silicon diodes as 'failsafe' feature on tube rectified amp?

[HTH]

is is ok to solder a pair of 1N4007 diodes directly on a GZ34 socket with the GZ34 still running in the socket - call it a failsafe feature in case the glass recto fails. anyone?

[dai h.]

I guess you're talking about the mod where you install diodes before the tube rect. so you still have diodes in case the tube rectifer shorts out? I don't think it would matter if they were directly on the socket or not as long as they were in the right place (in line before the tube rect.).

[stokes]

I would prefer 4 diodes as you describe,but I think with this you would not even need the tube as you would likely lose what little sag the GZ34 has.I remember reading about a similar mod that involved using a couple of high watt resistors in the circuit to give you some sag that would somewhat replicate the sag of a GZ34.

[Tom Phillips]

Ian,
I’ve seen articles about putting SS rectifiers in series with the tube. If done correctly that could keep an amp running if the tube developed a plate to cathode short. The B+ will go up of course and the bias will shift. The failed tube must be left in the socket. If the tube fails open circuit (falls out / loses vacuum), the added diodes won’t help (Unless you put them in parallel with the tube in which case the tube will be just sitting there using heater power).

Personally, I’d go with either just the tube or the SS recto and keep appropriate spares on hand. (Including fuses)

Cheers,
Tom

[Rob Mercure]

While I feel this is probably unnecessary why not a single diode "downstream" of the tube recto? If it' your filters and OT/PT you're concerned with this should take of your concerns.

Rob

[R.G.]

It's not only OK, I think it's silly to NOT do it. Here's why.

The power transformer is the single most expensive part in your amp. A shorted rectifier tube has a few outcomes, all bad, but one of them is a melted down power transformer. For under $2.00 in rectifiers and a few minutes with a soldering iron, you can completely eliminate the possibility of this happening. The solid state diodes are much lower impedance than the tube rectifier, so the tube rectifier completely covers them up sonically. It does not affect the sound.

What's not to love about that set of scenario's? Cheap protection for an expensive failure?

I don't think it would matter if they were directly on the socket or not as long as they were in the right place (in line before the tube rect.).

This is absolutely correct. It's just handy to put them right at the rectifier tube by unsoldering the PT lead, soldering on the diode, slipping a bit of heat shrink on the PT lead, soldering the PT lead, then shrinking the tubing over the diode.

I would prefer 4 diodes as you describe,but I think with this you would not even need the tube as you would likely lose what little sag the GZ34 has.I remember reading about a similar mod that involved using a couple of high watt resistors in the circuit to give you some sag that would somewhat replicate the sag of a GZ34.

This misses the issue. The issue is not to replace the tube rectifier. It's to save the power transformer if you get a bad or failing rectifier.

I’ve seen articles about putting SS rectifiers in series with the tube. If done correctly that could keep an amp running if the tube developed a plate to cathode short. The B+ will go up of course and the bias will shift. The failed tube must be left in the socket. If the tube fails open circuit (falls out / loses vacuum), the added diodes won’t help (Unless you put them in parallel with the tube in which case the tube will be just sitting there using heater power).

True - but these consequences are not cataclysmic like having your PT burn out.

Personally, I’d go with either just the tube or the SS recto and keep appropriate spares on hand. (Including fuses)

It's always a good idea to have spares on hand. However, the spares will not get your amp running again if it's got an open high voltage winding. Personally, I have retrofitted every tube rectifier amp I have with safety diodes. These are not nice-to-have things. They are the moral equivalent of fire extinguishers. They limit damage when things are already in the process of going up in smoke.

[bob p]

RG, can you explain which tube rectifier failure mode leads to the meltdown of a power transformer? The two failure modes that I envision with rectifier failure are open circuit and short circuit. Which one causes meltdown, and aren't the line fuses supposed to help in the short circuit failure mode? Thanks.

[R.G.]

RG, can you explain which tube rectifier failure mode leads to the meltdown of a power transformer? The two failure modes that I envision with rectifier failure are open circuit and short circuit. Which one causes meltdown, and aren't the line fuses supposed to help in the short circuit failure mode? Thanks.

Good question.

The first thing to realize is that a line fuse is not there to protect the PT. It's there to stop fires. Line fuses are difficult to size well for the full range of situations they have to protect, and are not high precision devices in the first place. They are the last resort, what you count on to prevent a fire when the wiring inside the box or the PT have already gone into meltdown.

