Overkill reverse polarity, AC and over-voltage protection.

[Phoenix]

Hi everyone.

I've been toying with a circuit for reverse polarity, AC and over-voltage protection. It's based partly on RG's excellent article on Advanced Power Switching.
I've also included in this circuit a TC1044 charge pump in voltage inverter configuration for bipolar rails. Note that in this circuit I'm using surface mount multi-layer ceramic capacitors which are not polarized, if using polarized electrolytic caps, make sure to refer to the TC1044 datasheet for the correct polarity orientation of your caps.
In order to still use the ring-to-sleeve connection (labelled sleeve on the schematic) of the input jack for power switching without introducing ground hum, the PNP circuit is added so that the current pulses drawn by the TC1044 do not go through the audio ground, only the current required to turn on the PNP.
The TC1044 has a maximum input voltage rating of 12V, so I've added an 11V zener diode and polyfuse (PPTC resettable fuse) which trips at 150mA (which is well over the max normal current draw, but was also the lowest value I could find), but should be quickly reached in the case of >11V, thereby protecting the circuit from overvoltage conditions. Once the pedal is disconnected from the over-voltage and allowed to cool down for a couple of minutes it will behave normally again.

Just wondering if anyone has any thoughts? I know it's complete overkill, but I like things to be rock-solid and as close to impossible to kill as is reasonably possible.

[merlinb]

There's engineering and there's overengineering... Why not just use a Schottky diode in series with the incoming supply? Job done, with less than 0.25V forward drop. Do you really care about quarter of a volt?

[slacker]

I don't think you need the mosfet, the polyfuse and zener should already protect against reverse polarity and AC. The only time the mosfet will have anything to do is if you manage to blow up the zener.

[anotherjim]

Zener may well blow before the fuse does. And if it has blown open-circuit, but everything else has survived, you will never know it's blown, so it might as well not exist.
I like the series Schottky protection too, but I think those diodes are only low Vf at low current. They can do this...

[merlinb]

I like the series Schottky protection too, but I think those diodes are only low Vf at low current.

Correct, but low current is all we need. Try the SR150x series:

[R.G.]

There's engineering and there's overengineering... Why not just use a Schottky diode in series with the incoming supply? Job done, with less than 0.25V forward drop. Do you really care about quarter of a volt?

Schottky and Ge diodes get it into the realm of OK-ish for most situations, but folks who like batteries often find that a quarter of a volt does matter to them. Getting under 100mV, and under 50mV, can be important.

Besides, with the absolute frenzy of - um, fecal-matter polishing in the pedal world, who's to say that anything is too much. Just ask a modder. 

[Phoenix]

There's engineering and there's overengineering... Why not just use a Schottky diode in series with the incoming supply? Job done, with less than 0.25V forward drop. Do you really care about quarter of a volt?

Yep, overengineering at its finest. I don't really care about 0.25V, but I can lower that, and in this instance I feel like doing it, well, just because I can.

I don't think you need the mosfet, the polyfuse and zener should already protect against reverse polarity and AC. The only time the mosfet will have anything to do is if you manage to blow up the zener.

The TC1044 datasheet recommends never applying anything less-than-ground to an input or risk destructive latch-up. I don't know if -0.7V will result in this and haven't tested as such, but even if it doesn't, I'm not the manufacturer and can't guarantee that it will never happen, so with the pmos I'm heeding the manufacturers advice and making sure that the charge pump will never be reverse-biased, even a bit.

Zener may well blow before the fuse does. And if it has blown open-circuit, but everything else has survived, you will never know it's blown, so it might as well not exist.

The polyfuse I've chosen should blow in just under 1 seconds at 200mA (which would be reached very quickly), so any overload should only be relatively momentary. Even with a half-watt zener, that's only a x4 overload for 1 second which should be well-and-truly survivable, even under bright stage lights or outdoors in a black enclosure on a hot day. I could of course use a TVS diode instead which are really just zeners designed for that type of overload and are better documented for those conditions, but I think a plain zener will be just fine, or use a 5W zener, I have plenty of those lying around.
Anyway, if someone is connecting a power supply of too high a voltage, it's most likely to be 12V or 18V, so the polyfuse will trip much faster.

I like the series Schottky protection too, but I think those diodes are only low Vf at low current. They can do this...

Yes, but as Merlin mentions above, you just need to select the appropriate schottky. We're only dealing with low currents here.

Schottky and Ge diodes get it into the realm of OK-ish for most situations, but folks who like batteries often find that a quarter of a volt does matter to them. Getting under 100mV, and under 50mV, can be important.

Besides, with the absolute frenzy of - um, fecal-matter polishing in the pedal world, who's to say that anything is too much. Just ask a modder. 

Thanks RG! I think you get the spirit of this post. I set some entirely arbitrary design goals, and this is the topology that allows me to achieve it. It only matters to me.

Keep the comments coming, I'm enjoying the discussion.

[slacker]

The TC1044 datasheet recommends never applying anything less-than-ground to an input or risk destructive latch-up. I don't know if -0.7V will result in this and haven't tested as such,

Fair enough, I hadn't thought of that, maybe move the zener to after the mosfet then, that way reverse polarity or AC protection are handled by the mosfet, which to me seems a nicer way of doing it. Obviously your mosfet then has to handle the current in over voltage so it might be swings and roundabouts.

[Phoenix]

Fair enough, I hadn't thought of that, maybe move the zener to after the mosfet then, that way reverse polarity or AC protection are handled by the mosfet, which to me seems a nicer way of doing it. Obviously your mosfet then has to handle the current in over voltage so it might be swings and roundabouts.

Placing the zener after the mosfet would add the mosfet impedance to that of the zener impedance, so would slow down the trip time of the polyfuse, so I think that would be a mistake. The polyfuse should trip in under 1 second at 200mA, which is only 2.2W at 11V, so I could always just use a 5W zener which will be more than beefy enough. Of course, at those currents the zener voltage will also be higher than the nominal 11V, even approaching 12V, so I could also add a second 1/2W 11V zener following the mosfet to ensure that the TC1044 never sees more than 12V under any circumstances like slow trip time on polyfuse. How's that for belts and braces/suspenders and three pairs of underwear over-engineering?