Design rule ?

Use approximately 1.5 to 2 times the continuous maximum current consumption rating, and if you have a lot of capacitors which generate a large turn-on surge, choose a slow-blow type.

The current through a fuse during a fault is typically in excess of 10 times its rating, and can be thousands of amperes in a short circuit, even on residential lines. Moreover, a fuse takes on the order of 10ms to clear the fault, an eternity to semiconductors: always remember, semiconductors fail in microseconds, typically 5-200us, thousands of times slower than a fuse. If it takes 20us for a transistor to fail shorted, it will take another 9980us before the fuse does anything. A fuse will never protect semiconductors*, or the circuit downstream, its only purpose is to protect the wiring upstream.

*Only exception I know of: specially made fuses, with extremely rapid fault-clearing times (on the order of.... single digit miliseconds!), can protect very robust semiconductors, like rectifiers and SCRs. These "semiconductor fuses" (so named because they protect semiconductors, they don't contain any themselves) are only available in large sizes (amperes), and only useful if the I^2t rating of the fuse is smaller than the same rating of the semiconductor.

Tim

-- Deep Friar: a very philosophical monk. Website:

I

on

t.

at

ne

ng

And you are asking under the pseudonym of the spiv from Minder?

As far as I know, the rule is simple enough. The fuse must chosen so that any plausible short circuit in the circuit it protects will blow the fuse before there's been enough energy dissipated at the short circuit to start a fire - about 30 millijoule, as in 30mW for one second, or 300mW for 100msec.

-- Bill Sloman, Nijmegen

Thanks Bill. And yes, I always loved George Cole as an actor, and particularly his characterisation of Arthur Daley in Minder. My 'corrupted' spelling of the name is my little tribute. I have used it as a nic on usenet for many years now. You won't have seen it on here before, as it's not a group that I often frequent. I am normally to be found over on sci.electronics.repair

Arfa

OK. So it is a little bit 'finger in the wind' coupled with rule of thumb and experience. That's pretty much the way I have always done it ...

All noted and accepted. Pretty much my experience, too

Arfa

Fuses aren't for protecting the following circuitry, but to protect the preceding circuitry from short circuits in the following circuitry.

An MOV firing on a spike shouldn't blow a normally rated fuse unless its an ultrafast (FF) fuse. A "quick" blow fuse takes lots of milliseconds to clear.

I have

Use a "delay" (T) fuse if there are inrush currents. Fusing is a complicated subject. Read the data supplied by Bussmann, Littelfuse, et al.

--
"For a successful technology, reality must take precedence 
over public relations, for nature cannot be fooled."
                                       (Richard Feynman)

etc. Google for "fuse selection guide".

I spend about a week arguing over the fuse ratings for one of my early designs. The problem was that if you added up the worst case situation (high temp, inrush peak current, audio full blast, maximum DC input voltage, maximum TX power output, etc), the safe value for the fuse was rather large. The most likely condition that would require blowing the fuse was shorting of the RF power output transistors, which would not have blown the fuse. The debate was settled arbitrarily on the assumption that the worst case "normal" conditions would never happen simultaneously.

You can measure inrush current. Put a 0.1ohm resistor in series with the fuse and put a differential input scope probe across the resistor. (That's where you use two scope probes and A-B settings to avoid having the ground lead on the AC line).

"Fired" as in caught fire? MOV's fry because of too much voltage, not current. Are the MOV's rated correctly? Did you do anything that might have produced an over-voltage condition? Are you sure it's an MOV and not an inrush surge protection NTC thermistor?

Assumption, the mother of all screwups. Measurement works better. Measure the inrush current, measure the applied voltage, etc.

The right way: Measure the inrush and quiescent currents. Calculate the fuse from the various online guides.

The not so right way: Find a pile of fuses in your junk box. Starting with the largest fuse, try turning it on and off a few times. Replace with a smaller value and try again. Eventually, you'll find a fuse that blows. Double the rated value of this fuse and you're done.

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann     AE6KS    831-336-2558

My friend measured the inrush and running currents, and that's how he arrived at his 200 mA fuse rating, which in theory, *should* be plenty. However, in my experience, with one or two exceptions, it's very rare to see such a low value in a mains feed on the domestic equipment that I see a lot of

No .... "Fired" as in it performed its function and momentarily dropped its resistance significantly as a result of over - voltage. And yes, I am quite sure that it is an MOV. The designer is a friend, remember ?

