How do the diodes protect the OT?

[Alan0354]

This is how I see it:

The problem is when you have open secondary, when the power tube cut off from conducting 100mA of current, the plate is going to swing up until the MOV or TVS turns on abruptly. Then it will dump the whole 100mA through the clamping device until the energy in the transformer is drained out. Even if you connect the clamping between CT to the plate, you expect to swing up +B above CT. So 450VX0.1=45W. Also, the lower the turn on impedance the clamping device, the longer it takes to drain the energy of the transformer.

Time constant is T=L/R, the lower the resistance of the clamping device, the longer it takes for the current to decay. This is the characteristic of inductor. I designed a high speed switching magnet, I used resistor as flyback load to increase R to shorten the decay time. But the voltage will swing very high. We had to use a bank of 50W carbon comp resistor to do that. Either way you use low or high impedance clamping device, you still have to dissipate all the stored energy.

I worked a lot with TVS voltage clamping in HV environment. You have to be careful in doing this as you can really burn something. If you just have one pulse event, it's ok. But if you have the user keep hitting the strings, it will heat up really fast and start burning.

Remember, it's not about what voltage it starts to conduct 1mA. It's the energy stored in the Transformer that needs to be Totally dumped off. At open secondary, the Transformer has very high inductance and whole lot of energy stored. Energy stored is 1/2 LI^2.


Using TVS was actually my first instinct. Then when I started calculating the power, then I realize that does not pan out.

[trobbins]

"The problem is when you have open secondary..."

Are you discussing a one-off event (ie. failed speaker, or pulled out speaker plug, or someone using a speaker impedance switch that is not make-before-break)?

The worst case one-off event will be when no windings have a noticeable load (eg. for current in the secondary, then a pulled speaker plug, and both tubes in cut-off). The energy stored in a winding will try and couple to other windings. If there is some load on a winding (eg. one plate conducting at that time) then energy in the transformer will couple in to that - the other windings will then see reflected voltage plus any leakage inductance effect if those windings have dI/dt in them. So need to identify when a winding has large current and no other coupling path for its energy other than to raise its voltage via dI/dt. You haven't elucidated that situation for a PP amp - only for an inductor.

[Alan0354]

To my understanding, the most likely situation is people forgot to plug in the speaker in any amp with separate speaker cabinet. That's when you burn the OT and more. I don't think it's common to burn OT if you have a load connected to the amp. I can't say about big 18" bass speaker, but for normal guitar 12" speaker, I just never ever heard about OT burn from arcing.

Yes, I only talk about open secondary. I believe all the protections are for this sole purpose.

[g1]

For the case of people forgetting to plug in the speaker, the simplest solution is the shorting speaker jack, like used in most vintage Fender amps.

[trobbins]

For unconnected secondary, then you still have two plates connected to the primary, and unless both plates are in cut-off then one plate will be providing some degree of loading. I suggest you plot a loadline for the plate being turned on, and rationalise what happens for an applied signal to the output stage.

To my mind, over-voltage faults on OTs could be from a variety of causes, not just an unloaded secondary.

[Alan0354]

I don't have a lot of data, but all these years around players, the main cause of problem is when people forgot to plug in speaker for a tube amp. Marshall is particularly prone to that because they don't short out the secondary when there is no jack plugged in. Fender short out the secondary when you don't plug in the jack, they don't have that much problem. But shorting the output presents a tremendous load on the tubes, so it's not exact a cure.

When the power tube turns off, the plate is in high impedance. Any load from the plate is likely due to arcing inside. I just don't think under normal running condition, the spike is a problem no matter how hard you drive.

Back to the inductance, if you have speaker load, the transformer transforms the impedance. The only inductance is the leakage and parasitic inductance.....which is much much lower. You don't have enough energy to do much damage. Its only when there is no load, then the OT behaves like an inductor with many henrys. As I explained before, it's the stored energy that burns. Energy = 1/2 LI^2. That's what cause the HV spike that punch through the insulation in the transformer.

Any well built OT should be able to stand off something like 2KV or more as long as you don't make it a habit of doing it all the time. It's only when you have a no load, you can have voltage spike so high until something goes wrong. It is possible you can spike up to 10KV+, that's how ignition coil work!!!

When pass a few KV, surface creepage take over and become more of a problem than insulation. Actually, insulation is really not a big problem, good insulator can withstand 400V/mil. Even the insulated magnetic wire can standoff 800V between two adjacent wires even the insulation is as thin as 1mil. Paper layer is way over 1mil, insulation is likely not a problem. But surface creepage is a totally different animal and that's likely the cause of HV failure when you pass a certain voltage. At high voltage, any contamination on the surface can be ionized and create a conductive path along the surface, corona can form and start conducting. We did a lot of test in this, if you turn off the light, you can even see the path of the surface conduction like a lighting bolt!!! I think that's the most likely cause of failure inside the OT when the spike is over a few KV.

