Fused HV

[Aurora]

I'm working on a ODS clone, - but on a more general basis, - what's your opinions on HV fuses?

[martin manning]

For FWB, fusing one leg of the HT is ok. For FW, fuse both legs of the HT, not the CT. Many many amps have been built with only a mains fuse.

[trobbins]

For FW, fuse both legs of the HT, not the CT.

Hi Martin, that is certainly the case for when a bias supply is generated off one of the legs, but otherwise I don't know of a good reason to not just fuse the CT ?
Ciao, Tim

[martin manning]

If there is a bias tap on the HT, then opening the CT to ground connection makes the bias tap an off-center tap, possibly causing damage to the bias supply components. Replacing the fuse could then create a situation where the HT is present, but the bias voltage is not. Another consideration is that a CT fuse will not protect the transformer if the rectifiers fail shorted, and this is also true for cathode biased amps. I should have qualified perhaps, but the question was asked in the context of a PP, fixed bias amp with Fender-type bias supply.

[Snicksound]

I've always put just the one after the rectifier, like Marshall and Fender before me. I don't like having the fuses hidden inside, and double fusing when using FW means I'd now have 3 fuse holders on the back (not counting if I decide to fuse heaters, or if I have a relay supply to fuse...). I know it's not perfect, but it's better than nothing as it MAY save the OT and or PT if a power tube goes bad in the wrong way.

I've had an issue recently where after replacing a PT with a lower voltage (but slightly higher current) one, the HV fuse kept popping when turning on. So I moved the fuse to after the first cap, like Marshall before me. Now if my cap fails it will take the PT with it. Oh well. Not like it happens very often.

A single fuse with FW also means no protection if rectifier fails. So I string 3 diodes a side. With a 2kV ceramic cap across each, Valve Wizard style. One diode failing? Possible. 2 diodes failing on the same side? Unlikely. 3 diodes failing on the same side? I can live with those odds. Only thing is I can't know if a diode as failed. But again, thousands if not millions of Marshall and Fender amps in the field in the same situation.

When using a tube rectifier, I add diodes just before as well. Funny thing is this makes the rectifier tube totally unnecessary from a functional point of view, but it still does its "feel" thing by limiting current, also the soft turn on.
Interestingly though, a GZ34 can take 1500V peak inverse voltage, 1.5x what a 1N4007 takes, so in theory it's less likely to fail. In practice... tubes are highly variable and stringing tube rectifiers isn't exactly doable.

Anyway, in 25 years of playing tube amps, half of those actually building them for myself and friends, a fuse has never saved anything. But it's cheap insurance.

You can either stick with the mains fuse, or you can go all out fusing everything, or you can stay in the middle like most manufacturers.

[TonewolfAmps]

I usually fuse the HV secondaries before the rectifier but after any bias taps off the HV. I avoid panel mounting any HV fusing just in case.

[trobbins]

The work-around for when there is a tap or take-off for a bias supply is to fuse the B+ directly after the rectifier and before the first filter cap. That location effectively see AC current, in the sense of the fuse current reaching zero which is key to applying a 250VAC fuse to HT/B+ power supplies.

Valve diodes will eventually degrade (like output valves), so adding ss protection appears to be common sense to me, and there are numerous forum threads and examples of failed power transformer secondaries that have a likely connection to an arcing rectifier. Valve diodes often have a higher PV rating than a 1N4007, so given any secondary higher than about 300-0-300V it is appropriate to use 2 series 1N4007 in series with each valve diode anode. If you have a batch/tape roll of 1N4007 then with suitable derating (such as the 300-0-300 tipping point) there is really no need for balancing caps for this type of mains frequency application. The valve diode still imparts its relatively high on-state voltage drop, and soft turn-off characteristic, so I don't see it as unnecessary at all.

So that leaves ss diode failure as effectively the last disadvantage to fusing a CT. The diode would have to fail as a continuous short circuit to damage the secondary winding, which for a small leaded package 1N4007 or UF4007 would have I'd suggest an extremely low failure mechanism rate (and I have yet to come across any forum comment on that as a failure mode). If the PT was unobtanium, then adding a redundant series ss diode would seem a reasonable action.

The concern I have with fusing each arm of a full-wave, apart from double the hardware and wiring, is that the what-if assessment starts to get complex.