As stated, that was done, and thus far, with the trial units that have been built for bench and in-service evaluation, the one that I have is the first to have failed as a result of a blown fuse.

Again, that's how it was done at the design stage, but as all of us at the sharp end know, including you, Jeff, the theory is often not borne out by the practice ...

Well, yes. But that's taking what I was intending to convey, a little out of context. When I say protecting the following circuitry, I am thinking from a repair person's point of view, so for example, if say a diode in the bridge on the back side of the mains tranny in a linear supply failed s/c, I would want the fuse in the mains side to blow before the transformer melted. To do any damage to the actual mains coming in, would take a substantial overload current, so in such a case, I would consider the fuse to be more protecting the circuitry that followed it - in this example, the transformer - than the mains wiring behind it. That said, I accept that in some instances with some equipment, your premise will be the more valid one.

The fuse that he initially specced and fitted to the unit that I have for evaluation, was an F200 mA. About all it feeds is a commercially made 5 volt brick, which claims to have its own internal protection. I would have expected that brick to have been tested to the ends of the earth by its manufacturer, such that it could be guaranteed never to produce any unexpected input surges. So about all that leaves is the MOV straight across the mains, and after the fuse. I appreciate that the firing of an MOV takes place in microseconds so shouldn't allow enough energy to blow the fuse, but that then leaves me at a bit of a loss to explain the fuse failure. It has not blown so violently as to blacken the glass, but never-the-less, still with enough violence to have rendered the entire length of wire down to a few tiny blobs of metal ...

Accepted that fusing is a complex subject - hence this post for advice from anyone involved in it all the time. I currently have a T400 mA in it, which seems a much more realistic value and type, based on many years of component-level electronic servicing of both industrial and domestic equipment. I am guessing that the 5 volt brick is a fairly conventional little switcher inside, in which case there will be a fairly small filter cap across the primary-side bridge. So this will cause there to be some inrush current at least, but I would not have thought very much, worst case.

Arfa

I don't trust anyone else's numbers unless I've checked them. I don't do much design work, but for a while, I was doing a few design reviews and sanity checks for former employers. The number of calculation errors were few, but the effects were often catastrophic. It's all too easy to slip a decimal point. Problems like your undersized fuse value were usually a red flag, where following industry practices (i.e. cloning the competition) and using sane component values (0.2A is too small) are good example. If it seems wrong, it probably is wrong. If I find one mistake, there are usually more mistakes.

If you want us to calculate the proper type and value, we would need those measurements and some detail on the power circuitry.

A while back, I did some testing of a new radio for a friend and customer. It was not intended for mobile use, but since it ran on "12v", I decided to see how well it worked running on vehicle power. I blew it up. A scope showed that the vehicle had a 40V spike lasting about 50msec when starting. After the usual excuses, the input circuit was "protected" with MOV's and fuses. It went back and forth

3 times before they were convinced it was a problem and got it right. I then replace the wall wart with a bench power supply and ran the voltage up and down. The design worked just fine at 12.0 VDC but produced some amazing noises out of the speaker at below 11.0 VDC. At lower voltages, various stages crapped out. A new test was added to make sure that it would at least remain functional at 10.0 VDC, which is the minimum for what one would expect from a very dead auto battery. Back and forth again 2 times before they got that one right. I bought a competing product, which worked just fine down to 9VDC, which was sufficient inspiration to fix the problems.

I think that's what you're friend expects you to do. Try to misuse it in any way that a customer might misuse it. Try to break it without actually doing something destructive. Find the faults. For example, put it on an variac and see what it does at low and high AC voltages. If it quits working at the low end, or catches fire at the high end, you've done your friend a big favor.

True. I'm used to 2A fuses on 117V circuits for the bulk of my hardware. UK 230V would be half of that. I suggest you look at similar equipment and get a consensus.

You shouldn't make assumptions. He's asked you to solve a problem for him. If you assume that he did everything correctly, then there shouldn't be a problem. Since there obviously is a problem, something needs to change. Handing him a solution or a least a theory of why the fuse blew is what he wants. No insult or criticism is implied, so it shouldn't matter if he's a good friend or the devil personified.

So, what are you going to do? Wait for the others to fail? Is there something different about your unit? Maybe something different with your AC power? Or maybe you can ignore the whole thing and assume it was a solder splash. That's easy to check as a dead short will usually cause the exploding wire inside the fuse cartridge to blacken the glass. An MOV protection fault will not do that.

