What Happens To A Tube Amp With No Load?

[khutch]

I'm an engineer and I am building a tube amp. I don't need to build a tube amp, I don't need to have a tube amp, I am doing it simply because I want to. But I believe in building things as bulletproof as possible. So I would like to know, in extreme detail, what happens inside a tube amp when you hit it with a signal and no load is connected. If I know what it is that causes unloaded tube amps to fail then perhaps I can fix that.

So if you know, let's talk.

I've been doodling around with tube amps in Spice simulations for a few months now. When I pull the load from one I can see things happening that could/would be bad. The things I see appear to have obvious solutions but from what I read here and everywhere else no one builds them that way. So maybe there are other things going on that are not so obvious.

I'd like to hear from anyone here who knows, what it is that goes wrong. I am not going to start the conversation by telling you what I see in simulations because I want you to tell me everything you know of that goes wrong. I don't want to focus the conversation on any one thing, I want to know everything.

If anyone is game....

 

[Son of Wobble]

I'm an engineer and I am building a tube amp. I don't need to build a tube amp, I don't need to have a tube amp, I am doing it simply because I want to. But I believe in building things as bulletproof as possible. So I would like to know, in extreme detail, what happens inside a tube amp when you hit it with a signal and no load is connected. If I know what it is that causes unloaded tube amps to fail then perhaps I can fix that.

So if you know, let's talk.

I've been doodling around with tube amps in Spice simulations for a few months now. When I pull the load from one I can see things happening that could/would be bad. The things I see appear to have obvious solutions but from what I read here and everywhere else no one builds them that way. So maybe there are other things going on that are not so obvious.

I'd like to hear from anyone here who knows, what it is that goes wrong. I am not going to start the conversation by telling you what I see in simulations because I want you to tell me everything you know of that goes wrong. I don't want to focus the conversation on any one thing, I want to know everything.

If anyone is game....

The Ampeg PF50-T is a 50 watt tube amp that can be used without a load. So if you can figure out how that manages it, maybe the concept can be scaled for 300-plus watts?

 

[Stumbo]

@Passinwind @agedhorse may stop by to give you a few tips or references.

 

[dbase]

 

[BassmanPaul]

There are ways to partly protect a tube amp from a no load situation. Fender, and some other North American manufacturers have the output jack short the tap to chassis when nothing is inserted. In the UK a resistor of 50-100Ω 5W WW was wired permanently across the load so that the OPT is still loaded without a cable plugged in.

I suspect that the Ampeg amp has a dummy load resistor to allow operation without a cabinet connected.

What exactly happens is not coming to mind, it has been a long time since I thought about this.

It involves voltage spikes that damage the output tubes and OPT insulation leading to shorted turns. One way to alleviate the spikes is to have three or four reverse connected silicon rectifiers connected from each plate to ground to protect both the tubes and OPT.

Hope this helps.

 

[throbbinnut]

The flyback nature of the output transformer primary means one end tries to go to infinity volts, so something arcs before it gets that high.

 

[throbbinnut]

Here's why it happens - if you put an open circuit as a load line on these 6L6 pentode tube curves, it's a horizontal line. So if you are sitting at -20V grid bias at 450V plate, and you swing the grid down to -25V, imagine how far to the right (plate voltage) you'd have to go to intersect that horizontal line with that gradually sloping-upward -25V bias line - you'd be way into the thousand volts range off the chart, and your output transformer would arc over and be fried. Now if you short the output, that's a vertical load line, and if you swing grid voltage around, the current goes nuts, but the voltage doesn't. That's the reason for the shorting output jack in a push-pull pentode amp like a Bassman.

And just to show another setup, here's the triode-wired curves. Note that if you do the same horizontal line, you don't have to swing to thousands of volts to get to the more negative grid voltage. This is because the triode has a lower internal impedance (chart bias lines sloping upward) versus the pentode's higher internal impedance (chart lines basically running almost horizontal).

Hopefully that explains it. Also, if the load was just resistive, you wouldn't have that insane voltage swing up above B+. It's the nature of an inductor to do whatever it has to to keep current the same, and in the case of the pentode trying to cut off current, the inductor voltage goes as high as it needs to to keep current flowing, or at least tries to.

