• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Shorting output transformer secondary

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I know that it's standard practice on guitar amps to have shorting jacks so that if the speaker is pulled and a no load situation occurs, the OT doesn't self destruct.

I thought i'd try it on a SE amp just to make sure it was safe!
With enough drive to create a square wave output into an 8 ohm load voltage across the cathode resistor dropped from 24V to 22V.
Tried the same into a short circuit and the voltage increased to 26V. Thought i'd share this info to dispel some apocalyptic fears propogated here. I only did it cos i was pretty sure no harm would be done.
If you short the output the primary sees zero ohms reflected back.
At first glance this looks like a damaging short but at DC the primary is virtually a short anyway and that doesnt cause a problem. Wit ha secondary short the primary impedance will just be DC resistance of primary winding.
It probably is a bit more complex if there is a large audio signal present.
Last edited:


Joined 2003
Paid Member
Shorting the output prevents arc-over damage in the OT. (wakculloch was being reckless driving a SE tube amp to full square output; it's how Randy Bachman killed a lot of OTs.)

With shorted load, the exactly-biased SE amp will take the SAME DC power at idle or at full audio input. (wakculloch's amp rose 8% because bias is never exact but also because tubes are curved.)

What is interesting is to measure the plate temperature. A cheap IR thermometer will give you a clue. Read the idle temperature. Run it to full signal with shorted load, no big change. Now load it proper and bring it to full clean Sine output. Plate will -cool-. Drive proper load to Square wave, plate will cool more.

The nicest thing you can do to an SE amp is PLAY IT LOUD.

On dummy load and an old tube, you hear it creak as it cools-down.

I had SE 6550 idled at 39 Watts. At squared output in proper load, I found 22W in load and 22W in plate, plate dissipation near half of idle.
The cathode resistor voltage isn't the problem, it's the AC across the primary and at the place with no connection to the secondary. It generates high spike voltages at transients (the rising and falling edges of a square wave) that can arc over in the transformer or even in the output tube. Look at the plate connection with an oscilloscope, preferably with a probe and scope gain setting that can handle 1000+ volts, and you should see the big square-wave voltage spikes with no load.

Start with an SE pentode amp that uses fixed bias and 50 mA Quiescent.
And with a load, RL, that gives 200mA when the grid to cathode volts = -1V.
Keep in mind that in this case (shorted secondary), the new RL is nearly as low as the Primary DCR.
That can occur if you apply a square wave signal level that is barely less than drawing grid current . . .
The square wave will cause the tube to be OFF for 50% of the time (0 mA).
But the same square wave will cause the tube to be ON 50% of the time (200mA).
In my book, that is an average current of 100mA, or 2X the 50mA quiescent current.
Hotter plate, yes?

All generalizations have exceptions.
Start with an SE pentode amp that uses fixed bias and 50 mA Quiescent. And with a load, RL, that gives 200mA
Single ended implies quiescent current is nominally 50% of full-load current - otherwise you get premature clipping. 50 mA quiescent and 200 mA peak puts you in class AB, not workable for a single-ended output stage!

In your example, you'd either need to bias the pentode to 100 mA quiescent (peaks swinging from 0 to 200 at the threshold of clipping), or choose a load to limit peak current to roughly 100 mA (50 mA quiescent, peaks of 0 mA and 100 mA at the threshold of clipping.)

Above numbers are simplified and assume perfectly linear pentodes, which don't exist. They also assume a perfectly regulated power supply, also fiction.

Reality - nonlinear pentodes and drooping power supplies - requires some small tweaks, but still, 50 mA quiescent and 200 mA peaks cannot co-exist along with class A, single-ended operation. Not without drastic clipping.

The attached 'scope screen capture shows one example of reality - the anode waveform I measured from a small-signal beam tetrode a few years ago. The input was a sine wave; the output is noticeably asymmetrical, with negative-going half cycles several percent bigger than positive going ones.

(Remember, this is anode voltage, and negative voltage peaks correspond to increases in anode current.)

Because of the slight asymmetry in current waveform, we may not need to bias to exactly half of peak current - we may be okay a few percent away from exactly 50%. But we're certainly not going to get away with only one-quarter of the peak currents, as your 50 mA quiescent / 200 mA peak example assumed.

