To fuse equipment or not to?

At the current involved I guess so.

Interesting case regarding fusing. We just had an email about one of our amplifiers that had failed, where the PSU pcb was destroyed by some internal fault elsewhere. The track got so hot in one place that the board was charred rotten and the track wasn't there anymore. FR4 material so no fire of course.

For ages one of my colleagues has wanted to reduce internal fusing. Now he can see the effects !

I'm just considering doing the maths for cost saving in manufacture vs cost of replacing trashed product. Not to mention customer confidence.

Graham

This stuff can't be used for low currents because it's too fragile. AFAIK, copper wire is used.

I read in sci.electronics.design that Pooh Bear wrote (in ) about 'To fuse equipment or not to?', on Fri, 25 Mar 2005:

That could indicated a failure to conform to IEC/EN 60065.

All the wall warts (and even some small transformers) that I have looked at have had a Woods-metal fusible link inside them, close to and in series with the primary. Used as an overtemperature failsafe.

fuse

i.e. hot !

And you do have to watch how you mount these. Open wire fuses are a goodie as long as you realise a) the metal can get _hot_ when it pops b) " and stay running that way for hours c) metal splatter can occur on blowing, though mostly its either oxide, or just so thin as not to conduct at low v.

You can put a fuse wire straight onto FR4, and it will improve fire protection, but it may still convert FR4 to charcoal. Simple heatsinking/isolating methods can prevent this.

Introduce your fuse-phobic friend to single sided pcbs plus the 2nd layer of tracks being nowt more than ultrathin bare wire links stuck into the holes with components placed on top. When every other track is a fuse, you're fairly safe re fault currents.

Did that for a one off project where space was a real squish. Was fitting a stereo into a matchbox. No, not one of those bigger matchboxes, the little ones. Persuading nearby components to act as output device heatsinks was fun.

NT

I read in sci.electronics.design that Tony Williams wrote (in ) about 'To fuse equipment or not to?', on Sat, 26 Mar 2005:

You reply does not relate to the quoted text but to my remark about 'real (low-melting point) fuse wire'. Note 'wire' - pellets are a different matter when it comes to fragility, obviously.

I was chewing that one over. Would that be a fail due to temp rise of the pcb ?

I happen to have EN60065 at home at the moment so any guidance as to finding the relevant clause would be helpful.

I would really like to put my colleague *right* about my preference for fusing - he's been on about reducing this for over 10 yrs. That's what they're there for - to stop fires.

Incidentally I've come across a similar scenario with a QSC amplifier range ( unfused DC rails ) - where a far eastern company made a copy ( omitting the fuses ) and their test house insisted they be included. I know test houses sometimes differ over interpretation but this seems a big one to disagree about.

Graham

I read in sci.electronics.design that Pooh Bear wrote (in ) about 'To fuse equipment or not to?', on Sat, 26 Mar 2005:

Yes.

If you have the 2002 edition, it's 11.2.3 to 11.2.6 and Table 3 b) and e), column 3.

It's not only that. If the board burns up under fault conditions, what should have been a simple replacement of a faulty component becomes a major cost or even a write-off case. That leaves you with a very disgruntled FORMER customer.

I don't have that edition but I'm sure that'll point me in the right direction.

Errr - quite ! I agree wholeheartedly.

Thanks, Graham

The only advantages I can imagine to lopsided winding losses would be

1. Thermal. Since cooling is critical to an impedance-limited transformer, it may be better to have the heat where it's easiest to conduct it out.

1. If you also want to protect against an internally shorted primary. Skinny wire will just fuse.

But otherwise, lowering the short-circuit current trashes the regulation in proportion, no matter how you distribute the losses. You can always model a transformer as an ideal transformer with all the losses lumped and reflected to a resistor on one side or the other.

John

I'm not convinced by this argument. You might as well distribute resistive losses equally between the two windings to distribute power loss rather that just in one winding. You'd also need more copper with putting all the resistive loss in one winding.

Another method of introducing impedance is inductive loose coupled windings but this makes the transformer more difficult design and hence more expensive. The resistive - lossy method is cheaper using less copper and hence becomes the chosen solution. It's a doddle with small transformers..

You're missing the point Fred. The idea is to make the primary 'impedance protected' like those Boxer AC fans of old.

Graham

I read in sci.electronics.design that John Larkin wrote (in ) about 'To fuse equipment or not to?', on Sun, 27 Mar 2005:

My comments were not about impedance-protection but the use of safely-fusible wire for the primary winding. I have been looking for a mathematical analysis, but the core geometry is involved, so no general solution is possible.

You can't get as good regulation with any other winding area distribution than that which makes winding resistances proportional to turns-squared. But the loss need not be disastrous.