Fast blow or Time delay fuses?

[Randall]

Or medium blow for that matter. I generally use slo blow in tube amps, mainly because some amps specify this. But how to know what to install if it's not spec'd?

[DRH1958]

If not spec'd, then I like slo blo because of the temporary inrush current when the power switch is thrown. When first powering up, the PT is initially close to a short circuit. As voltages build up to normal operating values, this of course subsides and the amp settles in. The slo blo action will allow this temporary overcurrent condition to occur without nuisance blowing of the fuse. If a short or other catastrophic event occurs, the fuse will act normally and blow. If a fast blow fuse is used, you might need to up the value of the fuse a notch to allow for this initial inrush current.

If an overload condition happens, which is a slow increase to a current level higher than normal such as might happen with a slowly failing component, the slo blo fuse will offer better protection because the rating is closer(smaller) to the value needed, whereas the fast blow you upped the value on will allow a higher overload current before it blows. So the two different types of protection a fuse offers needs to be kept in mind (short circuit and overload).

[Jazz P Bass]

In a previous post, I believe G1 spelled it out as "if the type is not listed, it is a fast acting fuse".

[Enzo]

Agree, if it isn't stated slow blow, then it is not slow blow. I always use the listed fuse type for an amp unless I have a damned good reason to change it.

Some OEMs use say a 3A slow blow fuse, where some other might use a 5A fast fuse. And these would be similar amps. One designer used the slow blow at the lower current to cover the inrush surge. The other used the fast blow, but raised the current a couple amps to handle the inrush. Different approaches. if though you decide to change that 5a to a slow, now you removed a lot of the fuse protection.

[trobbins]

Tube amps normally incur transformer inrush and heater inrush. If the amp uses ss diode supply then there is also a capacitor inrush. All of those inrush currents are best managed by a slow-blow or T (for time delay) labelled fuse.

If the amp doesn't come with a fuse spec for some reason, and you have the schematic and enough technical interest, then you could attempt to reverse engineer it. Similarly, if you add a secondary side B+ supply fuse then you can have a go at designing the fuse spec. Interestingly, the fuse spec design process has some common design aspects as how to choose an NTC if you want to curb inrush.

http://dalmura.com.au/projects/Valve%20amp%20fusing.pdf

[DRH1958]

And there's this directly from Littlefuse's site FAQ section: "How should I select the proper fuse amperage when I know the circuit amperage?
Always follow NEC guidelines for applying low-voltage fuses. Generally, the MINIMUM fuse size should be based on 125% of the circuits full load current. Time-delay fuses should be used for inductive loads and fast-acting fuses used to protect non-inductive loads". End FAQ.

The amp provides an inductive load in the form of the power tranny, so it's my understanding there should be a time delay fuse in the 120v supply according to that answer.. It's hard to argue with the company that makes the fuses. Since Enzo did this for a living and Jazz P posted G1's response, I tend to agree with them that this is the accepted practice. But consideration needs to be given as to what is the recommended way by the manufacturers. And of course adding an HT B+ fuse helps to give more protection as well. But for just a primary only fuse, I like the sound of having a 3 amp time delay fuse rather than a 5 amp fast blow. In an overload situation, this gives 2 amps more protection.

[R.G.]

All are correct views of pieces of the elephant.

Fuses blow because their filament melts. In turn, that means that the filament temperature was heated by I-squared-R resistive heating and could not get rid of the heat fast enough to keep its temperature below the melting point. There are two ways this happens: (1) the slow, let's-nobody-get-excited way where the current gently rises to the point where the mechanical structure of the fuse can't conduct and convect the heat away from the filament, and it slowly, gently sags and opens, and (2) the hammer-blow of a current pulse. Well, OK you could have a gentle average level with a hammer blow on top of it, but we're talking start up here usually.

It takes some time for heat generated in a part of a device to travel to other parts of the device. If you hold a bar of copper in your hand and stick the other end in a flame, it takes some time for you to have to drop it because it's so hot. You don't instantly get blisters. That's thermal time delay. If you use a steel rod in your thought experiment, it takes longer for the rod to get too hot to hold, because steel is both a higher specific gravity (more mass to heat up per volume) and a poorer conductor of heat, higher thermal resistance. The length of both rods can be thought of as a series of series resistors (the thermal resistances per length) and capacitors to "ground", the mass to be heated up. This is why it's easier to weld steel than copper - you can heat the steel to melting in a small puddle because the surrounding steel can't conduct heat away very fast. Copper is tougher to do this with because the surrounding copper sucks so much heat out.

In a fuse, the heat is generated in the filament. If it's a slow, gentle rise in heat, some of it escapes by conduction to the end cap connections of the filament, and some is conveyed away by the internal gasses in the fuse circulating and carrying it off. If it's a pulse, the heat doesn't have time to drift slowly off into the sunset, and the temperature rise depends on the mass of the filament to be heated (n.b. and a few other mechanical issues).

So t make a time delay fuse, you make a higher mass filament, to let it soak up more internally generated heat from pulses before it melts. Then, you diddle with the metal alloy to make it melt open at a temperature that opens it at the slow, gentle rate.

