Thoughts on cathode current sensors

[R.G.]

There's a lot of stuff compacted in this one thread.

IMHO, using higher wattage cathode resistors and not trying to also use them for fuses makes more sense. This is because:
1) From a technical perspective, designing a fuse to blow at a certain energy level - not current - is challenging, and fuses are designed to do this properly.
2) Resistors are not designed to do fusing at any certain level, so the uncertainty about where and when they will "open" is vastly more questionable.
3) Worse, the exact results when a resistor "opens" is vastly questionable. There are conditions where the resistors first shorts, then burns open, or where conductive byproducts are produced in the cooking process, with deleterious effects.
So to the informed designer, using a low-power resistor for a fuse is what I'd call "feel-good design". It posits a possible way where your amp *might* be protected without any real knowledge that it is. If your house catches fire, do you want to know that the fire department might show up or will show up?

That puts me on the side of using bigger, sturdier cathode resistors and not "maybe they'll be fuses too!" small resistors.

And that leads right into the question of how to protect the PT and OT. As mentioned, if you want to protect a transformer, you fuse each winding. This idea was not new with Merlin, nor me, nor O'Connor where I also saw it, nor my professor in my AC power course. It comes from the early half of the 1900s when AC power and transformer design was on the cutting edge of electronics.

Simply put, the uncertainties about when a fuse will blow, instead of when it will not blow are too big. A fuse rated at 1A is not guaranteed to blow at 1.01A. It's guaranteed NOT to blow at 1A, to carry at least that much. Small overcurrents may blow in hours. Protection by fusing is a game where the fuse can be counted on to protect against massive overcurrents, not "soft shorts", mild overcurrents.

To protect a transformer, you fuse the windings just a bit over their normal operating currents, allowing for startup transients and full-load operation. You then count on the transformer being a much more massive object, and therefore much slower to heat up. The difference in mass and thermal properties then get the fuse to open before the transformer can overheat and die.

It is a misconception to think that the AC power line fuse is there to protect the power transformer. It is not. It is there to prevent fires and electrocution in case of massive disasters after the fuse, and that's why it is (or should be!) the first part the power line wire hits after the AC cord enters the chassis.

I had a weeks-long "discussion" with a semi-technically-literate fellow over at MEF about this topic and led him as gently as he'd allow me through the rationale that showed that a soft short on the heater windings could burn out the heaters in a PT without blowing the AC mains fuse. It is possible - maybe, sometimes - to protect a single-secondary transformer with a single primary fuse, but when you get into multiple windings, your only course for a doing a decent job is to fuse each separate winding section.

Yes, that's a PITA. Yes, it's discouraging and difficult to find and mount fuses to do it. But the PT and OT are the first and second most expensive parts in most amps. And small fuses are available, as are in-line fuse holders. It's probably not necessary to put all these fuses on the back panel, anyway, just on the basis that if one blows, you would have had to pull out the chassis anyway to replace the transformer (or resistors if you did fusing that way) if the fuse wasn't there.

And you have a much better chance that what you replace will be a fuse, not a transformer.

The solution I actually prefer is not fuses. It's an internal current monitor that looks for over-currents and shuts things down with a massive-overkill power device. Massive-overkill MOSFET switches are cheap these days, probably cheaper than a few fuses. It takes some additional electronics and such to operate this well, but there's a high payoff for it. I'm working on a suitable solution for DIYers at my usual (...slow...) pace.

[xk49w]

I have some Littelfuse 2AG (5x15mm) axial through-hole fuses I was intending to place in series with each leg in the rectifier chain. They should never blow in normal operation. If they do the amp would require servicing. Of course, without them it would require servicing also, only for more $$ if a transformer were to be replaced. Finding the right value is the trick. So far I have picked 0.5A, 0208.500MXEP. This for a 50W output stage, 2x6L6.

-- bradley

[rp]

If you fuse both sides of the HT secs then you don't need to fuse the CT or the B+ after the rectifier, right? It's one or the other?

[pdf64]

If you fuse both sides of the HT secs then you don't need to fuse the CT

Yes, there's no benefit in fusing the same current twice, though no detriment either.

[RWood]

I'm working on a suitable solution for DIYers at my usual (...slow...) pace.

I for one will be keeping my eyes open for that, so for now I'll keep the 2w k-sensor.
This post is more or less the info I was seeking... Thanks R.G.

[gingertube]

Many ways to skin the cat.

I use 10R 0.6W MRS25 Metal Film for the current sense resistors between cathode and 0V because I know (for experience) that they do blow open when you get a tube short.
I use 10R in preference to 1R to make sure they blow up before the Output Transformer.

For a CT High Voltage Winding on the power tranny fusing just the CT is "fraught".

I had a lovely vintage SA412 HiWatt come to me with shorted high voltage rectifier diodes, a destroyed power tranny and an intact fuse (of the correct value) in the CT.

For new builds I fuse both sides of the HV winding which probably means my 10R 0.6W current sense/fuses in the output tube cathodes could safely be made higher wattage and/or reduced to 1 Ohm.

Cheers,
Ian