| Electronics > Beginners |
| Solenoid driver MOSFET flyback protection |
| (1/3) > >> |
| HwAoRrDk:
I want to drive a 12V vacuum solenoid with a MOSFET, and to drive it I had in mind to use a nice component I came across a while back, the ON Semi NUD3124 - an N-channel MOSFET, protection diodes and bias resistors all packaged conveniently in SOT-23. However, the solenoid I have draws more current (>420mA) than the NUD3124 can handle, so that is unfortunately out. What I thought to do instead was replicate the integrated parts of it with discrete components - a bigger MOSFET, diodes, resistors, etc. But, I am not quite 100% on understanding why some of the NUD3124's integrated component parts are like how they are, and I was hoping someone could offer some explanation. The diagram of the integrated components, from an associated app note, is as follows: As I understand it, this is an active clamping arrangement; when the solenoid's turn-off flyback voltage spike occurs, this passes through the 28V clamping zener and turns the gate back on, thus routing transient current through the MOSFET. My questions about this are: - Why are there two pairs of 14V gate zener diodes? Obviously, they are not exactly in parallel - there is a 10K resistor between - but it still seems to me the left-most pair is redundant. :-// - Also, why are those zeners 14V when the datasheet says the Vgs has an absolute maximum of 12V? - What's stopping all the transient current flowing through the 28V and 14V zeners to ground, instead of partially through the FET as described? Also, if I want to reproduce this arrangement with discrete components, does it particularly matter what value of zener diodes I use on the gate, so long as it is above my gate drive voltage (5V) and below the maximum Vgs? I already have some 5.1V zeners, so was thinking of using those. Also an 18V bi-directional TVS diode for the clamping between drain and gate. |
| Benta:
The NUD3124 probably has a lot more protection than you need, it's designed for automotive applications where a load dump carries a lot of power. In the simplest application (like your solenoid), the upper 28 V Zener is likely all that's needed. The leftmost 14 V Zeners are not there to protect the device, but the circuit driving it. The next set of 14 V Zeners are for protecting the MOSFET gate. And the lower 28 V Zener is for clamping a negative spike during a load dump. You decide, it depends on your application. |
| HwAoRrDk:
--- Quote from: Benta on November 26, 2018, 05:33:48 pm ---The NUD3124 probably has a lot more protection than you need, it's designed for automotive applications where a load dump carries a lot of power. --- End quote --- Ah, well, see, it will actually be used in an automotive application. :) --- Quote from: Benta on November 26, 2018, 05:33:48 pm ---The leftmost 14 V Zeners are not there to protect the device, but the circuit driving it. --- End quote --- Okay. I still don't understand why one set of 14V zeners doesn't serve as double duty for protecting both the gate and the driving device (i.e. MCU). Or this just a belt-and-braces approach? |
| Ian.M:
The left-most Zeners protect the 10K and 100K resistors against flash-over during ESD events which could erode the (presumably thin film) resistance element and eventually cause them to go open circuit. 5.1V zeners are a *bit* too close to 5V CMOS logic levels. Consider what happens if they are 5% under their nominal zener voltage and also what happens if the5V rail is slightly higher than expected. A TVS diode, unless very low energy handling capacity, is likely to have excessive junction capacitance, resulting in significant extra dissipation in the MOSFET during switching due to the Miller effect. |
| HwAoRrDk:
--- Quote from: Ian.M on November 26, 2018, 11:36:09 pm ---The left-most Zeners protect the 10K and 100K resistors against flash-over during ESD events which could erode the (presumably thin film) resistance element and eventually cause them to go open circuit. --- End quote --- I presume the resistors are susceptible to that because they are on-die and so are physically very small? So I guess that pair aren't at all necessary in a discrete component circuit, because there is little chance of such happening to an SMD or THT resistor. --- Quote from: Ian.M on November 26, 2018, 11:36:09 pm ---A TVS diode, unless very low energy handling capacity, is likely to have excessive junction capacitance, resulting in significant extra dissipation in the MOSFET during switching due to the Miller effect. --- End quote --- Not familiar with that phenomenon, will have to look it up. As a wild stab in the dark, I would guess capacitance in the diode causes turn-on and -off to be too slow, making the FET spend too much time in the linear region? I plan to PWM the solenoid to reduce hold current, so such a side-effect would probably be exaggerated under that condition? I have some regular 18V zeners, though - just thought a bi-directional TVS would be convenient as it's both diodes in one component. |
| Navigation |
| Message Index |
| Next page |