The pulse-energy a fuse can absorb is hidden its I-squared-T rating on its datasheet. This is a very big deal when selecting a fuse to protect a power device, like a triac, SCR, or power transistor. You have to make the fuse's I^^2*T rating smaller than the equivalent value of the device. If it's bigger, the semiconductor will give its life to save the fuse.

Pulsed currents are why we have slow-blow fuses, as noted. Transformers and motors are always heavy pulse currents; shunt capacitors are as well. A transformer feeding rectifiers and empty caps are a double shot of pulse-love. I think it would be very, very unusual for any designer worth his pay to specify a fast blow fuse for a tube amp. You have to run the carry current so high that it doesn't do much for normal fire prevention (again, as noted). But as Enzo says, use what the name plate says.

Transformers work by filling up the core/iron with magnetic field energy, converting the AC line voltage/current into M-field energy to make things work. At first power on, that tank has to be filled as fast as the line can fill it, and you get a big surge. I once instrumented a variac in the power supply lab to find out why it popped breakers about every fourth power-on. "Normal" power-ons were about 60-70A in the first half-cycle of AC. Sometimes, when the phase of the AC line and phase of the moon aligned, I say over 200A as the breaker was opening.

And remember, the purpose of AC line fuses is NOT to protect the power transformer or anything else inside the box. It's to prevent a fault inside the box from starting a fire.

[g1]

I'll have to be honest here and say I sometimes use slow-blow in tube amps where it is not specified, so I didn't mean that statement as "written in stone".
But I respectfully disagree with the idea of considering anything with a power transformer as an inductive load, I'm not sure that is what they (littelfuse) meant.
Consider the vast number of consumer electronic products that have power transformers, and the small percentage of those that have slow-blow fusing.

[Mike Sulzer]

Transformers work by filling up the core/iron with magnetic field energy, converting the AC line voltage/current into M-field energy to make things work. At first power on, that tank has to be filled as fast as the line can fill it, and you get a big surge. I once instrumented a variac in the power supply lab to find out why it popped breakers about every fourth power-on. "Normal" power-ons were about 60-70A in the first half-cycle of AC. Sometimes, when the phase of the AC line and phase of the moon aligned, I say over 200A as the breaker was opening.

The energy goes in and out of the magnetic field twice per cycle. An inductor draws current slowly, not with a surge, and the inductor is a large value so that the magnetizing current is generally smaller than load current. It is the capacitors on the other side of the diodes that draw the surges.

[R.G.]

I'll have to be honest here and say I sometimes use slow-blow in tube amps where it is not specified, so I didn't mean that statement as "written in stone".
But I respectfully disagree with the idea of considering anything with a power transformer as an inductive load, I'm not sure that is what they (littelfuse) meant.
Consider the vast number of consumer electronic products that have power transformers, and the small percentage of those that have slow-blow fusing.

Like everything, it's a matter of degree. When you start a transformer by connecting it to the AC line at any point on the AC voltage wave, it does two things. First, it instantly transfers energy to the secondary load through transformer action, and it also starts pumping up the field in the core. Ignoring the secondary load for a moment, we can look at the current needed to supply the core's M-field.

It is quite rare to have a transformer core without some remaining magnetic field from the previous turn-off. It can happen, but usually you have to de-magnetize the core with a steadily decreasing AC voltage/current to run the remanence down to nearly zero. The incoming voltage feeds the M-field, the core's ampere-turns increasing at a rate of di/dt = V/L. Neither V nor L are constants in a transformer power up, but we'll ignore the change in L for the moment and call it constant if there is no saturation. If you happen to hit the remanence of the core where the incoming AC reinforces the magnetic field direction already in the core, the Vdt integral ramps the M-field up from the remanence value.

Power transformers are designed so that for their maximum input voltage and minimum input frequency, the field density just stays below saturation. So for a full half-cycle of AC, the Vdt integral pushes the core to just below saturation, and that's where it runs, the core being run from just less than + saturation to just more than - saturation by the alternating AC wave. This is done to get the most use of the core iron and copper.

But if you get a remanence field plus a half cycle, you're guaranteed to push the core over into real saturation. At that point, L does change, dramatically, and the primary inductance quits opposing incoming AC line power. The current increases dramatically to a value limited only by what little inductance is left, the primary wire resistance, and any thing at the line cord and back to the generating station that limits current. This is what was happening with my variac. If I happened to turn it on near a zero crossing with the voltage headed in the direction that enforced the remanent field direction, it ramps up into saturation and gets a whacking big saturation pulse. All possible combinations of the incoming power on timing aiding and abetting the remanence are possible, as well as all possible values of remanance from exactly where in the AC wave the core was turned off last time. This matches the real world. I recorded inrushes from nearly zero to over 200A peaks on the same transformer.

The reason I only noticed this on the lab's variac was that it was a huge thing, probably a foot across, and it had a great big core (lots of energy stored in the remanent field) and big, fat primary wires that didn't limit currents much. It had a 15A slow blow fuse that didn't blow; the wall breaker went first. The bigger the transformer is, the bigger the core to be filled is, and the bigger the primary wire is, so its resistance doesn't limit current inrushes as much. The smaller a line transformer is, the more the primary wire resistance is a limitation. So yes, on consumer items with small transformers it's less of a problem.

