6

In the circuits below, what does the 100R resistor do?

In my understanding the 1K along with the 200 pF create an integrator topology to provide some stability but how about the 100R that feeds the gate of the MOSFET?
What does it do?

Wouldn't that delay the on/off time of the MOSFET's channel?
Wouldn't we want the MOSFET to instantly respond to every fluctuation of the channel's current?

Can this resistor be larger? 1K or 3K?

enter image description here enter image description here

Null
  • 7,448
  • 17
  • 36
  • 48
MrBit
  • 1,973
  • 4
  • 27
  • 54

3 Answers3

9

In my understanding the 1K along with the 200pF create an integrator topology to provide some stability

Correct and, particularly when using a MOSFET in these configurations there can be a significant amount of gate-source capacitance (2 nF for the Si9430) that can cause stability problems hence, the output of the op-amp is "isolated" from that gate-source capacitance by using the 100 Ω resistor.

Wouldn't that delay the on/off time of the mosfet's channel?

Yes it will but, without that isolation, the localized feedback capacitor cannot create an integrator. This would leave you with the only option of reducing the DC gain with a localized feedback resistor and, of course, that ruins output current regulation.

Wouldn't we want the mosfet to instantly respond to every flactuation of the channel's current?

That's not going to happen because of the necessity of the integrator.

Even though the MOSFET is used as a source follower, the gate-source capacitance (that is usually neutralized in an ideal MOSFET) still remains significantly high. BJTs are quite often used instead of MOSFETs and can be more easily made stable.

Can this resistor be larger? 1K or 3K?

The time constant of the integrator has to dominate the time constant of the gate resistor and residual gate-source capacitance so, probably not even worth trying. Of course, you might find a limited set of output currents where it might work but, if you want a generally half-decent constant current source, then no.

Andy aka
  • 434,556
  • 28
  • 351
  • 777
  • Wouldn't that delay the on/off time of the mosfet's channel causing an extra delay in ciruit's feedback loop? – MrBit Apr 28 '22 at 17:53
  • 1
    Yes it does @MrBit but, you need the op-amp to act as an integrator and, if the gate is directly connected to the output you can't get localized op-amp feedback that makes an integrator. – Andy aka Apr 28 '22 at 17:54
7

Power MOSFETs have very large input capacitances (thousands of pF). In conduction with the output resistance of the opamp that will cause significant phase delay that would affect the feedback loop to make it unstable.

The 100 ohm resistor isolates this capacitance form the opamp to avoid its effects.

The 1k resistor from the current sense resistor allows the feedback to be effective.

JYelton
  • 32,302
  • 33
  • 134
  • 249
Kevin White
  • 32,097
  • 1
  • 47
  • 74
4

It is common practice to put a resistor of low value before a MOSFET gate. It is particularly important when the MOSFET is to be switched ON/OFF at high frequency. MOSFET have an input capacitance that is of significant value, especially at high frequency. So putting a resistor at the gate will limit possible high current and damage to the MOSFET input.

In this case where the device is not submitted to high switching frequency, it may be avoided but at the price of a resistor, why not be on the safe side?

Fred Cailloux
  • 799
  • 7
  • 15