Chris

 

[bluesfordan]

View attachment 2996435

dang, somebody beat me to it

 

[mesaplayer83]

there are ways to partly protect a tube amp from a no load situation. Fender, and some other North American manufacturers have the output jack short the tap to chassis when nothing is inserted.

Was about to post that...

 

[amper]

I'm an engineer and I am building a tube amp.

Can I just ask, if you are an engineer, why are you asking this question? You should already know the answer, unless you aren't an electronics engineer. Lack of a load on a transformer-coupled tube amplifier will cause high voltages to develop in the output transformer and it will meltdown and/or short, and fail, likely taking your power tubes with it, and if you are lucky, that's all that will happen.

 

[Bugeyed Earl]

Can I just ask, if you are an engineer, why are you asking this question? You should already know the answer, unless you aren't an electronics engineer. Lack of a load on a transformer-coupled tube amplifier will cause high voltages to develop in the output transformer and it will meltdown and/or short, and fail, likely taking your power tubes with it, and if you are lucky, that's all that will happen.

Maybe he's an engineer in training, or maybe he's one of these:

 

[rickdog]

Vacuum tube electronic design is becoming the lost wisdom of the ancients. Anyone who got an EE degree in the last, say, four or five decades, studied transistor circuits.

With no load on the secondary, the primary of the transformer acts as an inductor. Changing the current through an inductor results in a voltage opposing the change, proportional to the inductance and the rate at which the voltage changes. An output transformer has a large inductance, and thus generates large voltages as the output tubes try to change the current.

 

[Old Garage-Bander]

Can I just ask, if you are an engineer, why are you asking this question? You should already know the answer, unless you aren't an electronics engineer. Lack of a load on a transformer-coupled tube amplifier will cause high voltages to develop in the output transformer and it will meltdown and/or short, and fail, likely taking your power tubes with it, and if you are lucky, that's all that will happen.

Maybe he didn't go to a Tube School? Not much call for that outside of geetar amps and high power broadcasting these days. Oh and maybe some medical electronics which is a lot like broadcasting only the equipment is cleaner, always new, they pay the techs a lot better and no one calls you in the middle of the night to say the MRI machine just took a dump.

Or maybe he drives the choo-choo train?

I try hard not to play engineer even though some people I work for insist on calling me that.
I ask them if they see an engineering degree hanging on the wall, if they say yes, they're in the wrong place.

Although, most real engineers (not the choo-choo ones, or people who get called that at work against better judgement) that are not electrical, have had enough book learnin' and have enough brains, that with just a bit more info, they can figure a lot of it out. They all get a bit of Ohm's law along the way, and that's pretty much what's needed to understand why good tubes go bad.

 

[Son of Wobble]

Vacuum tube electronic design is becoming the lost wisdom of the ancients.

I guess that can happen quite quickly, huh? An audio maven pal of mine, 45, graduated from a school where he learned how to make records with "tape and a razor blade," as he put it. Within 18 months of his graduating the program, he says wryly, the industry transitioned to digital and all he'd learned was largely useless.

I went to UCLA film school in the 80's and we cut 16mm on upright moviolas. Ruinously expensive and a royal pain in the ass.
Sometimes you'd find yourself in a trim bin trying to located a frame! Non-destructive digital is where it's at, the kids today got it easy with their DSLR and D class.

 

[Mr. Foxen]

Generally what happens is the voltage spike arcs over the path of least resistance, causing other stuff to go wrong and the fuse to blow. Often that path is inside a valve, but it can also explode filter caps (usually when old), or go across a valve base, and you wind up with something a bit like this, although this was from a failed grid resistor.

There isn't really a way to prevent it, good discipline, shorting connector on the amp end helps, but a lead falling off the cab end, including internally to the cab, can do it. But you can arrange a most convenient path of least resistance to get that fuse blown before anything burns up, which is known as a spark gap.

 

[beans-on-toast]

The amp should be designed to withstand being operated without a speaker cabinet connected. This is accomplished in different ways. Add this requirement to your design specification.