[pdf64]

Rather than protective diodes, Ian advises the use of 'tether' resistors between OT primary legs and ground, see http://music-electronics-forum.com/t33030/

It seems a good idea to me, so I've started using this protection arrangement, rather than a resistor across the secondary.

I noticed that the diodes noticeably affect the overdriven tone and waveform, whereas primary or secondary resistor don't; this may be due to the capacitance of the diodes?

With regard to a protective resistor across the secondary, bear in mind that at bass resonance the speaker's impedance may be >10x nominal, and ~3x nominal at a few kHz; hence the use of a value of ~30x nominal for the protection resistor.
Pete

[R.G.]

Rather than protective diodes, Ian advises the use of 'tether' resistors between OT primary legs and ground, see http://music-electronics-forum.com/t33030/

It seems a good idea to me, so I've started using this protection arrangement, rather than a resistor across the secondary.

That's fine if you like the idea and the amp can stand the extra DC power spent in the resistors, heat in the case, and extra dissipation in the OT, all of which have different issues. The extra DC power could be thought of as a bleeder resistor on the first filter cap, so it may not be all that much more heat in the case.

I noticed that the diodes noticeably affect the overdriven tone and waveform, whereas primary or secondary resistor don't; this may be due to the capacitance of the diodes?

Possibly. In theory, they should not. It makes me wonder whether the diode selection was done well, and whether the flyback of leakage inductance is being caught by the protection diodes in the overdriven tone.

I've never liked the use of "catch" diodes on an OT. I prefer to clamp the primaries directly with a TVS or MOV. These have capacitance as well, but I haven't noticed tone issues with them. One bidirectional clamper across each half of the primary in a PP amp works well.

Since the voltage across each half of the primary has to be less than B+ in normal operation, Anything more than B+ is being caused by flyback. It's possible that this is something that adds its spikes to the sound you hear, but I wouldn't think so. A TVS or MOV at maybe 150% of B+ but less than about 1.5-2.0kV ought to be OK.

I designed this into the Workhorse amps having discarded diodes for these reasons.

[J M Fahey]

Rather than protective diodes, Ian advises the use of 'tether' resistors between OT primary legs and ground, see http://music-electronics-forum.com/t33030/

Ouch!!! Terrible idea.
They are dissipating a ton of perfectly safe DC all the time
If they are large enough so this does not hurt, they are too large to protect anything; id they are low enough, they are a terrible waste.
If you want protective resistors, put them in parallel with windings, so they both get the same instantaneous voltage .... that's the point.

I noticed that the diodes noticeably affect the overdriven tone and waveform, whereas primary or secondary resistor don't; this may be due to the capacitance of the diodes?

No, of course they affect waveform, hence "tone"... they clamp peaks that overshoot either ground (downwards) or 2X B+ (upwards).
The wave becomes more traditional squarewave.
It loses some bite.

I noticed that long ago; those who use them (Fender/Peavey/Ampeg/etc.) must have too, I guess they also have "boring Engineers" in the payroll , must even (gasp !!!! ) scope them, , but don't talk much about that, because protecting an expensive OT is a good thing and neurotic Musicians will get a nipper and clip them off the very microsecond some Guru tells them that it's the path to "Guitar Godness" or whatever.

In fact, I introduced natural overshoot in my transistor amps, go figure , for that same reason.

As of MOVs, they are interesting for one reason: diodes are too strict, they clamp everything 1V below ground or above 2X +B ; while you can select MOVs to allow, say, 100V or more either way, which is still well within safety levels.

The normal overshoot I see on overdrive and speaker loads is, say, 10% to 20% above the "squarewave" peak, not more.

Maybe our friend Loudthud will overdrive a tube amp at a couple selected frequencies (say, 100/400/1500/4000Hz) and show overshoot on the secondary, the one which goes to the speakers and we actually hear.

And as I said, what's noticeable is a little extra "bite".

By the way, the exact opposite of those nicely rounded waveforms so many draw to show "tube distortion" .
Oh well.

[Alan0354]

Rather than protective diodes, Ian advises the use of 'tether' resistors between OT primary legs and ground, see http://music-electronics-forum.com/t33030/

It seems a good idea to me, so I've started using this protection arrangement, rather than a resistor across the secondary.

I noticed that the diodes noticeably affect the overdriven tone and waveform, whereas primary or secondary resistor don't; this may be due to the capacitance of the diodes?

