Overvoltage protection circuit

[ldkimball]

I'm working on a comparator overvoltage protection circuit for my 12V DC circuit to shut off the power supply to the downstream components if the voltage exceeds around 16V. A 5.6 zener diode sets the reference voltage and a voltage divider divides the line voltage by three to compare to the reference. I also have a window comparator circuit with two open collector comparators connecting to a MOSFET to monitor that the voltage stays between 0V and 5V in a certain area (AD1OUT). I've attached pictures of both of these (the first is the overvoltage protection and the second is the window comparator circuit).

This is my first time designing voltage protection circuits so I want to make sure I'm doing it right. Is there any reason these wouldn't work? Also, is there a downside to having two mosfets in series on the line voltage?

Thanks!

[Kim Christensen]

Try reattaching your diagram. Sometimes the forum bugs out and you need to edit your post so that it "remembers" your attachment.

[ldkimball]

Can you see them now?

[Kim Christensen]

Yes... Not a fan of those types of schematic symbols though. Draw your MOSFETs and OpAmps using standard symbols like you did with Q2.
You can't use N-FETs (Q2) to switch the positive rail without driving the gate voltage above the positive rail.
The simple solution is to use a P-FET like you did with the SI4435 in the 1st circuit. You need to limit the max gate-source voltage to less than 20V.
Your OpAmps should have some hysteresis (Add resistor between output and + / noninverting input) so that the load doesn't rapidly turn on/off when the input voltage is close to the trip point.

[ldkimball]

Thanks for your reply!

Good catch re the N-FET. I tried swapping the inverting and noninverting inputs of the op amps earlier so that the window comparator circuit would be compatible with a P-FET but since the comparators are open collector I decided they would pull low outside of the the window instead of high since pulling "high" is just a high impedance floating state and the low of the other comparator would override. Am I wrong about this?

[Kim Christensen]

Correct.
For the window comparator, you could leave the OpAmps as is and use a small N-FET, or NPN transistor, to invert the gate signal of the P-FET.

[ldkimball]

That makes sense. Is this what you're suggesting?

Thanks!

[Kim Christensen]

Yes, that's the general idea.
EDIT: You could use a LM393 instead of the two LM311.

[ldkimball]

Perfect, thanks!

[PGPG]

When in 2004 I was for the first time in EMC lab they haven't some capacitor that should be used to test Surge at signal lines (and may be low voltage DC supply). We tested all my signal inputs and 12V supply with 50us 1kV pulses connected to device pin via 40Ω resistor (brick size).
We (Poland) just entered EU few months before so CE and EMC tests were new to us. They said that I was the first person in their history who came from outside (not their designer) and just passed all the tests.
Since than I am designing all inputs to pass 50us 25A (1kV/40) pulse. The key is SMB 600W 18V transil (some manufacturers specify it as 18V, other as 15V depending if they use in name working or break voltage.

My protection is good against Surge pulse but not against mistaken connection of, for example, 24V power supply. So as always everything depends...
You have to start from specifying against what you are designing your protection.

Your input certainly will not pass 1kV 25A test as LM311 will get 1kV supply voltage.

[ldkimball]

I opted to go with the 311 because the voltages in this part of the circuit can sometimes swing negative and my impression of the 393 is that it can't handle negative inputs. \

I'm not worried about surge protection in this case.

[Kim Christensen]

If you connect the LM393's GND pin (4) to -6V then it'll work the same way as you've got the LM311 wired. (You could increase R45 to minimize loading on the -6V rail when Q3 is being turned off)
The 2n7002 will get a negative voltage on it's gate when the comparator output is low, but that's OK and the LM311 circuit does the same thing in your schematic.

[Konkedout]

I recommend you have a look at the TL431.  They are amazingly cheap, common and useful.  Once upon a time (and maybe still?) almost every switcher power adapter had a TL431 to control the output voltage feedback.  I don't think I am exaggerating to say they have been used in the billions.  I think of how many people might have several power adapters (each containing a TL431) in their house.

If doing a window comparison, one reasonable way could be to use two TL431s, each as reference + scaling + gain.  Another reasonable way could be to use one TL431 as a reference and one LM393 dual comparator.

I think of them as a bit like an NPN transistor with a precise Vbe of 2V495.

[Konkedout]

I opted to go with the 311 because the voltages in this part of the circuit can sometimes swing negative and my impression of the 393 is that it can't handle negative inputs. \

I'm not worried about surge protection in this case.

The way you have the LM311 connected, I do not think you will get any advantage with the LM311.  In either case, powering from +/- 6V, the output will swing to -6V.

You can connect the LM311 emitter output to ground, and then the output will only go down to ground.   But driving those MOSFETs should not be a problem either way.

[fzabkar]

You might like to consider this IC (P14C5N):

https://www.hddoracle.com/viewtopic.php?p=24492#p24492
https://www.prisemi.com/pdf/2ovpgybhqx/P14C5N.pdf

[Solder_Junkie]

A Zener and SCR will blow a fuse and protect components from an over Voltage situation. This is fitted in many older power supplies and I have used the approach successfully for several years.

More recently I have built a number of LTC4368 based protection boards, the LTC4368 is a purpose designed protection IC. The datasheet is here:
https://www.analog.com/media/en/technical-documentation/data-sheets/ltc4368.pdf

I used DMT6002 MOSFETs and added a 3.9K resistor from the output of the MOSFETs to pin 9 and added a 220nF cap between pins 8 and 9. This avoids a switch-on surge from equipment connected to the power supply tripping the protection, it will still protect against over current in around 1mS. You cannot blow a "fast blow" 3A fuse by shorting out the PSU terminals when using a 1.5 milli Ohm current sense resistor (33 Amp current).

The downside to the LTC4368 is the size of the IC... It is tiny! Mouser stock them and the DMT6002 N Channel MOSFETs.

SJ

[fzabkar]

A Zener and SCR will blow a fuse and protect components from an over Voltage situation. This is fitted in many older power supplies and I have used the approach successfully for several years.

This old hard drive PCB has such a protection circuit on its 5V supply:

https://web.archive.org/web/20230729181139/http://www.users.on.net/~fzabkar/HDD/Tutorial_SP0411N.html