# Why did my fuse blow after 3+ years of no problems?

For several years, I've had a solenoid wired to the 24VDC output of a PLC (Rockewell Automation 1769-OB16).

To protect the PLC output card, a fast-acting 500mA fuse was installed between the PLC and the solenoid. The fuse has been working without any trouble for a long time.

Recently, this fuse blew. There was no change on the line, no abnormal or excessive usage of the solenoid, and nothing out of the ordinary. It just blew. I replaced the fuse with an identical one, and the solenoid works just as well as it did before the fuse blew.

I measured the current to try and find out why it blew, and found that the solenoid is actually pulling 530mA. I let the solenoid continue to pull that much for over 20 minutes, and the fuse held.

Why does the fuse not blow even though the load is pulling more than the fuse is rated for? And why would it blow only now, after more than 3 years, and not sooner?

• The environmental temperature has an effect; if warmer, then the likelihood of the fuse blowing (somewhat) increase. – Daniel Griscom Jan 3 '16 at 17:28
• Thanks, I was aware of that fact originally, and fortunately the fuse was in a room 65-70 (F) and has stayed that way. – CaptJak Jan 3 '16 at 17:30
• @DanielGriscom, conversely, mechanical parts often get stiffer in the cold. Depending on what the solenoid was driving and whether there's any feedback for a reduced hold voltage, it could be running flat out for longer than normal. OK it doesn't sound like it in this specific case, but this could be implemented with a swicth attached to the solenoid. – Chris H Jan 4 '16 at 9:13
• Exceeding the nominal current leads to premature aging. Even if the new fuse didn't blow within 20 minutes, it won't last as long as it is expected to. – Dmitry Grigoryev Jan 4 '16 at 10:38
• Don't think of a fuse as a current limiter. Think of it as catastrophic fault protection. It's there to prevent surge and short failures from turning into fires, not protect a slightly overloaded power supply. – Passerby Jan 5 '16 at 1:26

## 4 Answers

The rating of a fuse is the amount of current that it will carry indefinitely without blowing. In order to guarantee this, most fuses will not blow until the current rises to 2× or more of their rating. In fact, if you look at the datasheet for a fuse, there will usually be a chart that relates time-to-blow to the percent (over)load. Most such charts rise to infinite time to blow somewhere near 200% load.

If you're putting somewhere between 1× and 2× the current rating of the fuse through it, you're in a gray area in which it may or may not blow, or it may just get weaker over time, eventually leading to a lower threshold for blowing.

There are other things specific to solenoids that could cause such a failure, too. At a simplistic level, you can think of a solenoid as an inductor, and its DC resistance limits its steady-state current. However, when the plunger is actually moving, the inductance is changing, and this creates an extra surge of current each time you operate it. If something in the mechanical load causes the operation to be somewhat slower than normal, this surge will last longer, and could potentially blow your fuse.

This is why slow-blow fuses are normally used with loads that have startup surges. It is also why you should use a PLC output that is rated to handle the surge.

• Thanks. The PLC I have is rated to handle the surge (1A constant, and up to 2A surge each 10 ms), I just wasn't the one who designed the system. Would replacing the fast acting 500mA fuse with a slow-blow 500mA be recommended? Or should I replace it with a fast-acting fuse of a higher rating? – CaptJak Jan 3 '16 at 15:35
• I would be inclined to use a 1A fuse, which is a good match for the rating of the PLC output. – Dave Tweed Jan 3 '16 at 15:37
• @DaveTweed but as you note, a 1 A fuse could carry up to 2 A for an indefinite time, which would be outside the PLC's 1 A max / 2 A surge rating, no? – nekomatic Jan 4 '16 at 8:53
• @nekomatic: No. As I said, the range between 1x and 2x is a gray area, protecting the margin of testing error and production tolerances for the manufacturer. The only guarantee of infinite time is at the actual rating. – Dave Tweed Jan 4 '16 at 11:45
• @DaveTweed 'indefinite', not 'infinite'. I think that means about the same as 'grey area'. It seems to me a 1 A fuse could allow the PLC output to be overloaded for an extended period of time. – nekomatic Jan 4 '16 at 13:05

## Fuse Design

Looking at this datasheet for fast 5x20mm fuses from ESKA, you find a table with "pre-arcing time limits" at the bottom. For a 500mA fuse, it states:

 2.1*500mA  =1050mA:         30min
2.75*500mA =1375mA:     50ms-2s
4*500mA    =2000mA:     10-300ms
10*500mA    =5000mA:         20ms


So, it is completely within the specs (of this fuse) that it doesn't blow at 530mA within 20min.

## Fuse Aging under constant conditions

On the other side, it's strange that the fuse DID blow. We once saw the same weird behavior and did a test. We had four 1A fuses in series, each with a diode in parallel. This was connected to a constant current source of 1A, and the voltage drop over each fuse was monitored. As long as the fuses were OK, the voltage drop was far below 0.7V, and all the current went through the fuses. A blown fuse would be indicated by a voltage drop of about 0.7V:

simulate this circuit – Schematic created using CircuitLab

We put five of this sets equipped with fuses from different manufacturers and batches into an oven with a constant temperature of 85°C to add some stress and got this quite interesting result:

All fuses started to "age" immediately, as the slowly increasing voltage drop shows.

• Set 1 has a high spread in voltage drop and the first fuse starts to increase the voltage drop at day 10 and finally blew 15 days later(!)
• Set 3 is a different batch from the same manufacturer. It took longer, but the result is the same.
• Set 4 started with a high voltage drop, but seemed to settle somewhere.
• Set 2 starts with a low drop and the slope starts to increase at day ~25-30
• Set 5 has a superior small voltage drop and really small part spread, but even here, the slope starts to increase very slowly around day 30. The price of this fuses is 3x the price of set 1 and 3...