The big-picture view is simple: a mathematically ideal class A amplifier draws a constant average current from the power supply, regardless of output power delivered to the loudspeaker. Therefore, when no power is being delivered to the load, all that power is dissipated by the active device (pentode, in this case.) However, at maximum (sinewave) output to the load, efficiency is 50%, half the incoming DC power is delivered to the load, and only half is dissipated by the pentode. Yup, the active device cools off, as PRR just explained.

There is good reason why pure class A output stages in audio power amplifiers were abandoned many decades ago, except by those unfortunates suffering from a painful (and apparently incurable) addiction to single-ended triode power amplifiers. :)



  • DSC_0562C_889_x_500.png
    773.8 KB · Views: 198

Look at a set of tube curves.
Draw a load line for the SE amplifier equal to the primary impedance as it was designed.
Yes, the max current is just a little more than 2x the quiescent current.
And, at the other phase of the signal, the current is approaching zero (a small percentage of the quiescent current, but not quite yet at clipping).

Now, on the same set of curves draw a load line equal to the DCR of the primary (almost straight up and down).
You will find that at the same signal voltages to the control grid, the maximum tube current may be more like 3, 4, 5, 6, or 7 times or more of the quiescent current, and that high current will be at or near B+. The tube will be dissipating up to 3, 4, 5, 6, 7 times or more of the quiescent power during that half of the cycle.
During the other phase of the signal, the current can not drop below zero, so as low as zero power.
So, we have approximately a 50% duty cycle.
3, 4, 5, 6, or 7 times quiescent power X 50% = 1.5, 2, 2.5, 3, 3.5 times the total plate dissipation of the SE tube.

Class A is only available during the non clipping operation of the amplifier.
And, in reality it is only available in an amplifier that drives the primary impedance, but not when it is driving just the primary DCR. That is not an amplifier when the secondary is shorted.

Does that explain the shorted secondary situation?

Do you have something against Triode SE amplifiers?
Do you have something against Pentode/Beam power SE amplifiers?

As to why doesn't anybody ever design a class A pentode amplifier anymore, How many 6V6 guitar clone amplifiers are being built by hobbiests?
I think some are still available commercially.

When does efficiency enter into whether or not to build an SE amplifier?
Do you criticize OTL amplifiers that way too?

The poor man's explanation of a class A amplifier is that the output tube stands all the power until a signal comes along and causes the passing of a small percentage of that power to the speaker.

Nuff' said.
Last edited:


Joined 2003
Paid Member

Start with an SE pentode amp that uses fixed bias and 50 mA Quiescent.
And with a load, RL, that gives 200mA when the grid to cathode volts = -1V.
....average current of 100mA, or 2X the 50mA quiescent current.
Hotter plate, yes?

An SE amp that idles at 50mA and roars at 100mA will have terrible transient behavior when working speech/music to clipping. Even for motor-driving we rarely tolerate 40% rise. Speech/music, 20% is typical. (There IS a trend to higher current, also the mis-load makes more impressive numbers.)

Will the plate get hotter or colder? Allow for power in the load. With a happy match (that does not jump around on every beat of the music), tube heat will go down. With your 2:1 mis-match, we need to know about the tube losses and how hard it is driven. An ideal device driven to square-wave will dissipate zero. Real tubes, "more".
Now, on the same set of curves draw a load line equal to the DCR of the primary (almost straight up and down).
Ah. You were talking about driving the amp into a nearly zero-ohms load the entire time, while PRR (in post #4) talks about what happens when you remove the short and then revert to an appropriate load impedance.

Mystery solved, the two of you were talking apples and oranges. :)
Do you have something against Triode SE amplifiers?
In audio preamps? Nope. As audio power amp output stages? Yes, they make as much sense as a Jello (TM) suspension-bridge. :)
Do you have something against Pentode/Beam power SE amplifiers?
Not at all - for guitar, where we want quite a lot of low-order harmonic distortion for good clean tones, a SE beam tetrode is my preferred choice - but used in a small-signal stage, not to deliver power to a loudspeaker.

The 'scope screen capture I posted earlier came from a little 2 W DIY tube guitar amp I designed and built a few years ago. The small-signal beam tetrode I used was not the output stage, however, but rather, earlier in the power amp, driving the cathodyne phase-splitter, which in turn drove the output stage. The output stage was push-pull class AB, not SE.

This topology gave me the SE beam tetrode sound I like for guitar "clean tone", along with the better efficiency and lower weight and cost of a class AB push-pull output stage.
As to why doesn't anybody ever design a class A pentode amplifier anymore, How many 6V6 guitar clone amplifiers are being built by hobbiests?
I specifically mentioned SE triode output stages, not pentodes. 6V6s are beam tetrodes, not triodes...