The winding pattern also makes a difference. The more leakage inductance a transformer has the more the leakage impedes current inrushes. Leakage inductances are by definition the magnetic flux in the air that does not couple to the core or other windings, so if the main core inductance suddenly changes, the leakage doesn't, and it helps cut down in the inrush. Small consumer transformers are almost always wound side by side these days, with primary in a lump beside the secondary. This is the highest possible leakage configuration. The bigger the transformer, the more likely it is to be wound primary over secondary or vice versa, the lowest leakage simple winding configuration.

Toroids are yet another special case. Variacs are almost always toroids. Toroids have the smallest air gap you can reasonably get, and so the change in core inductance with flux density is the biggest. E-I cores always have a distributed air gap at the ends of the Es, causing more leakage and more resistance to inrush pulses, but also poorer performance due to the leakage.

The energy goes in and out of the magnetic field twice per cycle. An inductor draws current slowly, not with a surge, and the inductor is a large value so that the magnetizing current is generally smaller than load current. It is the capacitors on the other side of the diodes that draw the surges.

Power on inrushes are a transient, time-domain phenomena, not a steady state phenomena because of the sudden change when the switch is thrown and the initial conditions imposed by remanence and where in the AC wave the voltage is removed and applied. Transformers can and do have inrushes all on their own. Motors do exactly this, but add on the problem of the rotor being locked by inertia at t=0. Locked-rotor is the mechanical analog of a big fat capacitor on a secondary.

[DRH1958]

I'll have to be honest here and say I sometimes use slow-blow in tube amps where it is not specified, so I didn't mean that statement as "written in stone".
But I respectfully disagree with the idea of considering anything with a power transformer as an inductive load, I'm not sure that is what they (littelfuse) meant.
Consider the vast number of consumer electronic products that have power transformers, and the small percentage of those that have slow-blow fusing.

Adding to what RG said(I wish I understood it more), there's no purely inductive load since wire has DC resistance, so it's a combination of the two and the total current is dependent on the vector sum of the resistance and inductive reactance(the purely inductive portion of the total) since they are not in phase with each other. Current allowed by pure inductive reactance is 90 degrees out of phase with current allowed by resistance so they are summed vectorally. Then as R G referred to, you need to consider the secondary's part in all this and of course the magnetic field. I think Llittlefuse refers exactly to the transformer as one type of inductive load since a purely inductive load cannot exist. Transformers and motors are as close as we can get. The alternative is a purely resistive load, a capacitive load or resistive/capacitive load.

[jmaf]

Came here to chip in 2 cents and walked out with 2 grand courtesy of RGpedia.

[g1]

My point was more about practical real world applications. If everything with a power transformer is considered an inductive load, then most consumer electronic units are doing it "wrong" according to littelfuse (at least prior to the advent of SMPS).
In practice, the only place we commonly see slow-blows are in tube amps. It was always my understanding that this was due to the heater surge at start up, not because they have PT's.

[DRH1958]

My point was more about practical real world applications. If everything with a power transformer is considered an inductive load, then most consumer electronic units are doing it "wrong" according to littelfuse (at least prior to the advent of SMPS).
In practice, the only place we commonly see slow-blows are in tube amps. It was always my understanding that this was due to the heater surge at start up, not because they have PT's.

Yes G1, I agree that like many threads, we have strayed slightly off course. I think we are debating the correct way to protect equipment when the OP just wanted to know what to do if there is no designation as to whether the fuse is slo blo or normal fast acting. If the standard practice is unless it specifically says sol blo, it is assumed to be fast blow then that's what the answer is. It totally makes sense. I'm sort of the new kid on the block here having been a commercial electrician all my life and am piggy-backing on some trade school knowledge of yesteryear to have fun in retirement with amps and I did agree in principle to that in post #6 to have this a the accepted practice. But when I did further research just for my own curiosity, I came across that Littlefuse FAQ and thought I should bring it up just as some pertinent info. It made me question "standard practice". Maybe that is possibly outdated and some new thought might be in order? Who knows. I guess we have one of those "whatever you think is best" situations. As R G first said in post #7, "all are correct views of pieces of the elephant". But I do think possibly since most consumer electronics is NOT what we would call high quality constructed solid state items, maybe they err on the side of safety and use fast blow and keep the size lower to not sacrifice electronic components in the name of saving the fuse. Again, who knows for sure. As I'm finding out in tube amps, there are a lot of subjects left to the discretion of the individual as to what they like to do. But it HAS been a nice learning experience to be sure. I've filled in many holes in my knowledge base here on MEF.

[mhuss]

In practice, the only place we commonly see slow-blows are in tube amps. It was always my understanding that this was due to the heater surge at start up, not because they have PT's.

I agree. Virtually everything has a transformer, and virtually none of these transformer-containing devices have slow-blowing fuses. I really think the slo-blo tradition is more about the cold heater surge than it is about transformer or capacitors.