(I have data for a longer period, but need to search it)

Here is also a picture of the fuses:

From left to right:

• New and unused
• After some time in one of our devices which often blew fuses, already has a slight yellow tone.
• Still working fuse, lots of black oxidation visible
• blown fuse, seems it blew relatively fast.
• Another blown fuse. It seems it took looooooooong to blew.

## Fuse aging by switching loads

The fuse wire gets hot and expands when current flows. At high temperatures, oxidation may happen, which weakens the wire mechanically, and may be electrically, too. Switching on/off a load means the wire is bent each time. This stress may cause a blown fuse at some point even when the current never exceeded the threshold.

## Bad circuit design

Of course, as you are writing about a solenoid, it's possible that there are short but large pulses across the fuse which will also damage it over time.

## Recommendation

A fuse usually is not there to protect the device in first place, but to protect the source or to avoid further damage e.g. by a fire.

The manufacturers said, it would be best to use a fuse rated for 1.5-2 times the max. expected current, though a too high rated fuse may not blow when it should.

However, aging still happens, and fuses blow from time to time without (external) reason.

• Makes sense. The data sheet for the fuse does show that it can go 30-60 min at 150% of the rating. The room where the fuse is kept stays at around 66 F, so I'm going to go with age and the nature of a solenoids operation as the cause. The graphs and illustrations were very helpful. – CaptJak Jan 3 '16 at 16:17
• How much do you think the ageing is accelerated by using 85 °C? Could it be the common factor of 2 per 10 °C (that would be on the order of 50 times)? – Peter Mortensen Jan 3 '16 at 17:40
• That's a really excellent answer. I had a car (Saab) which died after 10 years. The ignition circuit fuse had distorted and opened up while appearing good. The fuse element looked OK, although somewhat elongated, but crumbled to dust at the faintest touch. Swapped out a fuse from the A/C circuit (it was winter) and motored on. – WhatRoughBeast Jan 3 '16 at 18:05
• @PeterMortensen probably not. It completely depends on the activation energy for the reaction. That rule of thumb is based on reactions that have an appreciable rate close to room temperature, but the rate of fuse degradation at room temperature is pretty much negligible. If the process that leads to ageing is known, then its activation energy could be used to obtain an accurate (up to scale) temperature dependence. – Oleksandr R. Jan 3 '16 at 20:15
• @Rev1.0 In fact, the test was running at least 100 days, I just can't find the data... I tested this at university, the oven was free and we just let it run. In fact, the project is indeed a bit larger and the fuses were failing in devices which couldn't be accessed for months when in operation. While we just replaces 1A by 1.6A, this measurement was more or less interesting for us. – sweber Jan 5 '16 at 8:41

I have seen fuse melting curves with time against current, and those curves do NOT go to 'infinity'. Infinity does not exist on graphs. Instead, with currents somewhat around or under nominal, the time will go to some large number of hours, say 1000 to 10,000. And if the current is not 'on' at all times, as probably in your case, 3 years is a reasonable 'long' time to melt.

Another way to look at this: a fuse is like a light bulb, the old type with a hot filament. Although very hot, it still takes some amazing 1000 hr to melt. And even at under-voltage, burning lower, it will not light 'forever'.

To back up my argument, here is an arbitrary fuse time diagram, found with google. It shows a pretty straight line on logarithmic scale over 5 decades, from 0.01 s to 1000 s. From 1000 s to 3 years is another 5 decades.

As yet another argument: I have seen fuses blow after 25 years of service. Once, around 2010 or so, I replaced a blown household main fuse while in Iran only to discover a coin from before the Revolution (1979)! (I am not making this up) Without an apparent overload or short circuit event.

• Thank you. First, I am pretty sure that all participating in this thread know that it isn't actually "forever", it's just figurative for "indefinitely/a really long time". Second, do you have anything to back up your answer? Tests? Sources of information? I do not believe that 3 years is a reasonable time, especially considering that I have hundreds (literally) of older fuses which have not blown. – CaptJak Jan 4 '16 at 17:48
• @CaptJak I added some concrete data to my answer. About your fuses: do these all carry a constant current around the nominal value all the time? If not, you will see at the sample graphs that at a lighter (average) current load the melting time will increase significantly. – Roland Jan 5 '16 at 0:48
• It varies. Some of the fuses (like the one that blew) only carry the current momentarily, while others are constant. Some carry less than they are rated for (250mA) and others carry their rating or slightly higher. – CaptJak Jan 5 '16 at 1:35
• And I don't doubt that they will die eventually. But three years seemed too soon, especially considering that it was the only one of many fuses. – CaptJak Jan 5 '16 at 1:42

It is a common misconception that a fuse is supposed to melt immediately ('blow') at its rated current.

A fuse is used to protect your wiring from causing a fire in the case of excessive currents, and/or to protect against electrocution in case of isolation failure. A fuse may work in conjunction with earthing, causing a high short circuit current in case of a minor isolation failure. In the optimal case, the fuse will blow very fast because of the high short circuit current.

Fuses are known to be difficult to rate to protect against overheating your wiring in case of currents that are only marginally higher than designed. That is why electrical engineers need those complicated fuse graphs.

Looking at those graphs (e.g. see my other answer) you will see that a fuse rated 100 A will take several seconds to meld at a current of 10 times that value. This will explain your question as why does your fuse of 500 mA 'not blow' with a load of 530 mA. No, it will not blow at once, but may/will meld at some later time. A fuse of X Ampere is not by itself very useful for a load that nominally draws X Ampere. E.g. in my house I might turn on a lamp of say 10 Watt, while the installation is fused with 16 Ampere (at 230 Volt AC).