Alex , i will take blown transistors for 400 ...

[petey twofinger]

i bult an amz mosfet boost using this layout ;



i used a 1n4001 diode , a 1n4001 is not a zener diode !?

anyway , the transistor , a bs 170 blew up today . i was doing some tests and maybe i powered it up and off again a few times with the main breaker on the ac .  

not sure but i am guessing if i put a different diode in the , replacing the 1n4001 this will not happen again ?

i do have some nte139a's , not sure why i bought em , maybe for this ! but i do not know how to reverse cross reference an nte number ! doh .

if this is the issue , any 9.1 zener diode recommendations , easy to get , cheep , common , find one in a common appliance ... heh .

blowin stuff up , always a thrill !       ( at least i THINK i may have figured it out )

[R.G.]

The 1N4001 is definitely not a zener diode. In this circuit, it offers no protection to the gate at all for positive-going transients.

Finding a 12V zener in random electronics is going to be hard. Here's a possible make-do until you do another order for parts: take the base-emitter junction of an ordinary NPN transistor, and hook it up with the emitter to the MOSFET gate, base to MOSFET source. Clip the collector lead off and make sure it doesn't contact anything.

The base-emitter junction of a modern NPN happens to have a breakover voltage of about 7V for most of them. This is a common side effect of modern processing for "good" NPN transistors. It's a really sloppy zener, but it will keep you from killing MOSFETs until you can get the right thing.

[Earthscum]

If you have a Radio Shack nearby, they should still carry Zeners in their minuscule selection parts bins... actually, all they carry are 5.1V and 12.1V zeners. They come in packs of 2.

[tiges_ tendres]

Can you use LED's also for this?

[R.G.]

You can use anything that will clamp the gate of the MOSFET from getting more than 15V or so from the source. That being said, by adding stuff to it, you have to watch and be sure you're not adding other things that will keep the MOSFET from working the way you think it will.

Most MOSFETs have turn-on thresholds of about 3-4V. So nothing happens til it gets more than that. After that, they add current at a rate of 100ma to 2A per volt of additional gate-source voltage. The problem with transients is that the gate insulator punches through and is permanently damaged at about 20V (more or less) between gate and source, either direction.

So using LEDs means knowing the conduction voltage of the LEDs, which is from about 1V for old-school red ones to maybe 4V for some blues. Realistically, you'll need more than about 6V to ensure LED conduction is not robbing the gate of voltage, so count on stacking at least two, perhaps a lot more LEDs. That is for using LEDs in forward conduction as the safety clamp.

LEDs have a poorly-defined reverse voltage breakdown, and the makers warn you not to do that. So it it possible to use an ordinary diode anode to source and cathode to gate on the MOSFET to clamp the gate to less than 0.8V the "wrong" way, and a stack of LEDs to clamp it from overvoltage the proper drive polarity.

But it's a whole lot simpler to find a zener or something that can be used as one.

[tiges_ tendres]

Got it, just get a Zener!

Thanks, as always!

[petey twofinger]

well that is good news .

i could always run over to rat shack , but i am thinking i SHOULD just o with the NTE139a i have already purchased

that is IF it will function properly in the circuit ...

its 9.1 v
1 w
DO-41
zener



but i am GUESSING all of this ^ is overkill , seeing that in the original schemo , mr. oorman states clearly it calls for a 9.1v zener ...



phew , i didnt feel like having to answer a question like "what are you trying to hook up" or sign up for verizon ... ( i have a rotary phone , that suits me just fine , thank you very much )

and thanks for the replys !! 

[R.G.]

The reality is that the exact zener voltage doesn't matter much in this circuit. The current is so low that the MOSFET will be running with its gate only a fraction of a volt over the threshold/cut-in voltage of 2-4V. So any zener witn a voltage from about 6V up to less than the gate breakdown voltage works OK. If you have a 9V zener, it'll be fine. An NPN base-emitter junction breaking at 6-7V will be fine. A 12V zener will be fine. The zener just prevents voltages of perhaps 20V from happening at the gate and puncturing the insulation.

The zener power rating of 1W is massive overkill - which is always good  - in this circuit, because the zener's average current is zero, and until there is a transient, it never conducts.

If you have a 9V zener already, use it. If you don't, you can use one of the alternatives.

