I have to ask about this idea of placing diodes on the PT HT leads to the Rect tube.

[R.G.]

I read somewhere today that the same value fuse placed in the D/C end is less likely to blow as fast as it will if placed in the A/c side . I think the idea behind this is the current . In a 12 VDC car you are working with much higher current 6VDC even more current .

Welcome to session #37 of the continuing series "Why We Do Math".

Back when there was only DC, Smart Guys figured out from nothing that the amount of heating was proportional to the voltage times the current. Later, other Smart Guys figured out how to generate sine waves by spinning a loop of wire in a magnetic field. This was found to heat things up just like DC from batteries did, but the amount of heating was no longer voltage times current, because it was always changing.

Some of the math Smart Guys figured out that if you took the voltage times the current in tiny slices of time, slices so small that the voltage and current didn't really change much during that time, and then added up all those tiny power-time slices, you'd arrive at an average over time of the actual heating that would be the same as some DC current/voltage. A lot of (manual!!) math let to the idea that the equivalent of a DC current could be computed for a sine wave current by taking the square of the incremental current (one increment being a time slice, and power being proportional to I-squared times R) for a lot of time slices, averaging all those current-squared intervals, then taking the square root, you'd come up with a number which was the same number as a DC current which heated a resistor an equal amount to the AC current. That is, you took the square ROOT of the MEAN (i.e. average) SQUARE of the AC current, and got an equal heating.

Back at the rectifiers and fuses. The RMS value of any waveform can be computed with a little care and calculus. Fuses blow because of heating. A fuse which blows at, for example, 1.037A of DC will blow at 1.037A of AC **RMS* current too. Same heating ability.

So can a fuse blow more easily on DC than on AC? To a first order, no. They heat the same (yes, there are lots of caveats hidden there). But what can happen in rectifier circuits is that the rectifiers divert currents around. A full wave rectified AC waveform has the same RMS value as the AC waveform. But it does not have the same average DC value. And two half-wave rectified AC waveforms don't have the same RMS value as a full wave rectified waveform. So a fuse will blow very nearly the same on DC as an equivalent AC RMS current. But you do have to be careful to know what the RMS value is you're talking about. Comparing apples to apples can require some work.

[trobbins]

Fusing DC is much more onerous on the fuse, and normal fuses you find being used for AC aren't rated for DC. That's not to say that an AC rated fuse couldn't clear a B+ DC fault, but you are likely to get an internal fuse arc that leads to other issues (eg. shattered glass).

[catnine]

I have all I could find which is a 250 mA fast blow fuse in the B+ before the 1st cap , the PT is 100 mA so what good is the fuse?

[Enzo]

The PT may have a 100ma rating, but that means it can put out 100ma at the rated voltage all day long. The PT is capable of providing a lot more than 100ma. If you draw that "lot more" for an extended period of time, then the resistive heating of the wires inside will overheat the transformer, the insulation will break down, the plastic bobbin can melt, the PT would be destroyed. But momentary current surges won't burn out the winding. SO if there is a short across the first filter cap for example, or a power tube shorts to something, for a moment, that transformer can provide maybe an amp or two, and thus your fuse blows.

[trobbins]

PT rating is a continuous rating (I assume) - ie. the transformer doesn't overheat. If a short-circuit occurs, then the fuse will blow in a time that is relative to the max current that occurs - which depends on the nature of the short, and the impedance of the source - you may get a short circuit current well in excess of 1A perhaps. Fuses take time to blow, so best to use a fuse with a close rating to your max continuous, but midfull of surges. Usually the fuse gives out before the wire etc in the PT overcooks ;-)

[diagrammatiks]

you could 600 fuses in an amp if you wanted to.

you get a fuse block that's made for tinyfuses.

I don't think there's an amp chassis ever that couldn't fit enough fuses to fuse everything if you wanted to.

[kg]

Suffice it to say that you cannot guarantee that the 5V heater or 6.3V heater windings will blow the mains fuse.

in fact, i would almost guarantee it would NOT trip the mains fuse.

[ThermionicScott]

Most old amps say a SF fender champ , the 6.3 volt heaters are on one side grounded to the chassis , so you would have to remove the ground add a twisted pair and they are twisted for a reason then place the fuse either right near the PT of the pilot lamp , then if you have a 50 volt bias tap add another fuse. On these old factory builds there is just no place for all these fuses and trying to do so may well create a short or hum no matter if you you the snap in fuse holders in the amp of twist cap type and drill holes . So yes the diodes save form running A/C to the caps and rest of the amp and may save the PT yet if any other winding shorts for what ever reason even a rare one you are out shopping for a PT . Has anyone got a kit from any well know builder that hass all of this , I mean kits you buy to build or buy built ? Allen amps, Weber, Hoffman , Mission?