We both know that triodes with their large saturation voltages and large anode resistances make truly terrible power devices, right?
When does efficiency enter into whether or not to build an SE amplifier?
Seriously? Any time the active device is a triode valve, and any time the output power level required is more than a few watts (even with pentode output stages.)

There is very bad efficiency, and there is abysmal efficiency. SE triode output stages are in the latter category.
Do you criticize OTL amplifiers that way too?
OTL makes complete sense when the output devices are semiconductors, which operate comfortably at peak currents and voltages well matched to the requirements of 4 and 8 ohm speakers delivering typical SPL levels.

And yes, OTL amps make no sense at all to me with vacuum tube active devices and normal 4 ohm or 8 ohm speakers, which require far higher peak currents than reasonable-sized vacuum tubes can supply.

Phillips had to use 800 ohm speakers to make the things practical - meaning peak output currents were ten times lower for the same output power, compared to an 8 ohm speaker.

I was talking about a properly designed SE tube amp.

But someone thought that even if you short the output, the amp will be safe.
I argue that a couple of things can happen:

1. A full scale signal is applied to the amp, and the shorted output causes Higher than normal dissipation in the output tube (because of the extremely steep load line, current rises and the plate voltage stays near B+ and does not drop down).
Is that to hard to understand?
No, agreed, that is neither a normal operation, nor a good sounding operation.
But some people turn the volume up all the way because they do not hear anything.
That can happen with the speaker disconnected and no load on the amp.
And that can happen with a short on the amp output.
Your choice.

2. A short on an amplifier output that employs negative feedback that precedes the output transformer secondary (perhaps schade feedback, or perhaps output plate to driver cathode feedback, or other) might cause an instability with, or without signal applied.
In that case it might become a full power oscillator, and that again trying to drive DCR as the load.

Try those with every SE amplifier. Depending on the amplifier, the mileage may vary.

Since this is a tube part of this forum, I would refer others to such shorted output with amplfiers that are at the other part of this forum . . . solid state.
The problem with the original Tacoma narrows suspension bridge was twofold:
It was resonant, but too stiff to stand up when it was put in motion.
It exceeded the theoretical motion limits of concrete, before it blew apart.

All good bridges have some flex. Jello or not, there is flex.
Lack of flex may not be very good.

Many mechanical constructions have resonances.
I think the key is using damping in order to deal with the resonances.

I am sorry, I did not make it clear, but because this part of this forum is vacuum tubes, so what I meant was vacuum tube OTL amps.

I am not sure how efficient 800 Ohm voice coils are, versus modern day voice coil loudspeakers.
Can you show me a link that references the efficiency of such an 800 Ohm loudspeaker, in modern terms that todays loudspeakers use, dB/watt at 1 meter?
I am tired of re-calculating the old dB/mW at some large distance, to get dB/W at 1 meter, although I have done that many times.

I believe you meant that triodes have small plate resistances (not large plate resistances of a pentode or beam power tube), right?
Last edited:

Interesting reading!

I got the idea you were more into using beam tetrode for large output power, versus using pentode for large output power.

I may have read it wrong, but I thought you meant that decision was in terms of efficiency; and not so much a consideration of possible sound differences.
Or was the selection of beam tetrode both for efficiency and sound, versus using pentodes?
Did I understand that correctly, or not?


I do not know much about guitar amps, but I am learning.

For Hi Fi . .
I like SE amps, and push pull amps.
I think all the multi grid push pull amps I have had are beam tetrode amps, except when I had EL84 amps and EL34 amps.

I am too slow at typing, and writing. So sometimes when I post, there were 1 or more posts after I got started, so I did not even know it, and that might have caused me to change, edit, or add to what I said, Sorry.
Last edited:
I got the idea you were more into using beam tetrode for large output power, versus using pentode for large output power.
True...an entirely subjective preference on my part. I'm only interested in tubes for guitar amps, and the traditional choices at the roughly 5 - 15 watt output power level are the EL84 pentode, or 6V6 beam tetrode. Of those two, I find I always seem to prefer the sound of the 6V6 amps, particularly the SE ones.

diyAudio memberTubelab (George) once wrote something like "EL84s rock, while 6V6s sing the blues". That neatly describes my preference for the 6V6, I think; smoother, less extreme, less harsh guitar overdrive with a very gradual onset as you turn up the playing dynamics.