Yep. That which is most obviously correct, beyond any need of checking, is usually the problem. Welcome to the sci.electronics.design.philosphy newsgroup.

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann     AE6KS    831-336-2558

"Fred Abse"

** In the majority of cases, protecting the following circuitry is just what they do.

The context here is an AC supply fuse, which if correctly sized will protect the equipment from suffering any more damage than the original failure.

** Not all "T" fuses are alike and the size of the inrush is surge is hard to quantify as it varies with the exact moment of switch on also prior events. ** For sure.

... Phil

"Arfa Daily"

** Be nice to know the power rating of that " brick" - Arfa.

A F200mA fuse is not gonna stand a half cycle surge of even 1 amp peak more than a few times - so on the face of it is totally WRONG !!

With all AC to DC supplies, the use of a "T" or slo-blow fuse is ESSENTIAL !!!

SMPS always surge at switch on, so it is absurd to blindly assume that your one does not.

BTW:

To estimate the ( worst case) peak current at switch on, just divide 330 by the total resistance in series with the first filter cap in the 5V brick.

... Phil

"Bill Slowman"

As far as I know, the rule is simple enough. The fuse must chosen so that any plausible short circuit in the circuit it protects will blow the fuse before there's been enough energy dissipated at the short circuit to start a fire - about 30 millijoule, as in 30mW for one second, or 300mW for 100msec.

** Wot completely absurd crapology !!

Da Slow Man has totally LOST IT !!

... Phil

I hate to admit it. but Phil is right (as he often is).

What I was thinking of was intrinsically safe equipment

and even then I got the energy wrong - 30mJ is what it takes to ignite finely divided coal.

Hydrogen in air can be ignited by a lot less energy - 20uJ - and methane is not much better at 160uJ.

I should have remembered it better - when I worked at Kent Instruments in Luton (1973-1076), they were active in intrinsic safety, and the people who were most involved apparently all moved over to MTL as some point.

-- Bill Sloman, Nijmegen

You've got to remember that this is a power supply. If it doesn't go, for any reason, including niusance fuse opening, it's not doing it's job.

Using a quick blow fuse smaller than 3A in any power supply with capacitive inrush is asking for trouble. Slow blow fuses are better suited for this job. As long as it protects against fire in single-fault abnormals, it's doing it's job.

You haven't indicated the type of fuse involved, or the MOV marking. It's a fairly common mistake to use incorrect MOV's due to different mfr labeling practices and underestimation of normal line peak variation.

RL

It's best to test it with a large number of on/off cycles and add a bit of safety factor for elevated temperatures and fuse tolerance . If you cut it too close the fuse will just sag a little bit on each 'hit' and eventually fail.

In aerospace design I was asked to ADD all the worst case numbers, because after all they could happen. At HP we took the approach of the sum of the squares, which more replicates the assumption of non correlated statistical distributions, because as you pointed out, the likelihod of every term being the max at the same time is unlikely. In the Electronic Game Industry, we used what worked, plugged in and didn't fail, and could be purchased at the local shops across the US.

You've got to remember that this is a power supply. If it doesn't go, for any reason, including niusance fuse opening, it's not doing it's job.

Using a quick blow fuse smaller than 3A in any power supply with capacitive inrush is asking for trouble. Slow blow fuses are better suited for this job. As long as it protects against fire in single-fault abnormals, it's doing it's job - don't ask for precision or even overload limiting.

You haven't indicated the type of fuse involved, or the MOV marking. It's a fairly common mistake to use incorrect MOV's due to different mfr labeling practices and underestimation of normal line peak variation.

RL

The aim of the fuse manufacturer and the designer is to avoid this, particularly in anything without user-replaceable parts. It doesn't have to be inevitable, or within the design life of the equipment.

RL

Nope, pretty easy to make it so.

A few weeks ago I had a Garmin GPS fail on me because the fuse hidden in the spring-loaded knobby of the cig lighter adapter failed- caused some inconvenience (I had two backup GPS units including the car nav system, but the first one had all the locations programmed in). I really don't like being lost in the Detroit downtown area.

With AC-powered devices there's the additional variable of when exactly in the cycle the power gets applied, and maybe some memory of what happened at the end of the last power-on (especially with toroidal power transformers).