Some thoughts:

1) No need to convert to twisted-pair heaters if you fuse both sides of the winding.
2) Fused bias supplies have an ugly failure mode if *just* the bias fuse blows -- I'd rather the main or B+ fuse popped.
3) A barely-adequate main fuse gives better global protection in non-critical applications (practicing at home or with the band). You can always switch to the regular value for the gig.
4) There's plenty of room in a Champ chassis for extra fuses.

- Scott

[trobbins]

One problem with sizing fuses is the impact of in-rush current, which can often push the rating used to the next highest level to avoid blowing the fuse. A suitable NTC in the mains side can sometimes allow the ac input fuse to come down closer to max continuous level.

[Enzo]

And that issue is precisely why slow blow fuses were developed.

[trobbins]

I agree that time delay characteristic fuses are designed for that pulsed application, however there is a subtle price to pay with respect to overload ratings and I2t which lowers the level of protection for faults which have low fault level ratios. The consequence obviously depends on the character of the fault, and whether it matters if it takes 1 second or 1 minute to trip the fuse, and mindful that fuse tolerance is very wide anyway.

[R.G.]

It all eventually comes back to the fact that mains fuses can not be expected to protect anything inside the unit. They are intended to stop a failure that's already happened inside the unit from starting a fire or electrocuting someone. They protect things outside the unit.

It sometimes happens that a mains fuse will stop one failure - say, a shorted rectifier or filter capacitor - from killing the power transformer, or causing a chain of destruction that kills almost everything. But it that happens, it's a happy accident.

The necessary compromises between passing a start-up inrush and tripping soon enough to save a transformer (or filter cap, or rectifier, or tube, or whatever) can hardly ever be resolved completely.

If you want to use a fuse to protect a specific part, you have to (1) first know what current and energy (that I²T thing) the part has to pass in maximum-normal operations, and (2) the amount the part will tolerate before dying, then (3) attempt to find a fuse (across available values and fuse tolerances, if you can find a specification on that) that will pass the maximum working current and I²T, but trip before the failure current and I²T that will kill the part.

If that sounds like a tricky job of calculation, it is.

Especially since it depends on the surrounding ambient temperatures, which a designer sometimes doesn't know all that well. Expecting that you can simply put one fuse in the primary that will take into account all of the inrush *and* normal *and* maximum *and* just-tolerable overloads in the whole unit and step in to protect all the parts just before they fail and not cause an otherwise-unnecessary false trip is not a thing one can rationally expect to happen.

What one can expect is to figure out what a single winding will live through (for instance) and use one fuse for that one thing.

The idea that one fuse in the AC primary can protect everything in the amp without being massively too sensitive and preventing otherwise normal operation is charmingly simplistic. It's a testament to the toughness of most tube amps that it even exists at all.

[Enzo]

The difference between fast and slow fuses us more than the time factor. The idea when using them is that a circuit that draws a half an amp might want a 1A fuse. But that 1A fuse blows maybe 10% of the time at power up. SO we install a 4A fuse to withstand that. Of course now it takes 4A+ to blow the fuse. By using a 1A slow fuse then, we can have the protection from overcurrent, but without the inrush nuisance blows. One would not just sub a 4A slow blow for a 4A fast.

[trobbins]

I agree. It may also be possible to use a 1A normal time response fuse with an in-rush reducer, and for that situation the 1A normal fuse would give 'some' improvement in theoretical protection than a 1A slow blow fuse.

I see protection as a collation of all design and practical build factors that could cause a failure. If one particular item can be improved or removed then that is a good aim. I don't subscribe to the 'thats how it has always been done, and lets not worry about it evere again' approach. I agree that fusing is a very broad brush type of protection and throwing too much engineering at it is somewhat futile, but if in general a 'better' spec fuse can be used then why not use it.

[diagrammatiks]

the only problem is that fusing is pretty much the wrong type of component to do any sort of in circuit protection as RG said.

it's there to prevent gross failures and fire hazards.

thermistor, varistors, time-delay relays, current sensing relays, anti-surge resistors, etc are the components for protection circuits if you actually want to protect components like the transformer, tubes or whatever.

you could hardly expect say an aerospace certified circuit to be protected by a 50 cent fuse.