Fuse on both windings of a center-tapped secondary

[greekie]

Great discussion - what of it that is on topic anyways!

I decided to construct my amplifier as such:

I fused the HT and heater taps with two fuses each, as they are both center-tapped. I've then fused the bias-winding (250mA winding) with a 250mA slow-blo fuse. Just after this fuse is a circuit consisting of a couple of capacitors, diodes, resistors, and a transistor that turns on a relay when the power is turned on. This relay then closes and connects the fused HT-taps with the rectifiers and the rest of the power supply. This way, if the bias winding fuse blows, the relay won't energize, and I won't be in a situation with a shorted bias winding and full HT. I placed a small diode with a largish resistor in parallel with the fuse, so the failure is easy to locate.

Thanks for all the help you guys!

[greekie]

Oh, and the IEC-socket with built in fuse is what I use. As well as building my own eyelet boards, I add eyelets for PCB-mounted fuse-sockets. This is super easy and looks great as well!

[defaced]

The bias winding thing gets tricky. If you lose a bias winding through a short, the output tubes go critical.

What about a relay in series with the bias winding which controls the HV? If the bias winding fuse goes, the relay opens and the HV goes too (or even the mains voltage, assuming coupling between the AC on the contacts and the coil isn't an issue). High sensitivity relays (especially higher voltage ones) take significantly less current than their low voltage counterparts, so I think the extra current for the relay coil could fall into the "insignificant" category.

[R.G.]

What about a relay in series with the bias winding which controls the HV? If the bias winding fuse goes, the relay opens and the HV goes too (or even the mains voltage, assuming coupling between the AC on the contacts and the coil isn't an issue).

That's definitely the kind of thinking that this problem needs: no bias, no current flow in the output tubes, by a means that's independent of the output tube setup.

High sensitivity relays (especially higher voltage ones) take significantly less current than their low voltage counterparts, so I think the extra current for the relay coil could fall into the "insignificant" category.

I would normally say yes to that. However, I'm a bit paranoid about output tube bias voltages. Some designs rely on the grid leak resistors being 100's of k ohms.

I'd probably want to go with some scheme that used a MOSFET's gate-source threshold voltage as the on-off criteria, as this can be 3-5V, scalable to whatever you want with a resistive divider, and with a current drain only determined by the resistor divider. That lets me be pretty sure the bias voltage sensor doesn't overload the bias voltage generator. The MOSFET would turn on when the raw bias supply was more than X, and that would enable something that let current through the output tubes.

That something could be a relay, but could also be a Big Freaking High Voltage MOSFET (BFHVM) that is either open (for "off") or saturated to under an ohm for operation.

It's about here that the Silico-Phobes start running for the exits. I'm pretty sure I can make a MOSFET switch on and off maybe 10⁴-10⁶ more times than a relay without failing.

[Steve Conner]

Morgan Jones has some hi-fi amp designs with loss-of-bias interlocks like this. In his monstrous "Crystal Palace" amp, he has a separate big PT just for the HT supply to the output stage. The mains supply to it is switched by a solid-state relay, powered off the bias supply. "Simples" as the meerkat says.

While we're talking heresies, my own personal "improvement" to the tube amp would be screen current limiting, using a
regulated screen supply with a limit built in. I believe it can greatly improve the chances of surviving an open speaker load. For silicophobes it could be as easy as a light bulb or two in series with the screen supply, and for extra vintage points have it shine on a Vactrol that compresses the input signal.

The trouble is, once you start getting innovative, the elephant in the room is, why are you even using tubes? It's a slippery slope that ends in turning into a plugin developer for Line 6.

[R.G.]

While we're talking heresies, my own personal "improvement" to the tube amp would be screen current limiting, using a regulated screen supply with a limit built in. I believe it can greatly improve the chances of surviving an open speaker load.

I like it. Good thinking. This could also be as simple as a floating current limit.

The trouble is, once you start getting innovative, the elephant in the room is, why are you even using tubes? It's a slippery slope that ends in turning into a plugin developer for Line 6.

