Triode vs. Pentode

Trelja...Because music is "dynamic, unpredictable, ever changing" it makes a lousy tool for testing. Results with simple signals like a 1KHz sine wave, or a square wave, or whatever, are repeatable and therefore useful to reveal the effect of various design changes that one might make. The test signals can also be correllated with particular amplifier characteristics that you think are important. Of course the ultimate test is how it sounds with music, but you would never get to a sound that you would like without a lot of work with test signals.

That's why they make vanilla AND chocolate!

Some people are more concerned with test tones, some prefer music - to each, his own...

>"Some people are more concerned with test tones, some prefer music"

>"Of course the ultimate test is how it sounds with music"

Glad we all agree on that. Now that we have nothing to prove to one another, what say we get back to the issue at hand...

Trelja, I don't think of anything you've said here as "flying in the face of conventional wisdom" or likely to "spark a tremendous amount of controversy". Subjectively, I've agreed with almost all of what you've stated about what you hear (about the only item I couldn't quite relate to from my own experience was your description of triode as having more "sparkle"). What I'm trying to do is offer some possible explanations for what we hear. (BTW, my comments about the volume compensation related to doing this by ear - I don't own a meter, although I suppose I ought to.)

There's another, older thread on this same subject that basically illustrates the evolution of my views on the tetrode/triode subject.

Reading over it again, it occurs to me that it really took me over one year and a change in output tube type to come to the conclusion I hold today - namely, that it was naive of me to ever suppose that by merely flipping the tetrode/triode switch, I was conducting a test that could inform me which mode was 'inherently' superior. Getting to that requires digging a little bit deeper.

When I started my comparisons, I essentially disregarded the fact that the output power increases 100% in tetrode, treating it as just an incidental factor, which one had to be mindful of in terms of compensating for matched volumes at the listening position, but no more than that.

That thinking was a mistake, as I eventually came to realize. The difference in power is an integral, confounding, factor - one which cannot be written off as beside the point when comparing modes. It so happens that in my system, through my speakers, tetrode does hold some advantages, and not just with certain music or at certain volume levels. But I now no longer ascribe those advantages to the mode itself, but rather to the increased power, an important distinction.

It's interesting to note that a good portion of the advantages demonstrated in tetrode (again, due - I believe - to the additional power) do seem to be largely independent of absolute volume level. Even at moderate listening levels, and despite the superiority of triode regarding most *musical* qualities, the extra power of tetrode still holds some of the cards regarding what I call the "physical" qualities of the reproduction.

This observation gets to the question of why higher power is desirable. We are always reminded that it's the first watt which counts the most, and I can't disagree with the fundamental astuteness of that aphorism. But the reason they make amplifiers way above 200 watts isn't only to play at 115dB in rooms the size of a small hangar.

I don't have the kind of wide audio exposure to say that I've done conclusive research in this area, but my suspicion is that you can almost literally never have too much power - even for playing average-sized, average-sensitivity speakers in a normal room, and even at moderate volume levels. Whether obtaining that power presents other problems which can compromise the theoretical benefit is a separate question.

Last night I did some fresh comparisons between tetrode and triode with my VTL 185's running KT-88EH's. The amps are rated at around 100w in triode and 200w in tetrode into 4 ohms, the nominal impedance of my Thiel 2.2's.

Even at volumes quiet enough not to wake my gal sleeping upstairs - and even though triode was unquestionably superior at presenting a natural-sounding broad middle of the audioband - tetrode was still superior for bass control, dynamic range, extension at the far ends of the frequency spectrum, and clarity of the space around images. I think if I had a similar amplifier available which doubled the power again to 400w, I would have continued to hear some improvements in those areas - at the same low volume.

Unfortunately, tetrode also carried its penalty along with the extra power: a less-organic, more 'electronic'-sounding portrayal, with the timbral balance tipped by an artificially bright scrim that stripped some meat off the bones and made images seem 'pointier'. It was less naturally compelling and believable, even though the tom toms did pop out of the mix a little more, the bass was more defined, and the soundspace air was a bit more 'see-through'. These qualities may not be terribly *musically* important in many instances, but improving them would constitute higher fidelity, were it not for the trade-offs.

So the downsides of tetrode I attribute to the mode, the upsides to the additional power. If I wanted mo' better than I've already got, I would need an amp that developed higher power in triode.

Exactly *why* doubling already fairly high rated power should make any improvements even when the volume is kept low is not a question I pretend to know all the answers for. But it seems logical to assume some combination of improved driver control and greater freedom from dynamic compression, despite that the average power draw (and I stress average - momentary peak power demand might be more of a mystery) at low volume must only be a watt or less.

While composing this post, I took a break to repeat an experiment which I first mentioned on the other thread linked above. I listened in triode mode to my amps with all 6 of their output tubes per monoblock installed, and then compared the sound after removing 2/3 of the tubes, leaving just the minimum of one pair per amp, still set for triode.

