1
\$\begingroup\$

If, for example, a Silicon diode rated at 100V is reverse biased at 200V (long term, not just very briefly) but in series with a few MegaOhm resistor such that the reverse current is limited to one microAmp (or a few uA), would that change the characteristic of the diode. I know, from experience, that the diode won't become faulty but, rather, I am trying to inquire whether this is an acceptable practice. Presumably the diode works in Avalanche region but can it recover without any issues?

This question can also be applied to other pn junction devices e.g. transistors (C-E rating exceeded but current limited to a microAmp or two via MegaOhm resistor). 31May'18: WhatRoughBeast, Elliot Alderson and BimpelRekkie have all really made me think harder. Thanks all. I think you all have raised useful points. This question is not a theoretical one since I actually noticed a somewhat similar scenario in one of the circuits used successfully over a number of years, the series (collector) resistor being 470k but the device in question was an NPN transistor (BC547B) rated at 45V, Common-Emitter config. The positive rail could surge well beyond the 45V. Sorry, I can't post the full circuit due to strict NDA with various companies I worked for. A diode (pn junction), once it reaches its breakdown voltage level cannot "allow" voltage across it to rise further (the I-V graph slope is very steep beyond the Breakdown voltage) so the series R will keep absorbing any further rise in supply voltage (like in a normal Zener diode). So, basically, my question effectively boils down to: "Can a diode (or pn junction in Transistors etc) be damaged, or deteriorate in its standard characteristics, if it sees Breakdown voltage repetitively over long term, keeping in mind it doesn't venture far into the Breakdown region, only slightly i.e. a few uA. If power dissipation is the only consideration then the answer is "No, it won't get damaged", but I had been suspecting a possible problem elsewhere, hence my question. BTW, other external considerations, such as Safety, Creepage/Clearances can be assumed to cope with higher voltage surge.

5 June'18: I have given it further thought and I think the answer also depends upon how rapidly the reverse voltage is applied. For a "very fast" rising (reverse) voltage, when the current is initially zero, all of the voltage will appear across the pn junction. This raises alarm since there is (internal) di-electric breakdown consideration too (perhaps, at the back of my mind, this is what I had been worrying about). In practice, it may not be a problem since such adverse conditions are not so realistic (?). Anyhow, if anyone of you does manage to carry out relevant tests, perhaps on a Fast Recovery and on a slow recovery diode, then please do enlighten us too. If I ever get an opportunity and time, I'll endeavor to do likewise...keeping in mind, a failure will provide us with a conclusion but a "pass" won't (necessarily....since it's just a one-off test!)

\$\endgroup\$
  • \$\begingroup\$ Would you actually be able to maintain 200V across the diode given that leakage current? \$\endgroup\$ – pjc50 May 30 '18 at 9:39
  • 1
    \$\begingroup\$ Zener diodes don't have any worries about this happening. \$\endgroup\$ – Andy aka May 30 '18 at 9:49
  • \$\begingroup\$ @pjc50: Good point. Initially, the diode will see the full voltage, then depending upon how many uA it decides to pass through, there will be a V drop across the MegOhms series R. For the sake of argument, let's assume the actual Breakdown threshold is just above 100V and the diode continues to see volts beyond its rating...but only uA of current. \$\endgroup\$ – Naseer Ahmed May 30 '18 at 13:19
  • \$\begingroup\$ Some diodes may be rated for this. Some may not. It would be best to ask the manufacturer about it. As you say, in practice, a couple of uA of reverse current will usually not damage a diode. Also, it would be a better question with a specific part number and datasheet. As is, the answer is "maybe" or "it depends." \$\endgroup\$ – mkeith May 30 '18 at 16:26
  • \$\begingroup\$ If you maintain constant current, then the voltage will vary somewhat with temperature. Maybe you should explain why you are asking about this. What is your ultimate purpose here? \$\endgroup\$ – mkeith May 30 '18 at 16:45
1
\$\begingroup\$

Can..?

Yes you can, as long as no damage occurs. If you limit the current and voltage such that then no damage (to the component) can occur then it won't.

It depends on the power rating of the component how much it can handle. For example a 1 A diode like the 1N4001 has a forward voltage (at 1 A) of 1.1 V so that gives a total of 1.1 W dissipation. In reverse at 200 V at 1 mA that would result in 200 mW dissipation, far below that 1.1 W. So you do not have to limit the current to a few uA, one mA might be safe as well.

A 100 V rated diode is not guaranteed to show breakdown at for example 200 V. It can be at 101 V or 1000 V. The manufacturer only guarantees that breakdown occurs above 100 V.

Basically almost all diodes can be used as a zener diode however a zener diode will exhibit breakdown at a specified voltage. Using an NPN's BE junction as a "poor man's zener" is indeed an option. You just should not rely on an accurate breakdown voltage.

\$\endgroup\$
  • 1
    \$\begingroup\$ Your second sentence is the best answer possible. \$\endgroup\$ – Elliot Alderson May 31 '18 at 1:09
0
\$\begingroup\$

In general, you should not operate any electronic component outside the envelope of maximum values specified by the manufacturer. It is true that you can get away with this in some cases but you shouldn't generalize to all semiconductor devices.

For example, you asked if the same rule would apply to a transistor's CE voltage, that you could apply a very high voltage as long as current and power were limited. Suppose you tried to apply 1000V, limited to 1 uA, to the collector of a small 2N3904 transistor in a TO92 package. Would it survive for any length of time? Would it be safe? I wouldn't bet my own money on it. My point is that there may be characteristics of a device other than PN junction leakage that limit its maximum voltage.

\$\endgroup\$
  • \$\begingroup\$ For a current limited case, the voltage will be that same no matter what "nominal" voltage is applied. The current source will absorb the difference in voltage between the nominal voltage and the diode. In effect, the diode will (sort of) act like a zener, but with a different breakdown mechanism. \$\endgroup\$ – WhatRoughBeast May 30 '18 at 13:13
  • \$\begingroup\$ @WhatRoughBeast Yes, but my point is that the failure mechanisms in a real, packaged device might include factors other than reverse leakage current or power dissipation. Even if current and power are limited you can't assume that any arbitrary high voltage can be safely applied to a device. The OP asked about exceeding the maximum rated CE voltage, not about some "nominal" voltage. \$\endgroup\$ – Elliot Alderson May 30 '18 at 14:14
  • \$\begingroup\$ You have missed the point entirely. If the current is limited, the voltage applied cannot be greater than the breakdown voltage at the limit current. Let's say (just for argument) that a diode will have a reverse voltage of 400 volts when 1 uA is propagating through it. ANY voltage higher than 400 volts will then produce a current greater than 1 uA, and the condition of limited current as specified in the OP does not apply. You can't have it both ways - if the current is limited this will fix the voltage actually appearing on the diode. \$\endgroup\$ – WhatRoughBeast May 30 '18 at 23:08
  • \$\begingroup\$ @WhatRoughBeast No, I understand your point. However, I disagree with your interpretation of the OP's question. I understand that the current is limited. The question, as I understand it, is whether or not the device can be operated reliably if the reverse junction voltage exceeds its specified maximum value (even if the maximum power limit is not exceeded). I contend that this is not a reliable situation, in general. \$\endgroup\$ – Elliot Alderson May 31 '18 at 1:07
  • \$\begingroup\$ Please see my original post - I have now edited it in light of some useful comments by all of you gentlemen. \$\endgroup\$ – Naseer Ahmed May 31 '18 at 7:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.