I think the elephant is already in the room.

As I used to get ugly replies to in my first career, it's a matter of objectives. The only reason to use tubes at all today is to hear the tubes. Given that good tubes are expensive and not easy to find, it makes sense to preserve and defend the ones you can get so you get the best use of them. In my mind that translates into identifying conditions which are both hard on the tubes and do not contribute to the sound, and making those conditions easier on the tubes. The slope has a nice, grassy, flat space on it - if you're not listening to the tubes for their sound, you have opted out of the ...er, room??

As an example, when output tubes have lost their bias, they are no longer making sound to listen to. It logically follows that whatever you do to them to prevent damage in that condition is not programming plugins for Line 6.

Anyway, it seems logical to me.

[imaradiostar]

I'm confused. How does the relay/mosfet save the power transformer if one of the two diodes in a FWCT arrangement fails? Does it cut the primary power? If so how does the amp ever turn on? A "start" switch like a Hammond organ?

Perhaps we can no longer see the forest...

I think a mosfet/bias sense is a great way to preserve output tubes. There are a few simple arrangements for this in Merlin's book and an example or two on the Champ Electronics web site.

I still prefer simplicity in lower power designs. I'd work to save a transformer before I'd work to save an EL84 any day. Weren't EL84's designed to be toasted?

jamie

[big_teee]

I like the Idea of the additional Secondary fuses.
Beyond that all the relay talk is getting way to Gifted and Talented for me.
I like the Keep It Simple Stupid,(KISS) Approach.
Just Tubes No Solid State, and No relays.
But, thats just me.
B_T

[Steve Conner]

Well, there's a lot to be said for the single point of failure approach. Just use good quality parts and build it really well, then you won't have to worry about what happens when such-and-such part shorts out, because it won't!

I've only ever seen one schematic showing a centre tapped rectifier with two fuses, in an ancient ham radio handbook. It always struck me as a good idea.

I once tried a protection circuit that needed a Hammond organ-style start switch. It seemed like a cool gimmick at the time, but it ended up being nothing but a pain.

The problem with RG's approach is that in classic tube circuits, the tubes are often being driven well beyond ratings and tend to have a nasty, brutish and short life by design. Analysing how a cranked Plexi sounds the way it does, leads to the same uncomfortable conclusions as pondering where meat comes from.

[Gingertube]

Guys,
The one thing we can take from this thread (so far) is the need to fuse both sides of a centre tapped high voltage winding. A single fuse in the centre tap just does not cut it. Think of both rectifier diodes going short and you can see why the single fuse in the centre tap will not give you the protection you need.
Cheers,
Ian

[R.G.]

I'm confused. How does the relay/mosfet save the power transformer if one of the two diodes in a FWCT arrangement fails?

It doesn't, not meant to. The relay/MOSFET is a particular solution to the "loss of bias" situation. It's not meant to be a panacea.

Does it cut the primary power? If so how does the amp ever turn on? A "start" switch like a Hammond organ?
Perhaps we can no longer see the forest...

But the forest is such a target-rich environment (to appropriate some fighter-pilot jargon).

It is conceptually easy to sense a number of things and then cut the primary power. It can look like it starts exactly the way the amp does today, with an on/off switch and a standby switch. Inside, I'd use a secondary power transformer that makes a watt or two of DC to run just the supervisory circuit, and have it then take care of all the rest.

...oops, another flotilla of tube-amp guys just headed for the door. I can hear them "AGGH!!! Too complicated! Has silicon in it! Can't cope!" while they read this on a computer with a dozen microprocessors in it. Sorry. I need to remember to speak softly.

I have thought about a comprehensive protection circuit suite before, but it gets more complex (although no more complicated that some pedals that DIYers build for fun). The tube-amp guys have a strong tendency to drop out as soon as it gets beyond "source followers" and near "standby switch". They self-select out of it and won't listen any more.

I think a mosfet/bias sense is a great way to preserve output tubes. There are a few simple arrangements for this in Merlin's book and an example or two on the Champ Electronics web site.