There's a little time lost to the mechanical process and also readjusting the bias, but not more than a few minutes, so I repeated this several times back and forth, auditioning the same cut. Of course volume must be compensated for at the preamp, but even cutting nominal rated power from about 100w down to about 30w didn't seem to require more than around 3.5dB's worth of adjustment to offset. I listened at a low overall volume because once again it's the middle of the night.

The differences are interesting. When configured for 1/3 power, the amps' tonal balance changes. The high treble becomes comparatively shelved-down a bit, and the mids take on new prominence. Overall, the balance sounds enjoyably 'fatter', but not like I'm listening through a pillow or anything. I won't say the bass increased in level, or even necessarily proportionately, but the mid-to-upper-bass seemed a bit plusher, the lower bass more vague. The combined effect was to tilt the spectrum to be weighted more heavily in the range of male vocals, with an intimate quality which was quite attractive replaying same.

Reinstalling the full tube complement (still in triode) produced sound which by comparison seemed more 'hi-fi' - the upper treble was more highlighted, the bass became a little more defined and correspondingly less cushiony, and the heart of the midrange was comparitively demphasized. Notice how closely this seems to resemble what we hear when switching from triode to tetrode! How much is the responsibility of the mode alone, and how much of the additional power it brings?

The sound reacquired basically the similar set of 'physical' attributes I described above in comparing higher-powered tetrode to lower-powered triode - more dynamic contrast, more extension, tighter grip, more explicit deliniation of space with a clearer atmosphere. But listening as I was to a rythym & blues song recorded in the early 60's and remastered rather brightly for CD (it was what happened to be perched in the diskolater at the moment - what can I say?...I'm just lazy, and I dug it), in some ways I actually prefered the lower-power rendering, which sounded perhaps more appropriate for the material. Maybe this is part of why people go gaga over really low-powered SET's.

Trying to set aside the balance differences between the two presentations, it did almost seem to me that possibly the 2-tube rendering had an extra bit of 'purity' vs. the 6-tube version.

This could appear to be so for more than one reason; the first that comes to mind is that with less power and the balance changes I've noted, I could no longer hear the problems with the CD mastering job quite as well.

Some might suggest that a single pair of output devices will function with better symmetry than multiple-paralleled pairs (probably the same people who would say that just a single output device is even better than a single pair).

Another possibility is that there was some advantage in the fact that at 1/3 power, the power supply was now highly overspec'ed, relative to how it is called upon to deliver at full tube strength. Certainly my ExactPower's WRMS readout showed the reduced heater filament demand dropped the total quiescent power consumption by about half. Remember, this amp's power supply is designed to handle all 6 tubes in tetrode for 200w, now only dealing with 2 tubes in triode good for 30w.

That's extra beef without doubt - but who knows what consequences can be traced to what causes? There are too many variables to be sure. For instance, the combined output impedance of the power tubes will presumably grow higher in the less-massively paralleled 2-tube configuration. Since VTL optimizes their power transformers to the power tubes' source impedance for each amp model, this ratio will be sub-optimal with 2/3 of the tubes removed, meaning that not only may power transfer be less efficient, but also that final output impedance may change a little, with possibly audible response modification at the speaker. You just can't take everything into account to draw firm conclusions by listening alone.

But still, this experiment does appear to lend some support for my theory about what higher power means to tetrode mode and the sonic comparison with triode. Anyway, if fidelity is what one is after, then for the time being I'm comfortable with my contention that triode and higher power are what one should want in a tube amplifier.

PS - By the way, for those with mode-switchable tube amps and a good FM tuner in their system, one of the best demos you can run for hearing whether triode or tetrode sounds more real and natural in its harmonic palette is to do some comparative auditioning using speaking voice of NPR desk-jockeys. Public radio doesn't add EQ, compression, and reverb to their on-air voices the way commercial stations (or most recording artists) do, and the human speaking voice is a highly diagnostic tool for assessing the veracity of timbral reproduction. This seems to be true even if we have never met the speaker in person.

Zaikesman and Trelja, this has been one of the most informative posts I have read - thank you. What determines which non-triode amps can be converted to Triode mode?
For example, Joe has told me that the Jolida 502B can be.

Hello Springbok. Most commercial push/pull tube amplifiers can be wired for triode output. Another great post here Zaike and some interesting observations. Depending on the type of output tube used for comparing each mode can also have an influence on your observation here. Example:observing the change in the sound of an amplifier employing a KT-88 output tube [beam power tube] when switching from ultralinear mode to triode mode may or may not be consistent with your analogy when the amplifier employs an el-34[pentode]output tube, not to mention the power factor or speaker load the amplifier is reacting to. The quality of the transformers and the amount of negative feedback employed in the output stage will be a major factor and will have a large influence as well.