[Earthscum]

On (IRF510), is the zener integral with the gate, or something you should still consider including in a circuit? I've been kind of confused about what the internal diode protects. In the datasheet, it seems to be just a voltage limiter/reverse polarity protection for the D-S, and doesn't do anything for the gate. Is this correct?

[amptramp]

Some people use a pair of antiparallel diodes in series with the zener just to isolate the junction capacitance of the zener.  I have seen audio filters made from inductors and zener diodes with the reverse voltage used for tuning.  The diodes are never driven into reverse conduction - just used as varicaps with a large capacitance in the thousands of pF.  Try out the zener on its own, but if you are losing treble, just put a pair of diodes looking the opposite way (that is, antiparallel with the anode of each connected to the cathode of the other) in series with the zener and the capacitance drops to the diode capacitance of a few pF.

[R.G.]

On (IRF510), is the zener integral with the gate, or something you should still consider including in a circuit? I've been kind of confused about what the internal diode protects. In the datasheet, it seems to be just a voltage limiter/reverse polarity protection for the D-S, and doesn't do anything for the gate. Is this correct?

The IRF510 datasheet strikes me as odd. There appears to not be any internal protection zener for the gate. The zener showed from source to drain would do nothing to protect the gate, as you say. The oddity is why they showed that as a zener. I believe that it's just the substrate body diode. In a sense, all diodes are either zener or avalanche diodes at some voltage; they'll all break somewhere. "Zener" diodes are actually two kinds; true Zener action at - I forget the limit, 6-8V or so - and avalanche breakdown, which happens at higher voltages, but is otherwise very similar. It may be that diodes bought for zeners are just sorted into pre-known voltage piles.

But no, the "zener" on the IRF510 doesn't look like it protects the gate. Probably not the drain, either.

[PRR]

IR's old IRF510 sheet does have typos.

Here's Vishay's newer sheet.

I think the body diode has an "avalanche" rating to catch the flyback in an unclamped switcher. I think the rating is not-high, not like the main FET ratings. I think they may be telling you (without saying so) that you should have a proper fast clamp diode kick-in before the body diode does.

Which is all moot in a low-volt audio amplifier.

I agree I do not see Gate protection. Quite odd. Even if the built circuit is not abused, I thought it was customary to have gate diode just so a dry day in the factory does not give 50% infant mortality (from static discharge in handling) and another 50% field failures (microdamage from non-fatal static discharge in handling).

Perhaps the lack of gate diode is to reduce capacitance, and at 180pFd it might matter. (Fatter MOSFETs go over 1,000pFd, so a little diode is nothing.)

> what the internal diode protects.

Common MOSFET processing ends up with a stray diode to body, effectively Source-Drain. In most applications it does no harm. In a few low-performance switching applications it may be useful. But it is slow and lossy, and usually the system designer makes another diode take the kick.

A protected-Gate MOSFET has a <20V Zener Gate to Source. The Gate is insulated by micro-thin glass (Aluminum Oxide). The thinner the better for gain. So thin that it will punch-through at roughly 40V. Once punched-through, aluminum shorts Gate to Source and you can't drive the MOSFET at all. The 40V is approximate and also 10 times bigger than what you really need to put on the Gate, so they tell you 20V is the limit. This allows for their production slack and your protection tolerances while still allowing massive overdrive if that's how you want to do it.

As R.G. says, your protection has to be more than 4V so the MOSFET can turn on, less than 20V to agree with the punch-through rating. Anywhere 5V-20V is fine. BJT B-E junction, stacks of LEDs, 5V 9V 12V Zener.

If only dealing with static discharge, the Zener power rating is irrelevant, but the peak current rating (not often cited) should be generous. You can get many-many Amps for a split-microsecond.

If plugging into loudspeaker outputs and other stupid tricks, the Zener is asked to absorb all the power available, which could be intense.

Just as important is to have "some" resistance between any external abuse and the Gate. Even 100 ohms will allow a 0.5W Zener to eat loudspeaker outputs. In guitar inputs it is not uncommon to have 10K-34K series as part of an RF filter, which with any reasonable Zener makes the input mighty hard to kill.

[POTL]

Bump

The power dissipation of a zener diode must be 1 Watt?
Can I use a lower value
For example 0.25 watts?

[R.G.]

Bump

The power dissipation of a zener diode must be 1 Watt?

No.

Can I use a lower value
For example 0.25 watts?

Yes.

[POTL]

Thanks