I still prefer simplicity in lower power designs. I'd work to save a transformer before I'd work to save an EL84 any day. Weren't EL84's designed to be toasted?

Good choice. The power transformer is the single most expensive part in a tube amp, followed by the output transformer.

Thing is, they're both tough. Fuses in each winding section that can be overcurrented are probably enough to keep them alive; as witness the long-term survival of many transformers with no protection at all.

The problem with RG's approach is that in classic tube circuits, the tubes are often being driven well beyond ratings and tend to have a nasty, brutish and short life by design. Analysing how a cranked Plexi sounds the way it does, leads to the same uncomfortable conclusions as pondering where meat comes from.

Yeah, there is that. It's the classic thing amp techs hear. "It really sounded great just before it quit working. Can you fix it and make it sound like that all the time?"

[imaradiostar]

Make it sound more "explodey."

I've read more than one article (Eric Johnson was one) that mentioned how 100 watt four input marshalls sound better than 50 watters because they sound like they're more likely to blow up- that ragged edge sort of tone. I think that's because statistically, with higher voltages and insufficient screen dropping resistors, they ARE more likely to blow up!

As you said, RG- how do you capture the sound of impending failure without the eventual failure?

jamie

[Gingertube]

The sound difference between the 50W and 100W is directly related to the fact that they run 4 output tubes rather than 2 and hence have double the gm (that is swing double the current through the output tranny for the same applied signal voltage). I have seen an article which claimed that Angus Young (ACDC) preferred the 50W - mostly a matter of the players taste.

I did'nt pay a lot of credence to this (change of sound with 4 output tubes) until I recently built an amp using a quad of 6V6. It still has some of that 6V6 "grind" but is much "smoother" than a single pair.

It might also be due to the fact that with 4 tubes, statistically you are likely to have better balance in the output tubes.

It might also be due to where the Class A to ClassAB transistion occurs (as one side cuts off). Personnaly I can't pick it but it is commonly accepted that good players can. With the 100W this will happen at twice the power level as compared to the single pair of tubes.

So the claim has merit.

Cheers,
Ian

[Gaz]

I was looking through Ultimate Tone III to see examples of his secondary fusing values, and they don't really seem to make sense because of their inconsistency:

Plexi 4 x EL34: 750ma fuse in each secondary winding
Bassman: 2 x 6L6: 750ma " ", tube rectified
JCM800: 4 x EL34: 750ma " "
Bulldog: 2 x EL34: 750ma " "
Portaflex: 2 x 6L6: 500ma in CT only (Why?)
AC30: 4 x EL84: 750ma in CT only, tube rectified (Why not in both secondaries?)
Bassmaster: 2 x EL34: 750ma in each secondary winding
Custom Special: 4 x EL34: 1.5A " "
Guitarmate: 2 x EL84: 500ma " "
AC15: 2 x EL84: 500ma in only one secondary, FWB (Merlin explains this in his book)

And the heater secondary values for the same projects:


Plexi 4 x EL34: 10A in each secondary leg
Bassman: 2 x 6L6: 6A " "
JCM800: 4 x EL34: 10A " "
Bulldog: 2 x EL34: 6A " "
Portaflex: 2 x 6L6: 4A in only one leg (Why?)
AC30: 4 x EL84: 5A in each secondary leg
Bassmaster: 2 x EL34: 5A " "
Custom Special: 7A " "
Guitarmate: 4A " "
AC15: 2 x EL84: 3A in only one leg (Why?)

So, as usualy Mr. O'Connor leave me with more questions than answers....

[imaradiostar]

The 4 versus 2 output thing is interesting. I didn't mean to literally say that they sound different because of different voltages. It's logical that there are other differences too.

Ian, I would have thought the increase of gm is swamped by the change of transformer plate impedance- twice the current gain driving half the load.

I have a small amp in the works (for over two years now, really) with 4x5686 tubes. I was hoping to capture some of that 4 output tube magic with only about 20 total watts of output.

jamie