emergency switch or disconnect

[Electro nS]

i have a dc powered circuit from 15-to 50 A . 24-48v and maybe higher later !

the microcontroller issues at some fault a disconnect order . this should disconnect batteries by relay .

the question , can i replace the relay by a N-channel mosfet on the low side. ?? so the gnd is disconnected ? is this a good practice ?

i always read that GND should be secured and Vcc disconnected but why ??????? is it the same ?

thanks in advance

 

[tpetar]

https://www.edaboard.com/threads/182857/

 

[WimRFP]

I feel your preference for a low side mosfet switch (ease of driving, no high side driver or Pchannel mosfet required).

Determine for youself what else can go wrong when the circuit is switched off (via the low side mosfet). Most people assume that when something is turned-off, the "plus" is safe to handle (for example with electrically conducting tools). Now there can be a path from the positive conductor via something else to grounded metal and then your switch is bypassed (depending on your circuit design).

If your analysis turns out that it isn't save to switch the ground wire, you can use a P-channel mosfet, or use a N-channel mosfet with high side driver (you very likely know that this requires a gate voltage that is above supply voltage).

When the mosfet is also used as overcurrent or shurt circuit protection, you need to select a mosfet with sufficient save operating area so that it will be able to disconnect a short circuit without smoke signalling.

 

[Electro nS]

i know about gate drivers and mosfets in high side configuration , and the need for internal charge pump to achieve 100% duty cycle ! and P-channel are hard to find in high current low Rds comparing to N -channel .
i think i will try the N-channel at the low side and see what might happen . but it is still unclear what could possibly go wrong if the ground is disconnected and the VCC is still supplied to the circuit , hope i would find out the easy way

 

[WimRFP]

This is not to offend you, but if you can't figure out whether or not something can go wrong when switching the "minus" terminal, I would stay away from such circuits.

When working with > 2 kW circuits: "i think i will try the N-channel at the low side and see what might happen" gives me a bad feeling.

When the "minus" terminal is connected to ground, a short between the "plus" and ground will result in a short circuit, no matter the condition of the mosfet. You need at least a floating "minus" terminal so that a short between "plus" and ground (or metallic casing) does not result in a short circuit.

 

[schmitt trigger]

No, it is not a good practice, don't do it.

One never knows how the "earth", "chassis ground", "signal ground", "common", or "negative" rails are all inter-connected between different boards and circuits. You most likely will have sneak paths thru one of them.

I could tell you many horror stories, I've seen amplifiers whose current returned thru an audio cable which overheated and caught fire; but the latest I saw was a sneak path thru an USB cable. The cable evaporated, and both the computer and test equipment were damaged.