I want to protect a board I'm working on from reverse supply polarity. I also want to be able to fuse some 12V power devices on my board in the case of a short or whatnot - however I want my MCU itself to remain active, if the 12V lines are fused. So - I've come up with the following, which includes reverse protection for both 5V and 12V circuitry. I also want 5V sensors to have a short protected supply, so - instead of using another regulator, I thought of using the following.
By rating F1 at half of F2, F1 can fuse to shut off any shorted sensors, whilst the MCU remains active. I will think of a way to add some circuitry in to signal to the MCU when F1 fuses. Therefore, 5V/1 is now a protected sensor supply, whilst 5V is now the 5V for the MCU.
Can anybody see any problems with this, or is there a way I can reduce component count more importantly?
I'd suggest simply fitting a bridge rectifier assembly in your power feed, then it doesn't matter which way round your power feed is and no intentional short-circuits (via diodes) trying to act as safety devices. Reverse polarity destruction is likely to happen far quicker than a fuse can blow.
An interesting point - but I would then lose the 12V supply to the board would I not due to the 2x diode drop. Not a problem into the 5v regulator clearly, but for actual 12V parts of the application, I'd no longer have 12V?
Use a full bridge to feed the regulator and everything else use a mosfet, just look up mosfet reverse polarity protection . This way the 5v will always be live but the unaffected 12v will only workk the right way and no damage in reverse
So, I've moved away from the idea of a bridge - and found a great document on the IRF site about protecting circuits with MOSFETs as recommended.
Now, I've ended up with the following;
Maybe I worded it wrong before - but I don't want my MCU to be powered if there is a reverse polarity, I only wanted it to remain active if the sensor lines or 12v powers lines fused, I believe the above does that. The P channel FET protects against reverse voltage, whilst F1 & F2 will fuse if overcurrent events occur on those lines, whilst the MCU will still remain active. I could indeed add a third fuse to kill power to the MCU if that itself got shorted.
I'd still like a reverse indicating LED, but it's not the biggest issue.
I also want protection against a load dump event, so I've added a zener across the power lines that should avalanche at whatever voltage I choose, and clamp load dump spikes to ground.
You might look at using a Transzorb for D4, as they have huge wattage ratings,
plus I believe you have the p-MOSFET in there backwards. Notice how the built-in
diode is defeating the purpose of the MOSFET.
Thanks for this, just taken a look at TVS diodes as you suggested - they seem to be exactly what I want.
Regarding the FET polarity - I'm not sure about what you mean? It seems in order to me. It is also installed as per the instructions of the application note from IRF that I posted.
In it's current configuration, the diode begins to conduct when polarity is correct, then the gate is driven to open the FET, thus getting rid of the diode drop and leaving a drop of only Rds * Ids. If the polarity were reversed, the diode would just block current, and no current flows.
If you are using D4 as either over-voltage or reverse voltage protection you should have a fuse in the feed line from your battery to provide cable protection (current will be rather high !!)
jackrae:
If you are using D4 as either over-voltage or reverse voltage protection you should have a fuse in the feed line from your battery to provide cable protection (current will be rather high !!)
D4 is there to provide load dump protection, by clamping high voltages to ground. I don't see that there needs to be a fuse now, as if the polarity is wrong - no current will flow as it is blocked by the body diode of the FET, and the FET will not switch on.
dhenry:
Both work as a switch.
The IRF approach provides protection against reverse polarity. The other approach will just fry your load with the wrong polarity.
Sorry, I'm not with you - "the other approach". The FET approach in the video is the same as the IRF note is it not?
jackrae:
If you are using D4 as either over-voltage or reverse voltage protection you should have a fuse in the feed line from your battery to provide cable protection (current will be rather high !!)
D4 is there to provide load dump protection, by clamping high voltages to ground. But ah, yes - I see, I need a fuse. EDIT -- I have just done a little more research, no fuse is needed here. I can use a bidirectional TVS diode setup as depicted below. Here's a good read on it - http://www.protekdevices.com/Assets/Documents/Technical_Articles/ta1003.pdf
dhenry:
Both work as a switch.
The IRF approach provides protection against reverse polarity. The other approach will just fry your load with the wrong polarity.
Sorry, I'm not with you - "the other approach". The FET approach in the video is the same as the IRF note is it not?
is there a way I can reduce component count more importantly?
I hate to said it after all this redesign, but I think you're getting somewhat carried away.
Short of simply using a series diode in the Vin line, which I assume you abandoned from
the get-go, I'd go back to your very first ckt, and replace D4 with a unidirectional
transzorb.
The latest design includes as extra - just one FET. The one extra cap I don't think counts, as it should have been there anyway. The current design has only one TVS diode too, as well as losing a fuse from the list. Total cost is lower than before.
