Is rectifier tube arcing a problem?

So far so good. I will keep observing. Maybe the amps are just more demanding on the tubes, and some of the old tubes I tried are just bad.

@gte357s 

The 5U4, being a directly heated rectifier, is a far simpler tube than a KT88 or like; I really doubt the ones you had were bad. These tubes are generally pretty reliable because they are so simple! Keep in mind that tubes, especially older NOS tubes, can be quite forgiving when their specs are exceeded as long as those specs are not exceeded for too long.

Because this is a directly heated rectifier, one way the tube can fail due to arcing is the filament can open up- causing the tube to simply die with no (apparent) fireworks. This is because its two cathodes serving the two plates are in series.

Given your prior experience thus far I expect if you didn't make the changes in the power supply previously suggested that you'll be seeing this problem again.

Say from 220uf to 100 instead of going all the way down to 47uf.

@fiesta75 

You work out the timing constants. The amp likely does not have a timing constant in the audio circuit that is less than 5Hz since the output transformer won't go down anywhere near that. So it will not be able to modulate the power supply if the power supply has timing constants of below 5Hz. 220uF is way overkill; I don't know of a tube rectifier that will survive that.

Not only that and more importantly, the load on the power supply is constant owing to the circuit is class A- it draws the same power at idle as it does at full power. This makes the power supply requirements a lot simpler. That is why 47uf will be fine. I have an amp that makes the same power and it has only 30uF in this position. The power supply voltage dominates the equation when you're talking about energy storage to do work (like driving a loudspeaker). The formula is:

Work = 1/2 (C x Vsquared)

That is why tube amps don't need nearly the power supply capacitance as solid state amps do.

thank you very much. A couple more questions:

1) is it able to tell from the diagram if there other rectifier that can be used and handle the current drawn by the 220uF capacitor?

2) is it necessary or recommended to change the other two capacitors? If yes, what values should I use?

3) comparing to adding diodes, in your opinion, which way will yield better sound?

@gte357s 

There won't be another tube rectifier that will work better. The problem is that this circuit is designed to expect the voltage drop that exists with the 5U4 at the current at which its operating.

I gave you a range of capacitor values. But to be clear:

replace the 220uF units with 47uF. Replace the 100uF units with 22uF.

Once this is done you'll find additional diodes to be moot. They may also cause commutation noise by interacting with the inductance of the power transformer winding, which can manifest as an annoying low level buzz. If it were me I wouldn't bother and instead work on correcting the problem.

Yes, mine also has 220uF capacitor, and yes, the arcing only happens when the amp is turn on. After than, and if the fuse doesn’t blow, the amps sound fine.

Nailed it!

I would replace the capacitors after the choke with 50uF devices rather than the 220uf units. If anything you might hear a slight improvement since the smaller cap might perform better at high frequencies, if you install ones of similar quality. If you do this there may be no need to install the solid state rectifiers, although the caps in succeeding legs of the power supply are gross overkill (contributing to rectifier failure) as well. You could run 1/5th the value and be fine.

If it is, that 220uF value seems really high to me! An SET is a class A device so the load on the power supply is constant. If you cut all those values in half I think you’ll find that the rectifiers hold up just fine.

This appears to me to be a classic case of too much filtering (the filter caps need to have a time constant lower than that of the amp and in an SET that isn’t hard...)- when the amp is off, the voltages in the power supply are drained off. When you turn the amp on, as the rectifier warms up current flows to charge the caps. There is a certain point on the exponential charging curve (this is the charging curve of any capacitor) where the maximum current of the 5U4 is exceeded. In due time the tube, which is otherwise pretty forgiving, arcs. If I am right about this the arcing happens as the amp is warming up.

min the diagram, I need to add two diodes per tube, am I correct?

Yes.

I agree something is up in the amps- the 5U4 should easily handle the load all on its own. It sounds like something is causing an excessive load during warmup, which is a classic symptom of the filter capacitance at the output of the rectifier being too large. Both doing the same thing is not coincidence! What is the capacitor value at the output of the rectifier tube? There is no schematic in the manual...

A 5U4 should handle that load just fine. I think you've had bad luck with tubes... Is this happening in both channels? Do both channels sound the same?

@gte357s 

One thing that can cause rectifier tubes to arc is too much capacitance at the input of the power supply. What amp are you using?

I've seen a lot more arcing failure with Russian made rectifier tubes. Depending on the amp you may have to stick with NOS tubes to get reliable service.

In some cases its possible to run solid state rectifiers. A bit of attention has to be paid to the amp circuit since doing so will raise the B+ voltage which may not be a good thing. But if you have good rectifiers (like HEXFREDs) they can be nice and quiet. Amps that use solid state rectifiers usually sound more neutral on account of less power supply sag when more power is required. Tube rectifiers are certainly romantic but other than noise they really don't have a place in a hifi power amp; they are great for blues guitar though.