The idea that a line fuse might protect something inside the box is a common misconception. They may do that, and more power to them if they happen to. But they're put there to avoid burning the house down.

Second thing to realize: while many failures can be modeled as opens and shorts, not all can. What's going to kill your PT is when a short results in higher than normal current in any winding and that being less than what trips the line fuse.

So here's a scenario: Rectifier tube shorts, but the internal melting shorts the filament pins. The increased line current resulting from a shorted 5V winding is not enough to pop the line fuse, so the 5V winding just cooks until it burns open.

Scenario 2: Rectifier tube shorts one anode to cathode, high voltage AC is dumped into the filter caps. The caps are old and are high ESR, so the current is limited enough that the AC line fuse doesn't go. High voltage winding overheats, burns out its insulation. THEN the AC line fuse pops, properly stopping the fire after the PT is dead.

Scenario 3: Rectifier shorts, sputters a spray of metal onto the inside of the tube's elements, which acts like a high power, low ohms resistor. From here, it's Scenario 2.

Admittedly, there are many rectifier failures that will pop line fuses. But not all will.

Then there's that ugly I-squared-T thing for fuses. A 1A fuse does not blow at 1A. Nor at 1.1A. A 1A fuse is guaranteed NOT to blow at 1A. The clearing time on a fuse is proportional to the energy fed to it, the square of the current through it times the time the current acts. Slow blow fuses, which are usually what is used at AC line inlets may take tens of minutes to hours to clear on modest overloads.

So the fuse may be on its way to opening... and the PT may die first.

[Enzo]

And in response to Rob, RG's scenario #2 aside for a moment, one downstream diode won't protect things before it. If the rectifier tube shorts one anode to cathode, that leaves the remaining anode and the cathode as a diode now wired directly across the PT secondary. Your diode would at least keep that AC off the filters in #2.

And that is why when someone has a shorted diode in a rectifier circuit - solid state now - I always tell them to replace BOTH diodes. The one that shorted left the other one across the PT secondary, and even if it checks OK, it was under some serious stress.

[HTH]

thanks for all the replies - all I'm really wanting to do is have the GZ34 work as normal but for the diodes to take over if the GZ34 fails.

seems like putting a 1N4007 in line with each of the PT leads is the way to go. since the GZ34 only uses four of the pins on the socket, would it not be easier (and more sturdy) to mount the diodes directly to the socket and move the PT leads over to the pin connected to the anode side of the diode?

[Tom Phillips]

...Since the GZ34 only uses four of the pins on the socket, would it not be easier (and more sturdy) to mount the diodes directly to the socket and move the PT leads over to the pin connected to the anode side of the diode?

That will work. Good idea.
Tom

[R.G.]

That will work. Good idea.
Tom

That will work... as long as you can be sure that no tube ever put in those sockets ever have a pin in those positions that connect up to those positions internally. That may be a good bet for rectifier tubes, but it has been a problem for switching 6L6 and EL34 tubes. The EL34 uses a pin that is not used by the 6L6, so designs which solder stuff to the "unused" pin go PFFFT! when the EL34 is plugged in.

It may be fine for this one limited circumstance, but it's kind of like handling rattlesnakes - that can be done in perfect safety as long as you continuously, even on bad days, remember the basics about what is lurking inside there.

[HTH]

thanks for the help everyone,

I've just taken the GZ34 out, and soldered a pair of 1N4007s directly on the valve base pins.

There does appear to be more hum now - is this 'self noise' from the diodes? maybe some caps accross them would sort it out (?)

[Tom Phillips]

... as long as you can be sure that no tube ever put in those sockets ever have a pin in those positions that connect up to those positions internally.

R.G.,
You are correct.
Are you a lawyer? Of course my reply was limited to this situation. I just don’t have the time or the inclination to qualify every response with a list of possible mistakes people could make that would cause a problem.
Cheers,
Tom

[Tom Phillips]

I've just taken the GZ34 out, and soldered a pair of 1N4007s directly on the valve base pins.

If the amp is still working at all with the GZ-34 removed then you have installed the diodes in a different manner than we have been discussing. To which pins are the diodes connected? I suspect that you effectively "converted" the amp to a SS rectifier.

There does appear to be more hum now - is this 'self noise' from the diodes? maybe some caps across them would sort it out (?)

It's true that diodes can cause some noise but not hum if the circuit is working correctly. I think that something else is going on. If my first statement is correct, then your B+ voltage could be significantly higher.
Let's get the top question sorted out first.

Regards,
Tom