tested the circuit both on breadboard and on the final PCB, here the graph of Vin and Vds when the MOSFET turns off:

Oscilloscope output

Regarding the ringing:

  1. Why this huge ringing up to 70V the driver?
  2. What can I do to suppress it? Placing a 100 ohm resistor between driver output gate reduces the ringing to 50V. Adding a 0.1u ceramic cap between L1 connector and ground gives a further reduction to 15V. Should I place the gate resistor? Should I place a decoupling cap? A film cap could work better? What about an electrolitic one?

Regarding Vds with MOSFET off:

  1. Vds should be 24V. Why do I measure 9V? Is it due to the scope probe?
  2. If so, how should I measure this voltage?
  3. Placing the probe on the LED connectors results in a stable 24 V signal, without PWM, and the LEDs are fully switched on. Why this behavior?

UPDATE

Here the arrangement of the components on the breadboard. The MOSFET is a IRF520, but the behaviour is the same with the original MOSFET. I added also a 10 Ohm gate resistor and 4x 0.47uF decoupling cap on 24V rail. The circuit was also tested on a PCB with same results. The image shows also the scope probes (x10).

Circuit on breadboard

\$\endgroup\$
7
  • \$\begingroup\$ Point 2 - > "Placing a 100 ohm resistor between driver output gate ..." Generally, the value is 10 Ohm max ... \$\endgroup\$
    – Antonio51
    Commented Jan 31, 2022 at 15:57
  • \$\begingroup\$ Point 1 & 2: Have you tried placing a (1nF ?) capacitor between L1 and L2? If LEDs strip can be considered as a "line", this can explain why you have "high" pulse Vds (50 V). I don't have in my database the SQ2362 MOS-FET, so I tried with IRF6674 and a T-line of 5ns (1 meter). \$\endgroup\$
    – Antonio51
    Commented Jan 31, 2022 at 16:29
  • \$\begingroup\$ Point 1: how can you obtain 70V for Vds without clipping ... Vds max of MOS-FET is 60 V? \$\endgroup\$
    – Antonio51
    Commented Jan 31, 2022 at 16:39
  • \$\begingroup\$ @Antonio51 I tried to place a 10n cap between L1 and L2 and the ringing dramatically increased up to 130-140V. Anyway, I didn't understand what you mean with "If LEDs strip can be considered as a line, this can explain why you have high pulse" and the test that you done with IRF6674. Can you explain better? \$\endgroup\$
    – Marco
    Commented Feb 7, 2022 at 8:33
  • 1
    \$\begingroup\$ Please show your probe setup. \$\endgroup\$
    – winny
    Commented Feb 7, 2022 at 21:40

2 Answers 2

Reset to default
3
\$\begingroup\$

If one considers that LEDs strip is a "transmission line", here what happened.

Update: Diode must be a fast recovery diode as MURA220!
NB: see the condition \$ di/dt=50 A/us \$ --> \$ trr = 35ns \$. Interesting article
Always use a short wired decoupling capacitor (0.1uF->10uF...) onboard between pins supplies.
Always use short twisted wires for power supplies ...
With fast switching, ... don't use breadboard ... See last 2 pictures.

(simplified schematic, all strip is one line 1 meter - 5ns delay - line should be divided into more sections)

enter image description here

With diode ...

enter image description here

When strip divided in 5 sections ...

enter image description here

I add a simulation about wiring (only 10 cm)... as also @kruemi suggestion ...

enter image description here

Wires have no "inductance" ...

enter image description here

Wires have 1uH inductance ...

enter image description here

\$\endgroup\$
1
  • 1
    \$\begingroup\$ Comments are not for extended discussion; this conversation has been moved to chat. \$\endgroup\$
    – Voltage Spike ♦
    Commented Feb 7, 2022 at 22:01
1
\$\begingroup\$

The problem is your '24V'; nearly independent of the LED strip.

When the FET turns off and current was flowing from the 24V through the LED strip and the FET, the collapsing flux in the inductance of the line to the 24V will try to keep the current flowing in the FET -- this drives the drain to a v. high (damaging) voltage.

Although your schematic shows a Schottky diode clamping to '24V' -- what matters is the inductance between the 24 V supply and this circuit (in both the 24V line and the GND line). With (not on the schematic) inductance, you won't see the problem.

You could add a capacitor (1..10 uF ?) from 24 V on this board to GND, or you could add a Zener Diode across the supply on this board.

\$\endgroup\$
1
  • \$\begingroup\$ I tried to measure the +24V rail with the scope and I see ringing below 2V when the MOSFET switches off. I don't have 10uF cap, so I placed 10x 0.47uF polyester caps in parallel between +24V and GND, but nothing changed. I don't have either a 25V zener; can I place 5x 5V zener in series? \$\endgroup\$
    – Marco
    Commented Feb 8, 2022 at 13:30

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