 

[agedhorse]

And just to show another setup, here's the triode-wired curves. Note that if you do the same horizontal line, you don't have to swing to thousands of volts to get to the more negative grid voltage. This is because the triode has a lower internal impedance (chart bias lines sloping upward) versus the pentode's higher internal impedance (chart lines basically running almost horizontal).
View attachment 2996493

Hopefully that explains it. Also, if the load was just resistive, you wouldn't have that insane voltage swing up above B+. It's the nature of an inductor to do whatever it has to to keep current the same, and in the case of the pentode trying to cut off current, the inductor voltage goes as high as it needs to to keep current flowing, or at least tries to.

Chris

Are you sure you don't have this backwards? If you look at the V-I curves, the shallower slope correlates with a higher impedance.

Can I just ask, if you are an engineer, why are you asking this question? You should already know the answer, unless you aren't an electronics engineer. Lack of a load on a transformer-coupled tube amplifier will cause high voltages to develop in the output transformer and it will meltdown and/or short, and fail, likely taking your power tubes with it, and if you are lucky, that's all that will happen.

This is a topic that was not well covered in engineering, even back in the day.

It's also a topic that's full of mis-informaton and bad advice from otherwise well intentioned "experts".

There are several contributing reasons why this happens. The first reason is that even driving an inductive load (transformer, unloaded) with higher frequency signal, the faster the rise/fall time, the greater POTENTIAL for ringing (especially if the amp is only marginally stable under global feedback conditions).

The bigger and more concerning cause is clipping the amp while unloaded, which causes the global feedback around the transformer to open up. When the amp comes out of clip, the loop attempts to re-close. These are prime discontinuities which excite the transformer's magnetic field, which upon collapsing unloaded can generate a voltage.

It's not purely an unloaded transformer that's the cause, because if you look at the formula for the inductive spike, V = L x di/dt, but since there is little current flowing in an unloaded transformer, this is not the sole cause. Clipping can create additional currents with a high discontinuity at the corners (di/dt), and with no load on the secondary, the ringing reflects back to the primary rather than being absorbed by the load.

This subject is a lot more complicated than most simplistic arguments consider.

 

[Ross W. Lovell]

I'm an engineer and I am building a tube amp. I don't need to build a tube amp, I don't need to have a tube amp, I am doing it simply because I want to. But I believe in building things as bulletproof as possible. So I would like to know, in extreme detail, what happens inside a tube amp when you hit it with a signal and no load is connected. If I know what it is that causes unloaded tube amps to fail then perhaps I can fix that.

So if you know, let's talk.

I've been doodling around with tube amps in Spice simulations for a few months now. When I pull the load from one I can see things happening that could/would be bad. The things I see appear to have obvious solutions but from what I read here and everywhere else no one builds them that way. So maybe there are other things going on that are not so obvious.

I'd like to hear from anyone here who knows, what it is that goes wrong. I am not going to start the conversation by telling you what I see in simulations because I want you to tell me everything you know of that goes wrong. I don't want to focus the conversation on any one thing, I want to know everything.

If anyone is game....

I know it can have disastorous consequences, but I've been lucky.

Happened when I played guitar years back. Equipment was set up by me and I grabbed my guitar.

Hit a chord and Nada, bubkis. Checked my signal chain everything was fine.

Turned up the Master on my Hiwatt a little, still nothing, though I thought I heard a rattle.

"Dimed it" as the kids say nowadays and the rattle was definitely louder.

My drummer then informed me that every time I did that, the tubes glowed and then faded with the attack and delay of the chord.

Always better to make sure your cab is still plugged in.

That rattle, near as I could tell was something inside the tubes making the noise.

Still run the amp, the tubes were fine for the next 5 years.

I was very, very lucky.

Replicate Hiwatt's circuit or at least determine why it didn't go poof on me.

 

[LowActionHero]

 

[macmanlou]

The Ampeg PF50-T is a 50 watt tube amp that can be used without a load. So if you can figure out how that manages it, maybe the concept can be scaled for 300-plus watts?

This. If you can get the schematics for the PF50T or PF20T, you'll probably be able to figure it out.

One of the reasons I got a PF20T is to use it as a preamp with a GK MB500. The "no speaker load needed" feature helps to make that happen.