1
\$\begingroup\$

I am trying to use PMOS SSM3J334R for charging battery. But I notice that my charging mosfet is getting heated even at less then 1A current.

On checking waveform I found very bad off time of MOSFET. Can someone help me to find issues in my circuit.

Schematic

Mosfet Parameters

In plot below: blue - base, red - gate BJT PMBT3904 base and PMOS gate voltage plot

In plot below: blue - base, red - drain BJT PMBT3904 base and PMOS Drain voltage plot

In plot below: blue - gate, red - drain PMOS Gate versus drain voltage

\$\endgroup\$
5
  • \$\begingroup\$ You need to repalce that 470 ohm passive pullup with an active pullup. At this point, it might be worth investing in an actual gate-driver chip. \$\endgroup\$
    – Dave Tweed
    Dec 11, 2014 at 15:53
  • \$\begingroup\$ What the value ov your Vcc at the input, is a PWM or DC voltage? \$\endgroup\$
    – R Djorane
    Dec 11, 2014 at 16:12
  • \$\begingroup\$ By using the swiching biais to reduce the frequency noise and the power consumption. \$\endgroup\$
    – R Djorane
    Dec 11, 2014 at 16:17
  • \$\begingroup\$ Hello Dave, Can you suggest me some active pullup to be used here. Is it something like PMD3001D. \$\endgroup\$
    – user45805
    Dec 12, 2014 at 11:13
  • \$\begingroup\$ Hello Codo input is provided at connection of R14 and R1. Input is DC voltage from solar panel and goes upto 21 volt \$\endgroup\$
    – user45805
    Dec 12, 2014 at 11:15

2 Answers 2

1
\$\begingroup\$

The gate capacitance of the device is "holding-on" the MOSFET. This is because the only component trying to turn off the gate signal is a 470 ohm resistor. The time constant of the 470 ohm resistor and gate capacitance (around 1nF) is about 0.5 micro seconds. As we don't know what MOSFET you are using nobody can be sure about the gate capacitance but, it could be as high as 10nF for some devices.

This means the CR time constant is between 0.5us and 5us. Dave tweed has suggested using a proper driver (it will have a push-pull output that can discharge the gate capacitance in a few tens of nano seconds) but another alternative is to raise the top leg of the 470 ohm 15 volts (ish) above the power rail. This then causes the switch off time to be decreased because the "aiming voltage" of the " CR discharge" will substantially turn off the MOSFET in a fraction of the time.

\$\endgroup\$
4
  • \$\begingroup\$ Hello Andy, I agree. I have added my mosfet parameters in my original post. Can you suggest me some proper driver as you mentioned in your response. Will PMD3001D work for this \$\endgroup\$
    – user45805
    Dec 12, 2014 at 11:26
  • \$\begingroup\$ That device looks Ok to me but I think your mosfet may also be a little low power for your application - at a 5A drain current the device will dissipate 1.25 watts. I wouldn't run it more than 2 amps. What power supply rail do you have? \$\endgroup\$
    – Andy aka
    Dec 12, 2014 at 12:22
  • \$\begingroup\$ Hello Andy, I want to use it for around 1A current. When you say device OK, you mean PMD3001D ok? \$\endgroup\$
    – user45805
    Dec 13, 2014 at 8:38
  • \$\begingroup\$ Yes I believe it to be ok. \$\endgroup\$
    – Andy aka
    Dec 13, 2014 at 10:51
1
\$\begingroup\$

The effective capacitance that the driver sees is a combination of Gate capacitance and feedback between the Gate and Drain. The critical region occurs when Gate voltage approaches cutoff (about 2.6V in the SSM3J334) and the FET starts to turn off, then the effective capacitance increases dramatically and Gate voltage flattens out. To get past this point quickly you need a high current driver.

enter image description here

Adding a simple complementary emitter follower to increase drive current should reduce your turn-off time to about 1/10th of what it is now.

enter image description here

Waveforms with changes from Bruce, but 100ohm resistance between R14 and T5

Base - Gate plot(base blue, gate red) enter image description here

Base - Drain plot(base blue, Drain red) enter image description here

Gate - Drain plot(Gate blue, Drain red) enter image description here

\$\endgroup\$
6
  • \$\begingroup\$ Thanks Bruce as per your recommendation I will make changes and let you know \$\endgroup\$
    – user45805
    Dec 13, 2014 at 8:39
  • \$\begingroup\$ Hello Bruce just one question before I make changes, VGS max of SSM3J334R is 20V and VCC can go till 25Volt. Can it destroy mosfet as when T5 is off T6 is ON and gate voltage is 0V. \$\endgroup\$
    – user45805
    Dec 13, 2014 at 11:20
  • \$\begingroup\$ Voltage at the output of T6 and T7 follows the input, so just put a resistor between T5 and R14 (like in your original circuit). Since the FET only needs 10V gate drive you could make the resistance a bit larger to provide some safety margin. You might also consider using a more robust FET with higher voltage and current ratings. \$\endgroup\$ Dec 13, 2014 at 18:41
  • \$\begingroup\$ Hi Bruce I tested and captured waveform. There is lot of improvement but still off time is not order of 1/10th. Any further suggestions \$\endgroup\$
    – user45805
    Dec 14, 2014 at 10:00
  • \$\begingroup\$ Gate discharge looks to be about 5 times faster, but you are getting lot of bounce on the supply rail. Do you have a bypass capacitor on it? You might get some improvement by reducing R13, however discharge current will still be limited by R14 and current gain of T6. What transistor are you using? \$\endgroup\$ Dec 14, 2014 at 11:42

Your Answer

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

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