12VDC Reverse Polarity Protection - Low Side Diode?

[meshtron]

Working on the 2nd revision of my first-ever PCB design which is basically a "smart", high-side half-bridge driver for use in 12VDC automotive applications (aftermarket).

I'm using a nice IC for the load switching between Vbatt and Vout that has reverse polarity protection built in.  I pull power from Vbatt through an LM7805 to run the MCU and other logic-level stuff.  While researching reverse polarity protection, I found a recent discussion here with this comment at the end which was helpful.  I didn't want to hijack that thread so created this one to ask: is my use case here one where a low-side diode is a good choice for reverse polarity protection?

My 2-layer board uses almost the whole bottom surface as a ground plane, and with 4 wires on the thing I felt like GND was the logical, central spot to try to stop wrong-way flow.  It's hard to guess all the wrong ways someone could try to hook the thing up, so I put a B260S1F where shown on this lovely schematic (between the GND wire and the GND plane of the board).  I didn't want the voltage drop from the diode on - for example - the CTL (control) input wire even though I could've tolerated it on the input to the 7805.

Now I'm curious about where this falls on the spectrum of "unconventional" to "wrong in many ways" - appreciate any insights!

EDIT: Reworded slightly to try to make it more clear.

[Terry Bites]

I'm putting on my tinfoil hat to protect myself from the information overload this post.

[meshtron]

I'm putting on my tinfoil hat to protect myself from the information overload this post.

This part is always character-building for me.  Trying to figure out if I asked a bad question, asked it poorly, if it's so obvious that it's not worth answering, etc. 

[Manul]

I didn't want the voltage drop from the diode on - for example - the CTL (control) input wire

And what about ground diode? That also affects CTL levels, it does not really matter where the diode is. CTL input signal is not a magical one wire signal, it has return path and it is referenced to it.

I say, don't do it, looks ugly.

If you want reverse polarity protection for Vbatt, add series diode - perfect.

If you want protection for CTL signal, add series resistance and clamping. In essence - limit the current. With limited current and clamping you can make it tolerate even +-100V or more. That's what everyone does. If signal is supposed to be very fast, there are caveats, but in this case I assume you are not feeding in Mhz order signals.

[meshtron]

And what about ground diode? That also affects CTL levels, it does not really matter where the diode is. CTL input signal is not a magical one wire signal, it has return path and it is referenced to it.

I say, don't do it, looks ugly.

If you want reverse polarity protection for Vbatt, add series diode - perfect.

If you want protection for CTL signal, add series resistance and clamping. In essence - limit the current. With limited current and clamping you can make it tolerate even +-100V or more. That's what everyone does. If signal is supposed to be very fast, there are caveats, but in this case I assume you are not feeding in Mhz order signals.

Thanks for the input.  I already have 470 Ohms series resistance (behind a voltage divider to cut down under 5V) on CTL, so maybe it's already okay.

I hadn't thought about the fact that the diode set up this way would effectively raise my reference for external signals by Vf.  That is probably easy enough to manage with firmware, but your point stands that it's a pretty dumb thing to impose upon myself.

I appreciate the response!  Next go-round I'll use a series diode between Vbatt and my 5V regulator and add a clamp on CTL.  Your assumption is correct RE MHz signals, at most 10s of kHz is all I ever expect to see there.

[Manul]

Voltage divider itself will intrinsically limit current. Generally there is no need for extra resistors. With low frequency signals you may use much higher resistance than 470R. You likely can use 47k on top leg. With high enough series resistance even internal MCU ESD diodes will provide enough clamping to survive huge overloads, like even plugging to mains AC.

The only caveat here is that voltage divider output impedance (equal to both legs in parallel) and MCU input capacitance (plus parasitics) make a RC low pass filter. The cutoff must not be too low for your signal. Secondly, even if signal is of low fundamental frequency, you should consider it's rise and fall times (assuming it is digital). Slow rise and fall times and noise might very well cause trouble if you feed it into an input without schmitt trigger. For example, if interrupt is attached to that MCU input, you might get repeated interrupts due to slow rise time and noise. These things can be solved in software or hardware (signal conditioning). These concerns are valid not only with low pass filter present, but generally if the signal source is little controlled/external/unknown.