In Schade's white-paper about the design of the 6L6 beam tetrode, there is a section where he comments that by nature, beam tetrodes tend to have higher second harmonic distortion compared to otherwise-similar pentodes, while the pentodes tend to have higher third harmonic. I suspect that this might be the characteristic that makes me prefer the sound of the 6V6 (beam tetrode) over the similar EL84 (true pentode.)

From there, things get more complicated. I like the clean sound of a SE 6V6 guitar amp, but to my ears, the sound gets too harsh and ragged when overdriven hard. IMO, push-pull 6V6 output stages sound better when pushed hard than single-ended ones.

I speculate that this has to do with the fact that a symmetrical PP output stage only generates odd harmonics when driven hard, while an SE output stage is asymmetrical, and will generate lots of both even and odd harmonics when pushed deep into overdrive.

So I did some experimenting to see if I could get the sound of an SE beam tetrode when the amplifier was not being pushed hard, transitioning to the sound of a clipping push-pull output stage as the signal level increased.

The plan was to start by finding a small-signal beam tetrode (which would hopefully produce similar-sounding distortion to a 6V6, but at milliwatt power). That turned out to be harder than I thought, because the datasheets often say "pentode" even when the valve itself turns out to be a beam tetrode!

In the end, I bought a few different types of unloved one-dollar RF "pentodes" and inspected them with a magnifying glass, counting the internal grid support rods to see if there were three grids (pentode) or only two (beam tetrode); both the 6AG5 "pentode", and the "pentode" section of a 6JW8, turned out to actually be little beam tetrodes.

In the end, I think I got pretty close to the original goal with the topology I mentioned earlier; the power amp input stage is a 6AG5, followed by a fixed resistive attenuator, which drives a MOSFET used as a "FET-O-Dyne" phase splitter, which in turn drives the push-pull output valves.

At relatively small input levels, the 6AG5 provides an asymmetrical waveform with audible (and clearly visible on a 'scope) amounts of nice-sounding low order harmonic distortion, for nice "clean" guitar tones. The attenuator between 6AG5 and Fet-O-Dyne keeps the FET stage and output valves from being overdriven at this point.

As the input signal increases, the 6AG5 and the two output valves start to become overdriven (the FET stage has enormous headroom and doesn't clip). As signal level continues to increase, the output waveform becomes pretty symmetrical, as I was hoping it would, so that heavily overdriven guitar sounds are not as harsh as a SE Champ, but sound more like a typical PP output stage.
I do not know much about guitar amps, but I am learning.
Same here...I think the tricky part is that all the greatest artistic benefits seem to come from technical flaws in the design. So we're trying to design bad audio amplifiers that happen to sound good when a guitar is plugged into them...and even that is subjective, and depends on the tastes of the guitarist. :)


Wow! . . . I am impressed.
I am learning a lot about guitar amplifier considerations, and much from your post.


I mostly go for closer to the Blues (and softer Jazz sounds), rather than the hard and harsh rock sounds.

You really have done a lot of building and listening in order to do all of that research.

I am still working at all kinds of designs of Hi Fi amplifiers. As you say, so much different than guitar amps.

As to Beam Tetrode, versus Pentode, sometimes it is hard to find what you want:

The 6BQ5 is supposed to be a Beam Tetrode (or Beam Power) tube, same thing, different name.
The EL84 is supposed to be a Pentode.
The data sheet specifications are extremely similar.

Now, go and purchase a 6BQ5; and purchase an EL84.
But when you get them, they are Stamped with 6BQ5/EL84, and EL84/6BQ5.
That can not be right.
And you have to look at the construction to find out what each tube actually is.

Baaaaad. Should make the Manufacturer and the Vendor feel Sheepish.
I'm talking Self-Biased amplifier !!!!

Drive it HARD, the current often goes DOWN.
I've seen the behaviour you describe in a small-signal, fixed bias, beam tetrode gain stage. When driven hard, firstly, the average screen grid voltage fell (because of the increased screen current peaks), lowering anode current. Secondly, the cathode bypass cap started to charge positive, biasing the valve colder, and also lowering anode current.

Results of plate current and plate dissipation extremes when shorting the secondary on an SE amplifier according to the combinations of:

Triode output stage
Pentode or Beam Tetrode output stage
Triode Wired Pentode or Triode Wired Beam Tetrode output stage

Each of those with:
Self Bias
Fixed Bias

Just a few combinations to check.

The results will be several and varied.
Last edited:
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.