No, those two diodes go to the two power rails of the MPU. This ensures that the output pin voltage cannot get more than one diode drop higher or lower than the range between the power rails. Any voltage outside that range is then conducted by one diode or the other to the power rail itself, and away from the MPU output.

Placing a zener on the pin to protect (as in #10) is commonly done for input pins and I don't see anything wrong with it here. As long as the output is normally below the zener voltage, I don't see the zener doing any harm and it does offer protection.

I do not understand the objection to #10 from OBW0549. The zener in #12 would pop like a fuse if any significant voltage came along to challenge it. There's nothing to limit the current. Arrange as in #11, the 1kΩ resistor limits the current that the zener needs to sink to protect the MPU.

I assume in the last sentence, you meant #10 not #11. Just making sure. If I were to use that(#10 drawing), what zener value should I use?

Boy, these things are difficult

 

I assume in the last sentence, you meant #10 not #11. Just making sure.

Yes, sorry, #10. I'd like to hear from OBW0549 as to why he didn't like that arrangement before I recommend it without reservation, though.

The zener voltage should be just high enough that it never conducts any current if thing are operating normally. So if the MPU can put a regulated 5.0V on the output, a zener at, say 5.2V should be adequate.

 

Yes, sorry, #10. I'd like to hear from OBW0549 as to why he didn't like that arrangement before I recommend it without reservation, though.

The zener voltage should be just high enough that it never conducts any current if thing are operating normally. So if the MPU can put a regulated 5.0V on the output, a zener at, say 5.2V should be adequate.

Thanks!
How about these:
http://www.ebay.com/itm/100pcs-1-2W...878160?hash=item43e05b23d0:g:A90AAOSw9N1VuX-I

They are 1/2W 5.1V

So, what happens if the MOSFET is unlucky, and 50V comes out through the gate (or by butter fingers)? What does the zener
actually 'do'?

Thanks,
Nicholas

 

I do not understand the objection to #10 from OBW0549.

My objection is that the amount of protection this method provides depends on the relationship between Vz and Vdd; if Vz is too high, the μC's pin protection diode will conduct first, dumping most or all of the fault current into the Vdd bus and possibly blowing the pin's bond wire or the chip's metallization. OTOH, if Vz is too low, it will conduct during normal operation when the output pin goes high, causing unnecessary current draw. And the necessary balancing act between Vz too high and Vz too low is complicated by the Zener's voltage tolerance as well as its Vz temperature coefficient.

The zener in #12 would pop like a fuse if any significant voltage came along to challenge it. There's nothing to limit the current.

If anything external somehow manages to dump enough energy into this circuit to pop a 1-watt Zener diode, that voltage/current first has to punch through the MOSFET's gate-to-channel oxide, destroying the MOSFET. And with that much impulse power, the MOSFET is going to blow whether the Zener is there or not.

Bottom line, any of these protection schemes will work satisfactorily for the TS's purposes, and I'm not inclined to argue about which one is "best." I've used the Zener diode scheme as drawn in post #12 many times in the past and it works very well. (I'd also note that the Zener in that scheme helps protect the MOSFET gate from overvoltage.)

 

Yes, sorry, #10. I'd like to hear from OBW0549 as to why he didn't like that arrangement before I recommend it without reservation, though.

See my previous post.

The zener voltage should be just high enough that it never conducts any current if thing are operating normally. So if the MPU can put a regulated 5.0V on the output, a zener at, say 5.2V should be adequate.

Remember the tolerance on Vz, usually +/- 5%. 5.2V is cutting it awful close IMO.

 

See my previous post.

The zener voltage should be just high enough that it never conducts any current if thing are operating normally. So if the MPU can put a regulated 5.0V on the output, a zener at, say 5.2V should be adequate.[/QUOTE]
Remember the tolerance on Vz, usually +/- 5%. 5.2V is cutting it awful close IMO.[/QUOTE]

I think the zener approach is the way to go, it only uses one component and less space on a pcb. If 5.2V is cutting it close, then
my link to a 5.1V variant is out. Also, should I use 1/2W or 1W?

If I seem dense, please keep in mind that my background is in biotech/microbiology

 

Boy, these things are difficult

No, they're not. It's just that at this point, you are SERIOUSLY over-thinking this thing.

Quit agonizing over this: just use a 1N4735A Zener as drawn in post #12, or use Picbuster's Schottky diode clamp scheme that he described in post #16. Either one will work just fine.

 

Hi guys

Thus far I have used a 1n1001 normal diode in series from my controller (5V) to my MOSFET switching something like 30-50V.
I want to use a Schottky diode to get a smaller voltage drop from the controller, but I would like a recommendation as to which one.

The Schottky should protect the micro controller in case something goes wrong, and 50V is going out of the gate of the mosfets and
try to fry the micro controller. I have been adviced that this is very rare, but I want to be on the safe side

So, what Schottky should I use?

Thanks!

