Exact match. Not showing close matches.
'[EE] Zerners as fail-short protection'
I would be interested in knowing if anyone is aware of a formal
specification for fail-Short-Circuit behaviour for zener diodes or any
discussion on the subject.
Use of zener diodes as "fail-short" protection on power supply rails
If the rail voltage exceeds the zener voltage adequately the zener
overheats and fails short-circuited, thereby clamping the rail to near
ground and blowing the fuse which you had better have provided. The
intention is to protect equipment which may otherwise be damaged by
severe fault over voltages.
I have used 6 Watt SOD-64 glass bead Philips BZW03Cxx zener diodes for
this purpose for many years. These notionally will (in the right
circuit) clamp excess voltages in the sub-nanosecond timeframe. My
experience is that when used on a supply with a nominal 6 amps
capability these zeners will reliably fail short-circuit. On every
occasion when I have seen these fail they invariably go SC (short
circuit) and never OC (open circuit) .
The range of conditions under which a zener will fail SC and not OC
and, very importantly, the range of subsequent currents over which a
failed zener will REMAIN SC does not seem to be part of any
specification I have ever seen.
I note that much smaller zeners (eg 500 mW) also seem to always fail
SC under realistic provocation. It is not obvious what certainty I can
have that this will always be the case for any given circuit.
I would be interested in knowing if anyone is aware of a formal
specification for fail SC behaviour or any discussion on the subject.
Google knows about lots of people who mention the phenomena but
doesn't seem to know about anyone who discusses it usefully.
The sort of thing that would be useful is knowing that you can take a
given Vvolt Wmilliwatt zener of a certain construction and that if you
destroy it with less that Iamps/Pjoules or whatever (mainly a complex
function thereof) then you can expect it to go SC rather than OC with
a given degree of certainty.
That there are upper limits is demonstrated by considering a 5V1 500
mW zener placed across a 12V car battery. The zener may go SC
initially, but not for long. The fused junction will have a different
current carrying capacity than the leads and any interconnects.
Something will fail and rapidly. In a few informal tests of 500 mW
glass zeners across a 30V 5A lab supply failure was always SC with the
supply then sitting on the internal current limiter. I haven't pushed
this to see how repeatable it is or what the upper limits of reliable
In my target application I have 1360 uF at 150V behind a FET which may
itself fail SC and connect this supply to a 16V rail. The 3W zener
does a good job of stopping any damage if this happens. 500 mW zeners
are much smaller and cheaper, but .... .
I believe any component can breakdown once a certain
point is reached. In this case a zener can burn up
after the maximum temperature is reached.
Look at maximum junction temperature. 175 c. After
this temperature well........ I haven't seen a lab
test on this and can't fully backup my claim.
--- Russell McMahon <paradise.net.nz> wrote: apptech
On 3/18/06, Russell McMahon <paradise.net.nz> wrote: apptech
> I would be interested in knowing if anyone is aware of a formal
> specification for fail-Short-Circuit behaviour for zener diodes or any
> discussion on the subject.
I guess my experience is a bit different. Anyway, it is just for your reference.
SCR/zeners are often used for over-voltage protection. For zener barriers,
two or three zeners in parallel forms the over-voltage protection. A fuse will
prevent over-current. A resistor will protect the fuse.
For EN50020 EEx ia applications, I remember I did a lot of temperature
measurement to assure the certification body that the zener will reliably
clamping at the specified voltage and not fail open circuit. We need to
assume that the current will be 170% of the fuse current (fuse needs to
be certified, only a few vendors have the EEx certification), then the
zener junction temperature needs to be within the specified temperature
range and the power rating needs to be within 2/3 of maximum allowed
power rating and the PCB underneath the zener needs to be within the
specifications as well.
For higher power application, often SCR crowbar circuit is necessary for
over voltage protection since zener will not be good enough.
In all these cases, the fuse needs to blow if the SCR or zener conducts.
The certification body also wants to know the construction of the zener/SCR
and fuse as well as the resistor.
Anyway, Intrinsic Safety (IS) applications are in general too stringent. Still
this may give you some hint.
In your case, if you have a fuse with proper rating before the 5V1 zener,
I think the zener will be okay. If you have 5A before the zener, I think the
vendor will not guarantee that the zener will be okay since the power will
be too much for the zener.
On Sun, 19 Mar 2006, Russell McMahon wrote:
> I would be interested in knowing if anyone is aware of a formal specification
> for fail-Short-Circuit behaviour for zener diodes or any discussion on the
There is a discussion on this in some semiconductor books. In my
experience, all stud-mounted semiconductors fail short on overload (if
they do not blow up). Most end up as resistors between 0.5 and 50 ohms
after 'finished'. Semiconductors mounted with bonding wires tend to fail
open. This is most TO3 cases f.ex (argh) and nearly all ICs excepting
flip chip, with which I have little experience.
> Semiconductors mounted with bonding wires tend to fail
> open. This is most TO3 cases f.ex (argh) and nearly all ICs
> flip chip, with which I have little experience.
I have a good "success rate" of getting FETs (TO220, TO247) to fail
either GDS all shorted together (quite common) of DS short (less
common). In my applications failing OC is less common.
On Mon, 20 Mar 2006, Russell McMahon wrote:
>> Semiconductors mounted with bonding wires tend to fail
>> open. This is most TO3 cases f.ex (argh) and nearly all ICs excepting
>> flip chip, with which I have little experience.
> I have a good "success rate" of getting FETs (TO220, TO247) to fail either
> GDS all shorted together (quite common) of DS short (less common). In my
> applications failing OC is less common.
The FETs in TO220 and TO247 do not use bonding wires they use a piece of
thick wire. In my experience the die goes short, and stays short, and
then the bonding wire explodes. If it does not explode, then the device
stays shorted. This seems to be true over many many devices (say about
13 years worth of devices). This means that any case that can take the
short current will likely stay short.
Russell McMahon wrote:
> Use of zener diodes as "fail-short" protection on power supply rails
> is common.
"Protection" seems to be a safety feature. I'm not sure it's a good idea to
use a component way outside of spec'ed behavior (after failing) for a
Not what you asked for, of course, but maybe relevant. And in the lines of
"never put more than Vdd on a PIC IO pin, not even through a resistor" :)
Alan B. Pearce
> Use of zener diodes as "fail-short" protection
> on power supply rails is common.
Normally only as a clamp on the voltage going out of spec, not as a device
that goes short circuit though.
The common way of making a short circuit is to have the zener feed the gate
of a suitably beefy triac. I remember Power One supplies all having this on
5V supplies, would crowbar at about 5.5V which was sufficient for protecting
TTL devices which were common at the time.
More... (looser matching)
- Last day of these posts
- In 2006
, 2007 only
- New search...