With regard to a protective resistor across the secondary, bear in mind that at bass resonance the speaker's impedance may be >10x nominal, and ~3x nominal at a few kHz; hence the use of a value of ~30x nominal for the protection resistor.
Pete

I was going to comment about the diodes might change the sound. I did see the overshoot talked about when I really cranked up the amp even with resistor load. The spike is not very bad and I think it's safe. That might give a certain characteristic of of the OD sound. You put the diodes, you definitely affecting the overshoot and should change the sound.

This diode protection idea has been around for many years....as Prosonics is a very old amp over 20 years. There must be a good reason why most people don't put it in. I am going to experiment on the sound of this soon. For now, I think I am happy to do it the Fender way by shorting the output if no speaker connected.

[Alan0354]

I've never liked the use of "catch" diodes on an OT. I prefer to clamp the primaries directly with a TVS or MOV. These have capacitance as well, but I haven't noticed tone issues with them. One bidirectional clamper across each half of the primary in a PP amp works well.

Since the voltage across each half of the primary has to be less than B+ in normal operation, Anything more than B+ is being caused by flyback. It's possible that this is something that adds its spikes to the sound you hear, but I wouldn't think so. A TVS or MOV at maybe 150% of B+ but less than about 1.5-2.0kV ought to be OK.

My concern about TVS or MOV is in post #16. It is so common for people to strike the guiter, turn up the volume when they forgot to plug in the speaker. Takes a few seconds to remember to check the speaker connection. As I explained in #16, tremendous power would be dissipated in the TVS, it is very possible to burn the TVS.

From my experience, the P6KExx and 1.5KExx burn shorted......for good reason, they want to protect the circuit from over voltage even on it's last breath!!!! Must be designed this way. So this mean if you hit the guitar hard enough before you realize you forgot the speaker, you might burn the TVS and render the amp useless.

The power dissipation also applies to MOV even though the impedance is higher and they clip softer. One way or the other, the protection has to dissipate all the energy in the Transformer, there's no other way around this.

[Leo_Gnardo]

I was going to comment about the diodes might change the sound. I did see the overshoot talked about when I really cranked up the amp even with resistor load. The spike is not very bad and I think it's safe. That might give a certain characteristic of of the OD sound. You put the diodes, you definitely affecting the overshoot and should change the sound.

When I first encountered the suggestion of putting in "anti flyback diodes" I did do some careful listening. Yes I heard a slight difference, some of the nastier overtones went away with the diodes in place. Whether to risk your tubes & transformer, vs. retaining all the "natural" deep overdrive distortion including overshoot spikes is a personal decision. Regarding tone, hardly anyone besides the guitarist would know or care.

[Alan0354]

When I first encountered the suggestion of putting in "anti flyback diodes" I did do some careful listening. Yes I heard a slight difference, some of the nastier overtones went away with the diodes in place. Whether to risk your tubes & transformer, vs. retaining all the "natural" deep overdrive distortion including overshoot spikes is a personal decision. Regarding tone, hardly anyone besides the guitarist would know or care.

I just finished comparing sound of a Classic Tone Marshall OT to Classic Tone Vibrolux. I'll try the diode and see whether I can hear any difference. I'll update later.

My whole thread is for testing the two OT so I can switch back and fore hot. I did it and posted the result. It is very disappointing that I cannot hear any difference it all different setting I can think of. This together with the sound clip I posted on OT comparison, I really question why people worry so much about different brands components and all the fuzz. Why worry all the minor difference when changing value of some of the resistors and capacitors in the signal chain can drastically change the sound of the amp. It seems like people just use one single amp, leaving the circuit totally unchanged. Then just change one component to a different brand and claim this brand is better or not.

[trobbins]

Alan, your conclusion imho is not justified. You have only broached inductor protection experience, and we've all had experience with the LHC so can appreciate the level of quenching protection needed for that application. You should really go through what happens in a PP circuit, and address the influence of all windings, and the loadlines of the valves, to better appreciate what actually is being dumped in to a protection device like a MOV for that application.

I use small 7mm disk MOVs, and to my mind they far exceed the thermal stress that could be handed to them. Small MOVs have quite low capacitance, which is effectively negligible to frequency response of even a hi-fi amp, especially if 2-3 are placed in series. If really needed, the MOV capacitance can be tuned to the OT first SRF and a dampening resistor added - aka zobel network.

[Alan0354]

Can you explain what is the difference between the inductance of the OT with no load in a PP circuit compare to inductance of an inductor? Because I am sure any effect of inductance will dump the energy on the MOV.

I never complain about the capacitance as the extra 50pF is not going to hurt anything. It is common to put capacitance on the plate of the power tube. SF Fenders have 2000pF on the plates.