A Shottky diode will give you lower volt drop, but the reverse volt rating starts at only 20V and anything over 60V gets pricey.

There is a trick for using a MOSFET for reverse polarity protection. Basically; you connect the drain and source in series with the supply so that the body diode would normally conduct - then you bias the gate to enhance the channel. Some modern MOSFETs have RDSon of only a few milli-Ohms.

I don't remember the exact details - but its been incorporated in numerous magazine projects, frequently in EPE magazine who also published a page on how it works.

 

Isn't this situation one of the things a mosfet gate driver does? For some reason no one wants to use gate drivers. They are fairly reasonable in price and give isolation between the input and the gate of the mosfet. And will allow full gate voltage mosfets instead of logic level. And most use up less board real estate than discreet device methods.

 

Isn't this situation one of the things a mosfet gate driver does? For some reason no one wants to use gate drivers. They are fairly reasonable in price and give isolation between the input and the gate of the mosfet. And will allow full gate voltage mosfets instead of logic level. And most use up less board real estate than discreet device methods.

Good catch. I've been using these lately.

 

Only reason I can see for not using a gate driver is if it is a school project and your not allowed. But many people way smarter than me don't seem to even consider them? Guess it pays to be dumb sometimes.

 

Only reason I can see for not using a gate driver is if it is a school project and your not allowed. But many people way smarter than me don't seem to even consider them? Guess it pays to be dumb sometimes.

They're good for other things, too. They make good logic-level translators (within certain voltage limits) that are very fast and have high fanout; their outputs swing extremely close to the supply rails under light and moderate loads in case you need that functionality in an analog circuit; and they're handy as high-current LED drivers. Late last year in the "LED 'afterglow' time?" thread I used an MCP1406 in an LED test circuit for measuring the turn on/turn off time of a 3-watt power LED, driving it at 900 mA peak current with <20 ns rise/fall time. Worked like a charm.

 

The zener voltage should be just high enough that it never conducts any current if thing are operating normally. So if the MPU can put a regulated 5.0V on the output, a zener at, say 5.2V should be adequate.
Remember the tolerance on Vz, usually +/- 5%. 5.2V is cutting it awful close

I think the zener approach is the way to go, it only uses one component and less space on a pcb. If 5.2V is cutting it close, then
my link to a 5.1V variant is out. Also, should I use 1/2W or 1W?

If I seem dense, please keep in mind that my background is in biotech/microbiology

Sorry to confuse you. Maybe we should start a poll.
I like this one. http://forum.allaboutcircuits.com/threads/what-schottky-diode-for-protection.123653/#post-994517
It protects the micro from the 50 volts and protects from unplugging the micro from the FET.[/QUOTE]

 

Maybe we should start a poll.
I like this one. http://forum.allaboutcircuits.com/threads/what-schottky-diode-for-protection.123653/#post-994517
It protects the micro from the 50 volts and protects from unplugging the micro from the FET.

It depends. If the MOSFET's load is a big, beefy solenoid and the TS wants to make it go "clunk-clunk" a few times a second, then I vote for my own solution. OTOH, if it's an inductor whose current is being controlled via PWM at a 100 kHz+ rate, then I vote for shortbus's idea-- it's much better for turning the MOSFET on and off fast to minimize power dissipation.

 

It depends. If the MOSFET's load is a big, beefy solenoid and the TS wants to make it go "clunk-clunk" a few times a second, then I vote for my own solution. OTOH, if it's an inductor whose current is being controlled via PWM at a 100 kHz+ rate, then I vote for shortbus's idea.

With the zener on the FET side it will blow if the FET shorts and the micro will still see the 50 volts.
If you put it on the micro side it will probably work okay depending on the zener voltage and tolerance. To high and it doesn't protect, to low and it's on during normal switching.

 

With the zener on the FET side it will blow if the FET shorts and the micro will still see the 50 volts.

Yes, that's been said several times in this thread already, and as far as it goes, it's true. But is protecting the microcontroller from fault currents/voltages due to a blown MOSFET what this is really about? Depending on the particular microcontroller and FET used, it's quite possible that the FET could be just as expensive a part as the microcontroller-- or even more expensive. And if either one is blown, this circuit is going to be discarded, anyway.

Rather than worrying about heavy fault currents or high voltages blowing the μC away, I'm more concerned about glitches and/or noise from the outside world getting coupled in through the MOSFET's gate capacitance and making its way back into the μC and causing spurious resets or other erratic operation. Therefore my choice would be putting the Zener at the MOSFET gate and nipping such transients in the bud. That, or use Picbuster's proposed arrangement.

If you put it on the micro side it will probably work okay depending on the zener voltage and tolerance. To high and it doesn't protect, to low and it's on during normal switching.

True. And the Zener's ±5% voltage tolerance, all by itself, is enough to take you all the way from "too low" to "too high." IMO, having that degree of uncertainty would be unacceptable.

But that's just my opinion. YMMV, of course...

 

See